diff --git "a/QLoRA_french_sft/trainer_state.json" "b/QLoRA_french_sft/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/QLoRA_french_sft/trainer_state.json"
@@ -0,0 +1,18601 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 23217,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.001292156609381057,
+      "grad_norm": 0.9450203776359558,
+      "learning_rate": 4.999997711254574e-05,
+      "loss": 1.1858,
+      "num_input_tokens_seen": 81920,
+      "step": 10
+    },
+    {
+      "epoch": 0.002584313218762114,
+      "grad_norm": 0.6408165097236633,
+      "learning_rate": 4.999990845022488e-05,
+      "loss": 1.5568,
+      "num_input_tokens_seen": 163840,
+      "step": 20
+    },
+    {
+      "epoch": 0.003876469828143171,
+      "grad_norm": 0.5522646307945251,
+      "learning_rate": 4.999979401316311e-05,
+      "loss": 1.2099,
+      "num_input_tokens_seen": 245760,
+      "step": 30
+    },
+    {
+      "epoch": 0.005168626437524228,
+      "grad_norm": 0.5057278871536255,
+      "learning_rate": 4.999963380156999e-05,
+      "loss": 1.1902,
+      "num_input_tokens_seen": 327680,
+      "step": 40
+    },
+    {
+      "epoch": 0.006460783046905285,
+      "grad_norm": 0.40978753566741943,
+      "learning_rate": 4.9999427815738856e-05,
+      "loss": 1.0802,
+      "num_input_tokens_seen": 409600,
+      "step": 50
+    },
+    {
+      "epoch": 0.007752939656286342,
+      "grad_norm": 0.4875193238258362,
+      "learning_rate": 4.999917605604688e-05,
+      "loss": 1.0393,
+      "num_input_tokens_seen": 491520,
+      "step": 60
+    },
+    {
+      "epoch": 0.009045096265667399,
+      "grad_norm": 0.5074454545974731,
+      "learning_rate": 4.999887852295502e-05,
+      "loss": 1.1473,
+      "num_input_tokens_seen": 573440,
+      "step": 70
+    },
+    {
+      "epoch": 0.010337252875048455,
+      "grad_norm": 0.6053259968757629,
+      "learning_rate": 4.9998535217008054e-05,
+      "loss": 0.8868,
+      "num_input_tokens_seen": 655360,
+      "step": 80
+    },
+    {
+      "epoch": 0.011629409484429512,
+      "grad_norm": 0.5783246159553528,
+      "learning_rate": 4.999814613883459e-05,
+      "loss": 0.9585,
+      "num_input_tokens_seen": 737280,
+      "step": 90
+    },
+    {
+      "epoch": 0.01292156609381057,
+      "grad_norm": 0.5514815449714661,
+      "learning_rate": 4.999771128914701e-05,
+      "loss": 1.0241,
+      "num_input_tokens_seen": 819200,
+      "step": 100
+    },
+    {
+      "epoch": 0.014213722703191626,
+      "grad_norm": 0.6433015465736389,
+      "learning_rate": 4.999723066874154e-05,
+      "loss": 1.034,
+      "num_input_tokens_seen": 901120,
+      "step": 110
+    },
+    {
+      "epoch": 0.015505879312572683,
+      "grad_norm": 0.5091115832328796,
+      "learning_rate": 4.9996704278498185e-05,
+      "loss": 1.2244,
+      "num_input_tokens_seen": 983040,
+      "step": 120
+    },
+    {
+      "epoch": 0.016798035921953742,
+      "grad_norm": 0.40583527088165283,
+      "learning_rate": 4.9996132119380764e-05,
+      "loss": 1.266,
+      "num_input_tokens_seen": 1064960,
+      "step": 130
+    },
+    {
+      "epoch": 0.018090192531334797,
+      "grad_norm": 0.6629675626754761,
+      "learning_rate": 4.999551419243691e-05,
+      "loss": 1.1523,
+      "num_input_tokens_seen": 1146880,
+      "step": 140
+    },
+    {
+      "epoch": 0.019382349140715856,
+      "grad_norm": 0.6925719976425171,
+      "learning_rate": 4.9994850498798026e-05,
+      "loss": 1.1496,
+      "num_input_tokens_seen": 1228800,
+      "step": 150
+    },
+    {
+      "epoch": 0.02067450575009691,
+      "grad_norm": 0.43508079648017883,
+      "learning_rate": 4.999414103967934e-05,
+      "loss": 0.9302,
+      "num_input_tokens_seen": 1310720,
+      "step": 160
+    },
+    {
+      "epoch": 0.02196666235947797,
+      "grad_norm": 0.6177295446395874,
+      "learning_rate": 4.9993385816379876e-05,
+      "loss": 0.9475,
+      "num_input_tokens_seen": 1392640,
+      "step": 170
+    },
+    {
+      "epoch": 0.023258818968859025,
+      "grad_norm": 0.43350470066070557,
+      "learning_rate": 4.999258483028243e-05,
+      "loss": 1.0703,
+      "num_input_tokens_seen": 1474560,
+      "step": 180
+    },
+    {
+      "epoch": 0.024550975578240083,
+      "grad_norm": 0.8424047231674194,
+      "learning_rate": 4.999173808285362e-05,
+      "loss": 1.1235,
+      "num_input_tokens_seen": 1556480,
+      "step": 190
+    },
+    {
+      "epoch": 0.02584313218762114,
+      "grad_norm": 0.5010475516319275,
+      "learning_rate": 4.999084557564383e-05,
+      "loss": 1.3278,
+      "num_input_tokens_seen": 1638400,
+      "step": 200
+    },
+    {
+      "epoch": 0.027135288797002197,
+      "grad_norm": 0.5917531847953796,
+      "learning_rate": 4.9989907310287243e-05,
+      "loss": 1.1928,
+      "num_input_tokens_seen": 1720320,
+      "step": 210
+    },
+    {
+      "epoch": 0.028427445406383253,
+      "grad_norm": 0.40859103202819824,
+      "learning_rate": 4.998892328850181e-05,
+      "loss": 1.215,
+      "num_input_tokens_seen": 1802240,
+      "step": 220
+    },
+    {
+      "epoch": 0.02971960201576431,
+      "grad_norm": 0.7122620344161987,
+      "learning_rate": 4.9987893512089276e-05,
+      "loss": 1.3318,
+      "num_input_tokens_seen": 1884160,
+      "step": 230
+    },
+    {
+      "epoch": 0.031011758625145366,
+      "grad_norm": 0.6804201602935791,
+      "learning_rate": 4.998681798293516e-05,
+      "loss": 1.3226,
+      "num_input_tokens_seen": 1966080,
+      "step": 240
+    },
+    {
+      "epoch": 0.03230391523452642,
+      "grad_norm": 0.5440483093261719,
+      "learning_rate": 4.998569670300876e-05,
+      "loss": 1.1795,
+      "num_input_tokens_seen": 2048000,
+      "step": 250
+    },
+    {
+      "epoch": 0.033596071843907484,
+      "grad_norm": 1.6514626741409302,
+      "learning_rate": 4.9984529674363114e-05,
+      "loss": 0.9646,
+      "num_input_tokens_seen": 2129920,
+      "step": 260
+    },
+    {
+      "epoch": 0.03488822845328854,
+      "grad_norm": 0.46813198924064636,
+      "learning_rate": 4.998331689913506e-05,
+      "loss": 1.1236,
+      "num_input_tokens_seen": 2211840,
+      "step": 270
+    },
+    {
+      "epoch": 0.036180385062669594,
+      "grad_norm": 0.6068572998046875,
+      "learning_rate": 4.998205837954518e-05,
+      "loss": 0.9761,
+      "num_input_tokens_seen": 2293760,
+      "step": 280
+    },
+    {
+      "epoch": 0.03747254167205065,
+      "grad_norm": 0.4501520097255707,
+      "learning_rate": 4.998075411789783e-05,
+      "loss": 1.0846,
+      "num_input_tokens_seen": 2375680,
+      "step": 290
+    },
+    {
+      "epoch": 0.03876469828143171,
+      "grad_norm": 0.6213204264640808,
+      "learning_rate": 4.9979404116581104e-05,
+      "loss": 0.9604,
+      "num_input_tokens_seen": 2457600,
+      "step": 300
+    },
+    {
+      "epoch": 0.04005685489081277,
+      "grad_norm": 0.5889080166816711,
+      "learning_rate": 4.9978008378066844e-05,
+      "loss": 1.135,
+      "num_input_tokens_seen": 2539520,
+      "step": 310
+    },
+    {
+      "epoch": 0.04134901150019382,
+      "grad_norm": 0.5029394030570984,
+      "learning_rate": 4.997656690491064e-05,
+      "loss": 1.0162,
+      "num_input_tokens_seen": 2621440,
+      "step": 320
+    },
+    {
+      "epoch": 0.04264116810957488,
+      "grad_norm": 0.4100414216518402,
+      "learning_rate": 4.9975079699751825e-05,
+      "loss": 1.0928,
+      "num_input_tokens_seen": 2703360,
+      "step": 330
+    },
+    {
+      "epoch": 0.04393332471895594,
+      "grad_norm": 0.5180526971817017,
+      "learning_rate": 4.997354676531348e-05,
+      "loss": 1.4001,
+      "num_input_tokens_seen": 2785280,
+      "step": 340
+    },
+    {
+      "epoch": 0.045225481328336994,
+      "grad_norm": 0.47328633069992065,
+      "learning_rate": 4.997196810440239e-05,
+      "loss": 1.2969,
+      "num_input_tokens_seen": 2867200,
+      "step": 350
+    },
+    {
+      "epoch": 0.04651763793771805,
+      "grad_norm": 0.4311719834804535,
+      "learning_rate": 4.997034371990907e-05,
+      "loss": 1.0298,
+      "num_input_tokens_seen": 2949120,
+      "step": 360
+    },
+    {
+      "epoch": 0.04780979454709911,
+      "grad_norm": 0.8229668736457825,
+      "learning_rate": 4.9968673614807787e-05,
+      "loss": 1.1618,
+      "num_input_tokens_seen": 3031040,
+      "step": 370
+    },
+    {
+      "epoch": 0.04910195115648017,
+      "grad_norm": 0.4691760540008545,
+      "learning_rate": 4.9966957792156475e-05,
+      "loss": 0.8982,
+      "num_input_tokens_seen": 3112960,
+      "step": 380
+    },
+    {
+      "epoch": 0.05039410776586122,
+      "grad_norm": 0.539076566696167,
+      "learning_rate": 4.99651962550968e-05,
+      "loss": 1.15,
+      "num_input_tokens_seen": 3194880,
+      "step": 390
+    },
+    {
+      "epoch": 0.05168626437524228,
+      "grad_norm": 0.3812846541404724,
+      "learning_rate": 4.996338900685414e-05,
+      "loss": 1.3498,
+      "num_input_tokens_seen": 3276800,
+      "step": 400
+    },
+    {
+      "epoch": 0.05297842098462334,
+      "grad_norm": 0.4904411733150482,
+      "learning_rate": 4.996153605073756e-05,
+      "loss": 0.8624,
+      "num_input_tokens_seen": 3358720,
+      "step": 410
+    },
+    {
+      "epoch": 0.054270577594004395,
+      "grad_norm": 0.49068745970726013,
+      "learning_rate": 4.9959637390139814e-05,
+      "loss": 1.1246,
+      "num_input_tokens_seen": 3440640,
+      "step": 420
+    },
+    {
+      "epoch": 0.05556273420338545,
+      "grad_norm": 0.47231003642082214,
+      "learning_rate": 4.995769302853733e-05,
+      "loss": 0.8941,
+      "num_input_tokens_seen": 3522560,
+      "step": 430
+    },
+    {
+      "epoch": 0.056854890812766505,
+      "grad_norm": 0.4713283181190491,
+      "learning_rate": 4.995570296949024e-05,
+      "loss": 1.1223,
+      "num_input_tokens_seen": 3604480,
+      "step": 440
+    },
+    {
+      "epoch": 0.05814704742214757,
+      "grad_norm": 0.9544013738632202,
+      "learning_rate": 4.995366721664234e-05,
+      "loss": 0.7063,
+      "num_input_tokens_seen": 3686400,
+      "step": 450
+    },
+    {
+      "epoch": 0.05943920403152862,
+      "grad_norm": 0.43465104699134827,
+      "learning_rate": 4.995158577372107e-05,
+      "loss": 0.8703,
+      "num_input_tokens_seen": 3768320,
+      "step": 460
+    },
+    {
+      "epoch": 0.06073136064090968,
+      "grad_norm": 6.194292068481445,
+      "learning_rate": 4.9949458644537556e-05,
+      "loss": 0.8798,
+      "num_input_tokens_seen": 3850240,
+      "step": 470
+    },
+    {
+      "epoch": 0.06202351725029073,
+      "grad_norm": 0.4813326895236969,
+      "learning_rate": 4.9947285832986553e-05,
+      "loss": 1.154,
+      "num_input_tokens_seen": 3932160,
+      "step": 480
+    },
+    {
+      "epoch": 0.0633156738596718,
+      "grad_norm": 0.37111344933509827,
+      "learning_rate": 4.9945067343046494e-05,
+      "loss": 1.179,
+      "num_input_tokens_seen": 4014080,
+      "step": 490
+    },
+    {
+      "epoch": 0.06460783046905284,
+      "grad_norm": 0.473379909992218,
+      "learning_rate": 4.9942803178779396e-05,
+      "loss": 1.3203,
+      "num_input_tokens_seen": 4096000,
+      "step": 500
+    },
+    {
+      "epoch": 0.0658999870784339,
+      "grad_norm": 0.31630370020866394,
+      "learning_rate": 4.994049334433095e-05,
+      "loss": 0.95,
+      "num_input_tokens_seen": 4177920,
+      "step": 510
+    },
+    {
+      "epoch": 0.06719214368781497,
+      "grad_norm": 0.6394200921058655,
+      "learning_rate": 4.9938137843930466e-05,
+      "loss": 0.721,
+      "num_input_tokens_seen": 4259840,
+      "step": 520
+    },
+    {
+      "epoch": 0.06848430029719602,
+      "grad_norm": 0.4896668493747711,
+      "learning_rate": 4.993573668189083e-05,
+      "loss": 1.1505,
+      "num_input_tokens_seen": 4341760,
+      "step": 530
+    },
+    {
+      "epoch": 0.06977645690657708,
+      "grad_norm": 0.4696863293647766,
+      "learning_rate": 4.9933289862608584e-05,
+      "loss": 1.1169,
+      "num_input_tokens_seen": 4423680,
+      "step": 540
+    },
+    {
+      "epoch": 0.07106861351595814,
+      "grad_norm": 0.4316001236438751,
+      "learning_rate": 4.9930797390563834e-05,
+      "loss": 1.0893,
+      "num_input_tokens_seen": 4505600,
+      "step": 550
+    },
+    {
+      "epoch": 0.07236077012533919,
+      "grad_norm": 0.47244027256965637,
+      "learning_rate": 4.9928259270320295e-05,
+      "loss": 0.6521,
+      "num_input_tokens_seen": 4587520,
+      "step": 560
+    },
+    {
+      "epoch": 0.07365292673472025,
+      "grad_norm": 0.4212469160556793,
+      "learning_rate": 4.992567550652525e-05,
+      "loss": 0.9833,
+      "num_input_tokens_seen": 4669440,
+      "step": 570
+    },
+    {
+      "epoch": 0.0749450833441013,
+      "grad_norm": 0.49635788798332214,
+      "learning_rate": 4.992304610390955e-05,
+      "loss": 1.1048,
+      "num_input_tokens_seen": 4751360,
+      "step": 580
+    },
+    {
+      "epoch": 0.07623723995348236,
+      "grad_norm": 0.46898409724235535,
+      "learning_rate": 4.9920371067287645e-05,
+      "loss": 0.8898,
+      "num_input_tokens_seen": 4833280,
+      "step": 590
+    },
+    {
+      "epoch": 0.07752939656286342,
+      "grad_norm": 0.7145845293998718,
+      "learning_rate": 4.9917650401557505e-05,
+      "loss": 0.9906,
+      "num_input_tokens_seen": 4915200,
+      "step": 600
+    },
+    {
+      "epoch": 0.07882155317224447,
+      "grad_norm": 7.353865623474121,
+      "learning_rate": 4.9914884111700656e-05,
+      "loss": 0.9923,
+      "num_input_tokens_seen": 4997120,
+      "step": 610
+    },
+    {
+      "epoch": 0.08011370978162553,
+      "grad_norm": 0.7385011911392212,
+      "learning_rate": 4.991207220278217e-05,
+      "loss": 0.8434,
+      "num_input_tokens_seen": 5079040,
+      "step": 620
+    },
+    {
+      "epoch": 0.0814058663910066,
+      "grad_norm": 0.6693135499954224,
+      "learning_rate": 4.990921467995064e-05,
+      "loss": 0.6813,
+      "num_input_tokens_seen": 5160960,
+      "step": 630
+    },
+    {
+      "epoch": 0.08269802300038764,
+      "grad_norm": 0.4748353958129883,
+      "learning_rate": 4.9906311548438184e-05,
+      "loss": 0.8735,
+      "num_input_tokens_seen": 5242880,
+      "step": 640
+    },
+    {
+      "epoch": 0.0839901796097687,
+      "grad_norm": 0.5311540365219116,
+      "learning_rate": 4.990336281356042e-05,
+      "loss": 1.1035,
+      "num_input_tokens_seen": 5324800,
+      "step": 650
+    },
+    {
+      "epoch": 0.08528233621914975,
+      "grad_norm": 0.4842514991760254,
+      "learning_rate": 4.9900368480716466e-05,
+      "loss": 0.8329,
+      "num_input_tokens_seen": 5406720,
+      "step": 660
+    },
+    {
+      "epoch": 0.08657449282853082,
+      "grad_norm": 0.6414400935173035,
+      "learning_rate": 4.9897328555388943e-05,
+      "loss": 1.1088,
+      "num_input_tokens_seen": 5488640,
+      "step": 670
+    },
+    {
+      "epoch": 0.08786664943791188,
+      "grad_norm": 0.39652836322784424,
+      "learning_rate": 4.989424304314395e-05,
+      "loss": 1.0048,
+      "num_input_tokens_seen": 5570560,
+      "step": 680
+    },
+    {
+      "epoch": 0.08915880604729293,
+      "grad_norm": 0.6044394969940186,
+      "learning_rate": 4.9891111949631023e-05,
+      "loss": 0.9396,
+      "num_input_tokens_seen": 5652480,
+      "step": 690
+    },
+    {
+      "epoch": 0.09045096265667399,
+      "grad_norm": 0.4656500816345215,
+      "learning_rate": 4.988793528058321e-05,
+      "loss": 1.1961,
+      "num_input_tokens_seen": 5734400,
+      "step": 700
+    },
+    {
+      "epoch": 0.09174311926605505,
+      "grad_norm": 0.5340274572372437,
+      "learning_rate": 4.988471304181697e-05,
+      "loss": 0.9563,
+      "num_input_tokens_seen": 5816320,
+      "step": 710
+    },
+    {
+      "epoch": 0.0930352758754361,
+      "grad_norm": 0.4175763726234436,
+      "learning_rate": 4.988144523923221e-05,
+      "loss": 0.9487,
+      "num_input_tokens_seen": 5898240,
+      "step": 720
+    },
+    {
+      "epoch": 0.09432743248481716,
+      "grad_norm": 0.4877130389213562,
+      "learning_rate": 4.987813187881226e-05,
+      "loss": 0.9847,
+      "num_input_tokens_seen": 5980160,
+      "step": 730
+    },
+    {
+      "epoch": 0.09561958909419822,
+      "grad_norm": 0.5183830261230469,
+      "learning_rate": 4.987477296662387e-05,
+      "loss": 1.3118,
+      "num_input_tokens_seen": 6062080,
+      "step": 740
+    },
+    {
+      "epoch": 0.09691174570357927,
+      "grad_norm": 0.4533514678478241,
+      "learning_rate": 4.987136850881721e-05,
+      "loss": 1.0996,
+      "num_input_tokens_seen": 6144000,
+      "step": 750
+    },
+    {
+      "epoch": 0.09820390231296033,
+      "grad_norm": 0.4317517578601837,
+      "learning_rate": 4.986791851162582e-05,
+      "loss": 0.9023,
+      "num_input_tokens_seen": 6225920,
+      "step": 760
+    },
+    {
+      "epoch": 0.09949605892234138,
+      "grad_norm": 0.7972986698150635,
+      "learning_rate": 4.986442298136663e-05,
+      "loss": 1.0904,
+      "num_input_tokens_seen": 6307840,
+      "step": 770
+    },
+    {
+      "epoch": 0.10078821553172244,
+      "grad_norm": 1.0747971534729004,
+      "learning_rate": 4.986088192443995e-05,
+      "loss": 0.9374,
+      "num_input_tokens_seen": 6389760,
+      "step": 780
+    },
+    {
+      "epoch": 0.1020803721411035,
+      "grad_norm": 0.3895331919193268,
+      "learning_rate": 4.985729534732944e-05,
+      "loss": 0.8173,
+      "num_input_tokens_seen": 6471680,
+      "step": 790
+    },
+    {
+      "epoch": 0.10337252875048455,
+      "grad_norm": 0.6985827684402466,
+      "learning_rate": 4.98536632566021e-05,
+      "loss": 0.9162,
+      "num_input_tokens_seen": 6553600,
+      "step": 800
+    },
+    {
+      "epoch": 0.10466468535986562,
+      "grad_norm": 0.4658990502357483,
+      "learning_rate": 4.9849985658908296e-05,
+      "loss": 0.7986,
+      "num_input_tokens_seen": 6635520,
+      "step": 810
+    },
+    {
+      "epoch": 0.10595684196924668,
+      "grad_norm": 0.3981533944606781,
+      "learning_rate": 4.9846262560981674e-05,
+      "loss": 0.7034,
+      "num_input_tokens_seen": 6717440,
+      "step": 820
+    },
+    {
+      "epoch": 0.10724899857862773,
+      "grad_norm": 0.6919500231742859,
+      "learning_rate": 4.9842493969639215e-05,
+      "loss": 0.9466,
+      "num_input_tokens_seen": 6799360,
+      "step": 830
+    },
+    {
+      "epoch": 0.10854115518800879,
+      "grad_norm": 0.5051977038383484,
+      "learning_rate": 4.9838679891781214e-05,
+      "loss": 1.2844,
+      "num_input_tokens_seen": 6881280,
+      "step": 840
+    },
+    {
+      "epoch": 0.10983331179738984,
+      "grad_norm": 0.49048569798469543,
+      "learning_rate": 4.983482033439122e-05,
+      "loss": 0.7703,
+      "num_input_tokens_seen": 6963200,
+      "step": 850
+    },
+    {
+      "epoch": 0.1111254684067709,
+      "grad_norm": 0.29757964611053467,
+      "learning_rate": 4.9830915304536065e-05,
+      "loss": 1.2312,
+      "num_input_tokens_seen": 7045120,
+      "step": 860
+    },
+    {
+      "epoch": 0.11241762501615196,
+      "grad_norm": 0.4364021122455597,
+      "learning_rate": 4.982696480936586e-05,
+      "loss": 1.3658,
+      "num_input_tokens_seen": 7127040,
+      "step": 870
+    },
+    {
+      "epoch": 0.11370978162553301,
+      "grad_norm": 0.7192522287368774,
+      "learning_rate": 4.9822968856113926e-05,
+      "loss": 1.0453,
+      "num_input_tokens_seen": 7208960,
+      "step": 880
+    },
+    {
+      "epoch": 0.11500193823491407,
+      "grad_norm": 0.5548785924911499,
+      "learning_rate": 4.9818927452096855e-05,
+      "loss": 1.1852,
+      "num_input_tokens_seen": 7290880,
+      "step": 890
+    },
+    {
+      "epoch": 0.11629409484429513,
+      "grad_norm": 0.4914171099662781,
+      "learning_rate": 4.981484060471444e-05,
+      "loss": 1.2232,
+      "num_input_tokens_seen": 7372800,
+      "step": 900
+    },
+    {
+      "epoch": 0.11758625145367618,
+      "grad_norm": 0.4073733687400818,
+      "learning_rate": 4.981070832144967e-05,
+      "loss": 1.1144,
+      "num_input_tokens_seen": 7454720,
+      "step": 910
+    },
+    {
+      "epoch": 0.11887840806305724,
+      "grad_norm": 0.7221460342407227,
+      "learning_rate": 4.980653060986877e-05,
+      "loss": 1.0344,
+      "num_input_tokens_seen": 7536640,
+      "step": 920
+    },
+    {
+      "epoch": 0.1201705646724383,
+      "grad_norm": 0.5011032223701477,
+      "learning_rate": 4.9802307477621084e-05,
+      "loss": 1.1751,
+      "num_input_tokens_seen": 7618560,
+      "step": 930
+    },
+    {
+      "epoch": 0.12146272128181936,
+      "grad_norm": 0.4509833753108978,
+      "learning_rate": 4.9798038932439175e-05,
+      "loss": 1.1844,
+      "num_input_tokens_seen": 7700480,
+      "step": 940
+    },
+    {
+      "epoch": 0.12275487789120042,
+      "grad_norm": 0.6086503863334656,
+      "learning_rate": 4.979372498213871e-05,
+      "loss": 1.0034,
+      "num_input_tokens_seen": 7782400,
+      "step": 950
+    },
+    {
+      "epoch": 0.12404703450058147,
+      "grad_norm": 0.9258208274841309,
+      "learning_rate": 4.978936563461854e-05,
+      "loss": 0.9102,
+      "num_input_tokens_seen": 7864320,
+      "step": 960
+    },
+    {
+      "epoch": 0.12533919110996253,
+      "grad_norm": 0.5192177295684814,
+      "learning_rate": 4.97849608978606e-05,
+      "loss": 0.8154,
+      "num_input_tokens_seen": 7946240,
+      "step": 970
+    },
+    {
+      "epoch": 0.1266313477193436,
+      "grad_norm": 0.37925633788108826,
+      "learning_rate": 4.978051077992994e-05,
+      "loss": 1.0143,
+      "num_input_tokens_seen": 8028160,
+      "step": 980
+    },
+    {
+      "epoch": 0.12792350432872465,
+      "grad_norm": 2.0898265838623047,
+      "learning_rate": 4.9776015288974736e-05,
+      "loss": 0.7723,
+      "num_input_tokens_seen": 8110080,
+      "step": 990
+    },
+    {
+      "epoch": 0.1292156609381057,
+      "grad_norm": 0.667012631893158,
+      "learning_rate": 4.9771474433226194e-05,
+      "loss": 1.2866,
+      "num_input_tokens_seen": 8192000,
+      "step": 1000
+    },
+    {
+      "epoch": 0.13050781754748675,
+      "grad_norm": 0.4613621234893799,
+      "learning_rate": 4.976688822099861e-05,
+      "loss": 1.4743,
+      "num_input_tokens_seen": 8273920,
+      "step": 1010
+    },
+    {
+      "epoch": 0.1317999741568678,
+      "grad_norm": 0.7849400043487549,
+      "learning_rate": 4.976225666068932e-05,
+      "loss": 0.904,
+      "num_input_tokens_seen": 8355840,
+      "step": 1020
+    },
+    {
+      "epoch": 0.13309213076624887,
+      "grad_norm": 0.6418063044548035,
+      "learning_rate": 4.9757579760778697e-05,
+      "loss": 0.9119,
+      "num_input_tokens_seen": 8437760,
+      "step": 1030
+    },
+    {
+      "epoch": 0.13438428737562993,
+      "grad_norm": 0.47972187399864197,
+      "learning_rate": 4.9752857529830125e-05,
+      "loss": 1.1209,
+      "num_input_tokens_seen": 8519680,
+      "step": 1040
+    },
+    {
+      "epoch": 0.135676443985011,
+      "grad_norm": 0.4660983681678772,
+      "learning_rate": 4.9748089976489996e-05,
+      "loss": 1.2202,
+      "num_input_tokens_seen": 8601600,
+      "step": 1050
+    },
+    {
+      "epoch": 0.13696860059439203,
+      "grad_norm": 0.4959300756454468,
+      "learning_rate": 4.9743277109487674e-05,
+      "loss": 1.2102,
+      "num_input_tokens_seen": 8683520,
+      "step": 1060
+    },
+    {
+      "epoch": 0.1382607572037731,
+      "grad_norm": 0.541445791721344,
+      "learning_rate": 4.973841893763551e-05,
+      "loss": 0.9835,
+      "num_input_tokens_seen": 8765440,
+      "step": 1070
+    },
+    {
+      "epoch": 0.13955291381315416,
+      "grad_norm": 0.48562517762184143,
+      "learning_rate": 4.9733515469828795e-05,
+      "loss": 0.9155,
+      "num_input_tokens_seen": 8847360,
+      "step": 1080
+    },
+    {
+      "epoch": 0.14084507042253522,
+      "grad_norm": 0.4447590708732605,
+      "learning_rate": 4.972856671504576e-05,
+      "loss": 1.196,
+      "num_input_tokens_seen": 8929280,
+      "step": 1090
+    },
+    {
+      "epoch": 0.14213722703191628,
+      "grad_norm": 0.5334303379058838,
+      "learning_rate": 4.9723572682347566e-05,
+      "loss": 0.8913,
+      "num_input_tokens_seen": 9011200,
+      "step": 1100
+    },
+    {
+      "epoch": 0.14342938364129731,
+      "grad_norm": 0.376122385263443,
+      "learning_rate": 4.971853338087825e-05,
+      "loss": 1.1649,
+      "num_input_tokens_seen": 9093120,
+      "step": 1110
+    },
+    {
+      "epoch": 0.14472154025067838,
+      "grad_norm": 0.5344268083572388,
+      "learning_rate": 4.971344881986477e-05,
+      "loss": 1.084,
+      "num_input_tokens_seen": 9175040,
+      "step": 1120
+    },
+    {
+      "epoch": 0.14601369686005944,
+      "grad_norm": 0.8270778059959412,
+      "learning_rate": 4.9708319008616926e-05,
+      "loss": 0.9945,
+      "num_input_tokens_seen": 9256960,
+      "step": 1130
+    },
+    {
+      "epoch": 0.1473058534694405,
+      "grad_norm": 0.6520871520042419,
+      "learning_rate": 4.97031439565274e-05,
+      "loss": 0.8976,
+      "num_input_tokens_seen": 9338880,
+      "step": 1140
+    },
+    {
+      "epoch": 0.14859801007882156,
+      "grad_norm": 0.4764862656593323,
+      "learning_rate": 4.969792367307168e-05,
+      "loss": 1.063,
+      "num_input_tokens_seen": 9420800,
+      "step": 1150
+    },
+    {
+      "epoch": 0.1498901666882026,
+      "grad_norm": 0.48043620586395264,
+      "learning_rate": 4.9692658167808094e-05,
+      "loss": 0.8809,
+      "num_input_tokens_seen": 9502720,
+      "step": 1160
+    },
+    {
+      "epoch": 0.15118232329758366,
+      "grad_norm": 0.44339507818222046,
+      "learning_rate": 4.9687347450377755e-05,
+      "loss": 1.3418,
+      "num_input_tokens_seen": 9584640,
+      "step": 1170
+    },
+    {
+      "epoch": 0.15247447990696472,
+      "grad_norm": 0.40918466448783875,
+      "learning_rate": 4.968199153050457e-05,
+      "loss": 0.6324,
+      "num_input_tokens_seen": 9666560,
+      "step": 1180
+    },
+    {
+      "epoch": 0.15376663651634578,
+      "grad_norm": 0.5043272972106934,
+      "learning_rate": 4.967659041799522e-05,
+      "loss": 0.9144,
+      "num_input_tokens_seen": 9748480,
+      "step": 1190
+    },
+    {
+      "epoch": 0.15505879312572685,
+      "grad_norm": 0.47593727707862854,
+      "learning_rate": 4.9671144122739106e-05,
+      "loss": 0.8396,
+      "num_input_tokens_seen": 9830400,
+      "step": 1200
+    },
+    {
+      "epoch": 0.1563509497351079,
+      "grad_norm": 0.3913171887397766,
+      "learning_rate": 4.966565265470838e-05,
+      "loss": 1.0608,
+      "num_input_tokens_seen": 9912320,
+      "step": 1210
+    },
+    {
+      "epoch": 0.15764310634448894,
+      "grad_norm": 0.8421456813812256,
+      "learning_rate": 4.9660116023957906e-05,
+      "loss": 0.9964,
+      "num_input_tokens_seen": 9994240,
+      "step": 1220
+    },
+    {
+      "epoch": 0.15893526295387,
+      "grad_norm": 0.46809521317481995,
+      "learning_rate": 4.9654534240625225e-05,
+      "loss": 0.8427,
+      "num_input_tokens_seen": 10076160,
+      "step": 1230
+    },
+    {
+      "epoch": 0.16022741956325107,
+      "grad_norm": 0.5621390342712402,
+      "learning_rate": 4.964890731493057e-05,
+      "loss": 1.2353,
+      "num_input_tokens_seen": 10158080,
+      "step": 1240
+    },
+    {
+      "epoch": 0.16151957617263213,
+      "grad_norm": 0.4590226113796234,
+      "learning_rate": 4.964323525717681e-05,
+      "loss": 1.0439,
+      "num_input_tokens_seen": 10240000,
+      "step": 1250
+    },
+    {
+      "epoch": 0.1628117327820132,
+      "grad_norm": 0.40506160259246826,
+      "learning_rate": 4.9637518077749476e-05,
+      "loss": 0.914,
+      "num_input_tokens_seen": 10321920,
+      "step": 1260
+    },
+    {
+      "epoch": 0.16410388939139423,
+      "grad_norm": 0.8865328431129456,
+      "learning_rate": 4.96317557871167e-05,
+      "loss": 1.0189,
+      "num_input_tokens_seen": 10403840,
+      "step": 1270
+    },
+    {
+      "epoch": 0.1653960460007753,
+      "grad_norm": 0.37955254316329956,
+      "learning_rate": 4.9625948395829216e-05,
+      "loss": 1.0049,
+      "num_input_tokens_seen": 10485760,
+      "step": 1280
+    },
+    {
+      "epoch": 0.16668820261015635,
+      "grad_norm": 0.5044606328010559,
+      "learning_rate": 4.962009591452032e-05,
+      "loss": 0.941,
+      "num_input_tokens_seen": 10567680,
+      "step": 1290
+    },
+    {
+      "epoch": 0.1679803592195374,
+      "grad_norm": 0.6987674236297607,
+      "learning_rate": 4.96141983539059e-05,
+      "loss": 1.1209,
+      "num_input_tokens_seen": 10649600,
+      "step": 1300
+    },
+    {
+      "epoch": 0.16927251582891847,
+      "grad_norm": 1.0083003044128418,
+      "learning_rate": 4.960825572478436e-05,
+      "loss": 1.0478,
+      "num_input_tokens_seen": 10731520,
+      "step": 1310
+    },
+    {
+      "epoch": 0.1705646724382995,
+      "grad_norm": 0.3237228989601135,
+      "learning_rate": 4.960226803803664e-05,
+      "loss": 0.9342,
+      "num_input_tokens_seen": 10813440,
+      "step": 1320
+    },
+    {
+      "epoch": 0.17185682904768057,
+      "grad_norm": 1.4129977226257324,
+      "learning_rate": 4.959623530462617e-05,
+      "loss": 1.0727,
+      "num_input_tokens_seen": 10895360,
+      "step": 1330
+    },
+    {
+      "epoch": 0.17314898565706163,
+      "grad_norm": 0.3343133330345154,
+      "learning_rate": 4.9590157535598855e-05,
+      "loss": 1.0719,
+      "num_input_tokens_seen": 10977280,
+      "step": 1340
+    },
+    {
+      "epoch": 0.1744411422664427,
+      "grad_norm": 0.3956698775291443,
+      "learning_rate": 4.958403474208308e-05,
+      "loss": 1.0195,
+      "num_input_tokens_seen": 11059200,
+      "step": 1350
+    },
+    {
+      "epoch": 0.17573329887582376,
+      "grad_norm": 0.45558327436447144,
+      "learning_rate": 4.957786693528965e-05,
+      "loss": 1.1703,
+      "num_input_tokens_seen": 11141120,
+      "step": 1360
+    },
+    {
+      "epoch": 0.17702545548520482,
+      "grad_norm": 0.4159637689590454,
+      "learning_rate": 4.95716541265118e-05,
+      "loss": 0.8547,
+      "num_input_tokens_seen": 11223040,
+      "step": 1370
+    },
+    {
+      "epoch": 0.17831761209458585,
+      "grad_norm": 0.510883629322052,
+      "learning_rate": 4.9565396327125155e-05,
+      "loss": 1.056,
+      "num_input_tokens_seen": 11304960,
+      "step": 1380
+    },
+    {
+      "epoch": 0.17960976870396692,
+      "grad_norm": 0.4751374125480652,
+      "learning_rate": 4.955909354858772e-05,
+      "loss": 0.8731,
+      "num_input_tokens_seen": 11386880,
+      "step": 1390
+    },
+    {
+      "epoch": 0.18090192531334798,
+      "grad_norm": 0.3273867070674896,
+      "learning_rate": 4.955274580243987e-05,
+      "loss": 0.7233,
+      "num_input_tokens_seen": 11468800,
+      "step": 1400
+    },
+    {
+      "epoch": 0.18219408192272904,
+      "grad_norm": 0.38997891545295715,
+      "learning_rate": 4.95463531003043e-05,
+      "loss": 0.8429,
+      "num_input_tokens_seen": 11550720,
+      "step": 1410
+    },
+    {
+      "epoch": 0.1834862385321101,
+      "grad_norm": 0.4781230092048645,
+      "learning_rate": 4.953991545388603e-05,
+      "loss": 0.9205,
+      "num_input_tokens_seen": 11632640,
+      "step": 1420
+    },
+    {
+      "epoch": 0.18477839514149114,
+      "grad_norm": 0.39320600032806396,
+      "learning_rate": 4.9533432874972366e-05,
+      "loss": 1.0245,
+      "num_input_tokens_seen": 11714560,
+      "step": 1430
+    },
+    {
+      "epoch": 0.1860705517508722,
+      "grad_norm": 1.2111022472381592,
+      "learning_rate": 4.952690537543287e-05,
+      "loss": 1.161,
+      "num_input_tokens_seen": 11796480,
+      "step": 1440
+    },
+    {
+      "epoch": 0.18736270836025326,
+      "grad_norm": 0.42389845848083496,
+      "learning_rate": 4.952033296721938e-05,
+      "loss": 1.4307,
+      "num_input_tokens_seen": 11878400,
+      "step": 1450
+    },
+    {
+      "epoch": 0.18865486496963432,
+      "grad_norm": 0.3861338496208191,
+      "learning_rate": 4.951371566236597e-05,
+      "loss": 0.9983,
+      "num_input_tokens_seen": 11960320,
+      "step": 1460
+    },
+    {
+      "epoch": 0.18994702157901538,
+      "grad_norm": 0.39819157123565674,
+      "learning_rate": 4.9507053472988867e-05,
+      "loss": 0.927,
+      "num_input_tokens_seen": 12042240,
+      "step": 1470
+    },
+    {
+      "epoch": 0.19123917818839645,
+      "grad_norm": 0.47134819626808167,
+      "learning_rate": 4.9500346411286534e-05,
+      "loss": 1.0184,
+      "num_input_tokens_seen": 12124160,
+      "step": 1480
+    },
+    {
+      "epoch": 0.19253133479777748,
+      "grad_norm": 0.6282069683074951,
+      "learning_rate": 4.949359448953959e-05,
+      "loss": 0.9956,
+      "num_input_tokens_seen": 12206080,
+      "step": 1490
+    },
+    {
+      "epoch": 0.19382349140715854,
+      "grad_norm": 0.48138657212257385,
+      "learning_rate": 4.9486797720110746e-05,
+      "loss": 0.7369,
+      "num_input_tokens_seen": 12288000,
+      "step": 1500
+    },
+    {
+      "epoch": 0.1951156480165396,
+      "grad_norm": 0.42543670535087585,
+      "learning_rate": 4.947995611544489e-05,
+      "loss": 0.9916,
+      "num_input_tokens_seen": 12369920,
+      "step": 1510
+    },
+    {
+      "epoch": 0.19640780462592067,
+      "grad_norm": 0.6821240782737732,
+      "learning_rate": 4.947306968806896e-05,
+      "loss": 1.0502,
+      "num_input_tokens_seen": 12451840,
+      "step": 1520
+    },
+    {
+      "epoch": 0.19769996123530173,
+      "grad_norm": 0.7128183245658875,
+      "learning_rate": 4.946613845059199e-05,
+      "loss": 1.1209,
+      "num_input_tokens_seen": 12533760,
+      "step": 1530
+    },
+    {
+      "epoch": 0.19899211784468276,
+      "grad_norm": 0.5544474720954895,
+      "learning_rate": 4.945916241570504e-05,
+      "loss": 1.152,
+      "num_input_tokens_seen": 12615680,
+      "step": 1540
+    },
+    {
+      "epoch": 0.20028427445406383,
+      "grad_norm": 0.4839724600315094,
+      "learning_rate": 4.945214159618121e-05,
+      "loss": 1.1884,
+      "num_input_tokens_seen": 12697600,
+      "step": 1550
+    },
+    {
+      "epoch": 0.2015764310634449,
+      "grad_norm": 0.381164014339447,
+      "learning_rate": 4.9445076004875596e-05,
+      "loss": 1.1069,
+      "num_input_tokens_seen": 12779520,
+      "step": 1560
+    },
+    {
+      "epoch": 0.20286858767282595,
+      "grad_norm": 0.5115939974784851,
+      "learning_rate": 4.9437965654725264e-05,
+      "loss": 1.0457,
+      "num_input_tokens_seen": 12861440,
+      "step": 1570
+    },
+    {
+      "epoch": 0.204160744282207,
+      "grad_norm": 0.45792338252067566,
+      "learning_rate": 4.943081055874925e-05,
+      "loss": 0.8206,
+      "num_input_tokens_seen": 12943360,
+      "step": 1580
+    },
+    {
+      "epoch": 0.20545290089158805,
+      "grad_norm": 0.5342550873756409,
+      "learning_rate": 4.9423610730048495e-05,
+      "loss": 1.2474,
+      "num_input_tokens_seen": 13025280,
+      "step": 1590
+    },
+    {
+      "epoch": 0.2067450575009691,
+      "grad_norm": 0.6011587381362915,
+      "learning_rate": 4.941636618180586e-05,
+      "loss": 0.9485,
+      "num_input_tokens_seen": 13107200,
+      "step": 1600
+    },
+    {
+      "epoch": 0.20803721411035017,
+      "grad_norm": 1.3028993606567383,
+      "learning_rate": 4.94090769272861e-05,
+      "loss": 1.0329,
+      "num_input_tokens_seen": 13189120,
+      "step": 1610
+    },
+    {
+      "epoch": 0.20932937071973123,
+      "grad_norm": 0.4599340260028839,
+      "learning_rate": 4.940174297983581e-05,
+      "loss": 1.1409,
+      "num_input_tokens_seen": 13271040,
+      "step": 1620
+    },
+    {
+      "epoch": 0.2106215273291123,
+      "grad_norm": 0.4517963230609894,
+      "learning_rate": 4.93943643528834e-05,
+      "loss": 0.6724,
+      "num_input_tokens_seen": 13352960,
+      "step": 1630
+    },
+    {
+      "epoch": 0.21191368393849336,
+      "grad_norm": 0.4794328212738037,
+      "learning_rate": 4.938694105993914e-05,
+      "loss": 0.9835,
+      "num_input_tokens_seen": 13434880,
+      "step": 1640
+    },
+    {
+      "epoch": 0.2132058405478744,
+      "grad_norm": 0.43061280250549316,
+      "learning_rate": 4.937947311459503e-05,
+      "loss": 1.2099,
+      "num_input_tokens_seen": 13516800,
+      "step": 1650
+    },
+    {
+      "epoch": 0.21449799715725545,
+      "grad_norm": 0.4864104986190796,
+      "learning_rate": 4.937196053052486e-05,
+      "loss": 0.9276,
+      "num_input_tokens_seen": 13598720,
+      "step": 1660
+    },
+    {
+      "epoch": 0.21579015376663652,
+      "grad_norm": 0.40697136521339417,
+      "learning_rate": 4.9364403321484145e-05,
+      "loss": 1.4808,
+      "num_input_tokens_seen": 13680640,
+      "step": 1670
+    },
+    {
+      "epoch": 0.21708231037601758,
+      "grad_norm": 0.613544762134552,
+      "learning_rate": 4.9356801501310105e-05,
+      "loss": 1.248,
+      "num_input_tokens_seen": 13762560,
+      "step": 1680
+    },
+    {
+      "epoch": 0.21837446698539864,
+      "grad_norm": 0.25986501574516296,
+      "learning_rate": 4.934915508392164e-05,
+      "loss": 0.8261,
+      "num_input_tokens_seen": 13844480,
+      "step": 1690
+    },
+    {
+      "epoch": 0.21966662359477968,
+      "grad_norm": 0.4318770468235016,
+      "learning_rate": 4.9341464083319314e-05,
+      "loss": 1.2263,
+      "num_input_tokens_seen": 13926400,
+      "step": 1700
+    },
+    {
+      "epoch": 0.22095878020416074,
+      "grad_norm": 0.32169216871261597,
+      "learning_rate": 4.933372851358532e-05,
+      "loss": 0.8885,
+      "num_input_tokens_seen": 14008320,
+      "step": 1710
+    },
+    {
+      "epoch": 0.2222509368135418,
+      "grad_norm": 0.507994532585144,
+      "learning_rate": 4.932594838888347e-05,
+      "loss": 1.369,
+      "num_input_tokens_seen": 14090240,
+      "step": 1720
+    },
+    {
+      "epoch": 0.22354309342292286,
+      "grad_norm": 0.5168066620826721,
+      "learning_rate": 4.931812372345913e-05,
+      "loss": 1.1812,
+      "num_input_tokens_seen": 14172160,
+      "step": 1730
+    },
+    {
+      "epoch": 0.22483525003230392,
+      "grad_norm": 0.5198150873184204,
+      "learning_rate": 4.9310254531639235e-05,
+      "loss": 0.7136,
+      "num_input_tokens_seen": 14254080,
+      "step": 1740
+    },
+    {
+      "epoch": 0.22612740664168499,
+      "grad_norm": 0.4334443509578705,
+      "learning_rate": 4.930234082783225e-05,
+      "loss": 1.4446,
+      "num_input_tokens_seen": 14336000,
+      "step": 1750
+    },
+    {
+      "epoch": 0.22741956325106602,
+      "grad_norm": 0.4047377109527588,
+      "learning_rate": 4.9294382626528144e-05,
+      "loss": 0.7729,
+      "num_input_tokens_seen": 14417920,
+      "step": 1760
+    },
+    {
+      "epoch": 0.22871171986044708,
+      "grad_norm": 1.014719009399414,
+      "learning_rate": 4.928637994229834e-05,
+      "loss": 0.9155,
+      "num_input_tokens_seen": 14499840,
+      "step": 1770
+    },
+    {
+      "epoch": 0.23000387646982814,
+      "grad_norm": 0.5884613394737244,
+      "learning_rate": 4.9278332789795746e-05,
+      "loss": 1.0874,
+      "num_input_tokens_seen": 14581760,
+      "step": 1780
+    },
+    {
+      "epoch": 0.2312960330792092,
+      "grad_norm": 0.4626663029193878,
+      "learning_rate": 4.9270241183754637e-05,
+      "loss": 0.7714,
+      "num_input_tokens_seen": 14663680,
+      "step": 1790
+    },
+    {
+      "epoch": 0.23258818968859027,
+      "grad_norm": 0.42093148827552795,
+      "learning_rate": 4.9262105138990745e-05,
+      "loss": 1.3797,
+      "num_input_tokens_seen": 14745600,
+      "step": 1800
+    },
+    {
+      "epoch": 0.2338803462979713,
+      "grad_norm": 0.3722878396511078,
+      "learning_rate": 4.925392467040112e-05,
+      "loss": 1.2271,
+      "num_input_tokens_seen": 14827520,
+      "step": 1810
+    },
+    {
+      "epoch": 0.23517250290735237,
+      "grad_norm": 0.4904583990573883,
+      "learning_rate": 4.924569979296417e-05,
+      "loss": 0.9982,
+      "num_input_tokens_seen": 14909440,
+      "step": 1820
+    },
+    {
+      "epoch": 0.23646465951673343,
+      "grad_norm": 0.5454955101013184,
+      "learning_rate": 4.9237430521739626e-05,
+      "loss": 0.9374,
+      "num_input_tokens_seen": 14991360,
+      "step": 1830
+    },
+    {
+      "epoch": 0.2377568161261145,
+      "grad_norm": 0.9513818621635437,
+      "learning_rate": 4.9229116871868485e-05,
+      "loss": 0.9464,
+      "num_input_tokens_seen": 15073280,
+      "step": 1840
+    },
+    {
+      "epoch": 0.23904897273549555,
+      "grad_norm": 1.1613630056381226,
+      "learning_rate": 4.922075885857301e-05,
+      "loss": 0.8913,
+      "num_input_tokens_seen": 15155200,
+      "step": 1850
+    },
+    {
+      "epoch": 0.2403411293448766,
+      "grad_norm": 0.5714117288589478,
+      "learning_rate": 4.92123564971567e-05,
+      "loss": 1.3031,
+      "num_input_tokens_seen": 15237120,
+      "step": 1860
+    },
+    {
+      "epoch": 0.24163328595425765,
+      "grad_norm": 0.417524516582489,
+      "learning_rate": 4.9203909803004245e-05,
+      "loss": 1.0151,
+      "num_input_tokens_seen": 15319040,
+      "step": 1870
+    },
+    {
+      "epoch": 0.2429254425636387,
+      "grad_norm": 0.43594273924827576,
+      "learning_rate": 4.9195418791581504e-05,
+      "loss": 1.0122,
+      "num_input_tokens_seen": 15400960,
+      "step": 1880
+    },
+    {
+      "epoch": 0.24421759917301977,
+      "grad_norm": 0.5096331238746643,
+      "learning_rate": 4.918688347843549e-05,
+      "loss": 1.3312,
+      "num_input_tokens_seen": 15482880,
+      "step": 1890
+    },
+    {
+      "epoch": 0.24550975578240083,
+      "grad_norm": 0.4702720046043396,
+      "learning_rate": 4.917830387919434e-05,
+      "loss": 1.4312,
+      "num_input_tokens_seen": 15564800,
+      "step": 1900
+    },
+    {
+      "epoch": 0.2468019123917819,
+      "grad_norm": 0.49178382754325867,
+      "learning_rate": 4.9169680009567254e-05,
+      "loss": 0.8297,
+      "num_input_tokens_seen": 15646720,
+      "step": 1910
+    },
+    {
+      "epoch": 0.24809406900116293,
+      "grad_norm": 0.4930388033390045,
+      "learning_rate": 4.916101188534452e-05,
+      "loss": 0.9553,
+      "num_input_tokens_seen": 15728640,
+      "step": 1920
+    },
+    {
+      "epoch": 0.249386225610544,
+      "grad_norm": 0.44128185510635376,
+      "learning_rate": 4.9152299522397424e-05,
+      "loss": 1.2107,
+      "num_input_tokens_seen": 15810560,
+      "step": 1930
+    },
+    {
+      "epoch": 0.25067838221992506,
+      "grad_norm": 0.45994389057159424,
+      "learning_rate": 4.91435429366783e-05,
+      "loss": 1.3879,
+      "num_input_tokens_seen": 15892480,
+      "step": 1940
+    },
+    {
+      "epoch": 0.2519705388293061,
+      "grad_norm": 0.46608877182006836,
+      "learning_rate": 4.9134742144220394e-05,
+      "loss": 1.2823,
+      "num_input_tokens_seen": 15974400,
+      "step": 1950
+    },
+    {
+      "epoch": 0.2532626954386872,
+      "grad_norm": 0.45161423087120056,
+      "learning_rate": 4.912589716113794e-05,
+      "loss": 0.9816,
+      "num_input_tokens_seen": 16056320,
+      "step": 1960
+    },
+    {
+      "epoch": 0.2545548520480682,
+      "grad_norm": 0.5524086952209473,
+      "learning_rate": 4.9117008003626066e-05,
+      "loss": 0.8746,
+      "num_input_tokens_seen": 16138240,
+      "step": 1970
+    },
+    {
+      "epoch": 0.2558470086574493,
+      "grad_norm": 0.5476992726325989,
+      "learning_rate": 4.910807468796079e-05,
+      "loss": 0.8788,
+      "num_input_tokens_seen": 16220160,
+      "step": 1980
+    },
+    {
+      "epoch": 0.25713916526683034,
+      "grad_norm": 0.5770102739334106,
+      "learning_rate": 4.9099097230498974e-05,
+      "loss": 0.9458,
+      "num_input_tokens_seen": 16302080,
+      "step": 1990
+    },
+    {
+      "epoch": 0.2584313218762114,
+      "grad_norm": 0.4459473490715027,
+      "learning_rate": 4.909007564767831e-05,
+      "loss": 0.7193,
+      "num_input_tokens_seen": 16384000,
+      "step": 2000
+    },
+    {
+      "epoch": 0.25972347848559246,
+      "grad_norm": 0.43885576725006104,
+      "learning_rate": 4.90810099560173e-05,
+      "loss": 1.1239,
+      "num_input_tokens_seen": 16465920,
+      "step": 2010
+    },
+    {
+      "epoch": 0.2610156350949735,
+      "grad_norm": 0.42829430103302,
+      "learning_rate": 4.907190017211517e-05,
+      "loss": 1.2067,
+      "num_input_tokens_seen": 16547840,
+      "step": 2020
+    },
+    {
+      "epoch": 0.2623077917043546,
+      "grad_norm": 0.47339022159576416,
+      "learning_rate": 4.906274631265191e-05,
+      "loss": 1.0478,
+      "num_input_tokens_seen": 16629760,
+      "step": 2030
+    },
+    {
+      "epoch": 0.2635999483137356,
+      "grad_norm": 0.4274667799472809,
+      "learning_rate": 4.90535483943882e-05,
+      "loss": 1.0864,
+      "num_input_tokens_seen": 16711680,
+      "step": 2040
+    },
+    {
+      "epoch": 0.26489210492311666,
+      "grad_norm": 0.508808434009552,
+      "learning_rate": 4.904430643416541e-05,
+      "loss": 0.7331,
+      "num_input_tokens_seen": 16793600,
+      "step": 2050
+    },
+    {
+      "epoch": 0.26618426153249775,
+      "grad_norm": 0.39238959550857544,
+      "learning_rate": 4.903502044890551e-05,
+      "loss": 0.8874,
+      "num_input_tokens_seen": 16875520,
+      "step": 2060
+    },
+    {
+      "epoch": 0.2674764181418788,
+      "grad_norm": 0.5183860659599304,
+      "learning_rate": 4.902569045561113e-05,
+      "loss": 1.1042,
+      "num_input_tokens_seen": 16957440,
+      "step": 2070
+    },
+    {
+      "epoch": 0.26876857475125987,
+      "grad_norm": 0.45784205198287964,
+      "learning_rate": 4.901631647136543e-05,
+      "loss": 1.0121,
+      "num_input_tokens_seen": 17039360,
+      "step": 2080
+    },
+    {
+      "epoch": 0.2700607313606409,
+      "grad_norm": 0.43935465812683105,
+      "learning_rate": 4.900689851333216e-05,
+      "loss": 1.027,
+      "num_input_tokens_seen": 17121280,
+      "step": 2090
+    },
+    {
+      "epoch": 0.271352887970022,
+      "grad_norm": 0.3319588303565979,
+      "learning_rate": 4.899743659875556e-05,
+      "loss": 1.1295,
+      "num_input_tokens_seen": 17203200,
+      "step": 2100
+    },
+    {
+      "epoch": 0.27264504457940303,
+      "grad_norm": 0.5109902024269104,
+      "learning_rate": 4.8987930744960355e-05,
+      "loss": 1.3633,
+      "num_input_tokens_seen": 17285120,
+      "step": 2110
+    },
+    {
+      "epoch": 0.27393720118878406,
+      "grad_norm": 0.5351001620292664,
+      "learning_rate": 4.897838096935174e-05,
+      "loss": 1.1889,
+      "num_input_tokens_seen": 17367040,
+      "step": 2120
+    },
+    {
+      "epoch": 0.27522935779816515,
+      "grad_norm": 0.4509984850883484,
+      "learning_rate": 4.896878728941531e-05,
+      "loss": 0.927,
+      "num_input_tokens_seen": 17448960,
+      "step": 2130
+    },
+    {
+      "epoch": 0.2765215144075462,
+      "grad_norm": 0.7170317769050598,
+      "learning_rate": 4.8959149722717057e-05,
+      "loss": 0.7892,
+      "num_input_tokens_seen": 17530880,
+      "step": 2140
+    },
+    {
+      "epoch": 0.2778136710169273,
+      "grad_norm": 0.5630227327346802,
+      "learning_rate": 4.894946828690334e-05,
+      "loss": 1.0423,
+      "num_input_tokens_seen": 17612800,
+      "step": 2150
+    },
+    {
+      "epoch": 0.2791058276263083,
+      "grad_norm": 0.6323631405830383,
+      "learning_rate": 4.893974299970082e-05,
+      "loss": 1.2965,
+      "num_input_tokens_seen": 17694720,
+      "step": 2160
+    },
+    {
+      "epoch": 0.28039798423568935,
+      "grad_norm": 0.5983838438987732,
+      "learning_rate": 4.892997387891648e-05,
+      "loss": 0.9905,
+      "num_input_tokens_seen": 17776640,
+      "step": 2170
+    },
+    {
+      "epoch": 0.28169014084507044,
+      "grad_norm": 0.4904590845108032,
+      "learning_rate": 4.892016094243753e-05,
+      "loss": 1.1746,
+      "num_input_tokens_seen": 17858560,
+      "step": 2180
+    },
+    {
+      "epoch": 0.28298229745445147,
+      "grad_norm": 0.39486417174339294,
+      "learning_rate": 4.891030420823142e-05,
+      "loss": 0.818,
+      "num_input_tokens_seen": 17940480,
+      "step": 2190
+    },
+    {
+      "epoch": 0.28427445406383256,
+      "grad_norm": 0.48404547572135925,
+      "learning_rate": 4.89004036943458e-05,
+      "loss": 0.997,
+      "num_input_tokens_seen": 18022400,
+      "step": 2200
+    },
+    {
+      "epoch": 0.2855666106732136,
+      "grad_norm": 0.5132070183753967,
+      "learning_rate": 4.8890459418908476e-05,
+      "loss": 1.3072,
+      "num_input_tokens_seen": 18104320,
+      "step": 2210
+    },
+    {
+      "epoch": 0.28685876728259463,
+      "grad_norm": 0.5053073167800903,
+      "learning_rate": 4.888047140012737e-05,
+      "loss": 0.8763,
+      "num_input_tokens_seen": 18186240,
+      "step": 2220
+    },
+    {
+      "epoch": 0.2881509238919757,
+      "grad_norm": 0.4907422363758087,
+      "learning_rate": 4.8870439656290525e-05,
+      "loss": 0.8944,
+      "num_input_tokens_seen": 18268160,
+      "step": 2230
+    },
+    {
+      "epoch": 0.28944308050135675,
+      "grad_norm": 0.4988420307636261,
+      "learning_rate": 4.8860364205766006e-05,
+      "loss": 0.8613,
+      "num_input_tokens_seen": 18350080,
+      "step": 2240
+    },
+    {
+      "epoch": 0.29073523711073784,
+      "grad_norm": 0.6583264470100403,
+      "learning_rate": 4.885024506700195e-05,
+      "loss": 0.9404,
+      "num_input_tokens_seen": 18432000,
+      "step": 2250
+    },
+    {
+      "epoch": 0.2920273937201189,
+      "grad_norm": 0.4961051344871521,
+      "learning_rate": 4.884008225852644e-05,
+      "loss": 1.1465,
+      "num_input_tokens_seen": 18513920,
+      "step": 2260
+    },
+    {
+      "epoch": 0.2933195503294999,
+      "grad_norm": 0.62845778465271,
+      "learning_rate": 4.8829875798947554e-05,
+      "loss": 1.3089,
+      "num_input_tokens_seen": 18595840,
+      "step": 2270
+    },
+    {
+      "epoch": 0.294611706938881,
+      "grad_norm": 0.557864248752594,
+      "learning_rate": 4.8819625706953286e-05,
+      "loss": 1.0963,
+      "num_input_tokens_seen": 18677760,
+      "step": 2280
+    },
+    {
+      "epoch": 0.29590386354826204,
+      "grad_norm": 0.4075135886669159,
+      "learning_rate": 4.88093320013115e-05,
+      "loss": 0.6548,
+      "num_input_tokens_seen": 18759680,
+      "step": 2290
+    },
+    {
+      "epoch": 0.2971960201576431,
+      "grad_norm": 0.4397824704647064,
+      "learning_rate": 4.879899470086995e-05,
+      "loss": 0.7479,
+      "num_input_tokens_seen": 18841600,
+      "step": 2300
+    },
+    {
+      "epoch": 0.29848817676702416,
+      "grad_norm": 0.2952675521373749,
+      "learning_rate": 4.8788613824556194e-05,
+      "loss": 1.0112,
+      "num_input_tokens_seen": 18923520,
+      "step": 2310
+    },
+    {
+      "epoch": 0.2997803333764052,
+      "grad_norm": 0.46947795152664185,
+      "learning_rate": 4.8778189391377574e-05,
+      "loss": 1.3439,
+      "num_input_tokens_seen": 19005440,
+      "step": 2320
+    },
+    {
+      "epoch": 0.3010724899857863,
+      "grad_norm": 0.5178149342536926,
+      "learning_rate": 4.876772142042117e-05,
+      "loss": 1.0994,
+      "num_input_tokens_seen": 19087360,
+      "step": 2330
+    },
+    {
+      "epoch": 0.3023646465951673,
+      "grad_norm": 0.45673584938049316,
+      "learning_rate": 4.875720993085384e-05,
+      "loss": 1.2049,
+      "num_input_tokens_seen": 19169280,
+      "step": 2340
+    },
+    {
+      "epoch": 0.3036568032045484,
+      "grad_norm": 1.184935212135315,
+      "learning_rate": 4.874665494192206e-05,
+      "loss": 0.6961,
+      "num_input_tokens_seen": 19251200,
+      "step": 2350
+    },
+    {
+      "epoch": 0.30494895981392944,
+      "grad_norm": 0.4711746275424957,
+      "learning_rate": 4.8736056472951955e-05,
+      "loss": 0.8927,
+      "num_input_tokens_seen": 19333120,
+      "step": 2360
+    },
+    {
+      "epoch": 0.3062411164233105,
+      "grad_norm": 0.34465235471725464,
+      "learning_rate": 4.8725414543349326e-05,
+      "loss": 0.9728,
+      "num_input_tokens_seen": 19415040,
+      "step": 2370
+    },
+    {
+      "epoch": 0.30753327303269157,
+      "grad_norm": 0.7332857847213745,
+      "learning_rate": 4.871472917259947e-05,
+      "loss": 1.3036,
+      "num_input_tokens_seen": 19496960,
+      "step": 2380
+    },
+    {
+      "epoch": 0.3088254296420726,
+      "grad_norm": 0.4272123873233795,
+      "learning_rate": 4.870400038026728e-05,
+      "loss": 0.9335,
+      "num_input_tokens_seen": 19578880,
+      "step": 2390
+    },
+    {
+      "epoch": 0.3101175862514537,
+      "grad_norm": 0.47827640175819397,
+      "learning_rate": 4.869322818599714e-05,
+      "loss": 1.1586,
+      "num_input_tokens_seen": 19660800,
+      "step": 2400
+    },
+    {
+      "epoch": 0.3114097428608347,
+      "grad_norm": 0.5260847806930542,
+      "learning_rate": 4.868241260951289e-05,
+      "loss": 1.2544,
+      "num_input_tokens_seen": 19742720,
+      "step": 2410
+    },
+    {
+      "epoch": 0.3127018994702158,
+      "grad_norm": 0.5642371773719788,
+      "learning_rate": 4.867155367061781e-05,
+      "loss": 1.1045,
+      "num_input_tokens_seen": 19824640,
+      "step": 2420
+    },
+    {
+      "epoch": 0.31399405607959685,
+      "grad_norm": 0.2645663321018219,
+      "learning_rate": 4.8660651389194576e-05,
+      "loss": 0.5473,
+      "num_input_tokens_seen": 19906560,
+      "step": 2430
+    },
+    {
+      "epoch": 0.3152862126889779,
+      "grad_norm": 0.47711700201034546,
+      "learning_rate": 4.8649705785205224e-05,
+      "loss": 1.2962,
+      "num_input_tokens_seen": 19988480,
+      "step": 2440
+    },
+    {
+      "epoch": 0.316578369298359,
+      "grad_norm": 0.46864113211631775,
+      "learning_rate": 4.8638716878691125e-05,
+      "loss": 0.8808,
+      "num_input_tokens_seen": 20070400,
+      "step": 2450
+    },
+    {
+      "epoch": 0.31787052590774,
+      "grad_norm": 0.45676669478416443,
+      "learning_rate": 4.862768468977293e-05,
+      "loss": 1.2679,
+      "num_input_tokens_seen": 20152320,
+      "step": 2460
+    },
+    {
+      "epoch": 0.3191626825171211,
+      "grad_norm": 1.7806425094604492,
+      "learning_rate": 4.861660923865052e-05,
+      "loss": 0.9461,
+      "num_input_tokens_seen": 20234240,
+      "step": 2470
+    },
+    {
+      "epoch": 0.32045483912650213,
+      "grad_norm": 0.491797536611557,
+      "learning_rate": 4.860549054560301e-05,
+      "loss": 0.7236,
+      "num_input_tokens_seen": 20316160,
+      "step": 2480
+    },
+    {
+      "epoch": 0.32174699573588317,
+      "grad_norm": 0.46949175000190735,
+      "learning_rate": 4.8594328630988696e-05,
+      "loss": 1.2098,
+      "num_input_tokens_seen": 20398080,
+      "step": 2490
+    },
+    {
+      "epoch": 0.32303915234526426,
+      "grad_norm": 0.4671134948730469,
+      "learning_rate": 4.858312351524499e-05,
+      "loss": 1.2859,
+      "num_input_tokens_seen": 20480000,
+      "step": 2500
+    },
+    {
+      "epoch": 0.3243313089546453,
+      "grad_norm": 0.6649695634841919,
+      "learning_rate": 4.857187521888843e-05,
+      "loss": 0.9768,
+      "num_input_tokens_seen": 20561920,
+      "step": 2510
+    },
+    {
+      "epoch": 0.3256234655640264,
+      "grad_norm": 0.42521458864212036,
+      "learning_rate": 4.8560583762514594e-05,
+      "loss": 0.8317,
+      "num_input_tokens_seen": 20643840,
+      "step": 2520
+    },
+    {
+      "epoch": 0.3269156221734074,
+      "grad_norm": 0.6050983667373657,
+      "learning_rate": 4.854924916679811e-05,
+      "loss": 0.6581,
+      "num_input_tokens_seen": 20725760,
+      "step": 2530
+    },
+    {
+      "epoch": 0.32820777878278845,
+      "grad_norm": 0.6202778220176697,
+      "learning_rate": 4.8537871452492565e-05,
+      "loss": 1.248,
+      "num_input_tokens_seen": 20807680,
+      "step": 2540
+    },
+    {
+      "epoch": 0.32949993539216954,
+      "grad_norm": 0.4647495746612549,
+      "learning_rate": 4.852645064043053e-05,
+      "loss": 0.9416,
+      "num_input_tokens_seen": 20889600,
+      "step": 2550
+    },
+    {
+      "epoch": 0.3307920920015506,
+      "grad_norm": 0.5066781044006348,
+      "learning_rate": 4.851498675152346e-05,
+      "loss": 1.0115,
+      "num_input_tokens_seen": 20971520,
+      "step": 2560
+    },
+    {
+      "epoch": 0.33208424861093166,
+      "grad_norm": 0.252456933259964,
+      "learning_rate": 4.8503479806761684e-05,
+      "loss": 0.8474,
+      "num_input_tokens_seen": 21053440,
+      "step": 2570
+    },
+    {
+      "epoch": 0.3333764052203127,
+      "grad_norm": 0.4478939175605774,
+      "learning_rate": 4.84919298272144e-05,
+      "loss": 1.0834,
+      "num_input_tokens_seen": 21135360,
+      "step": 2580
+    },
+    {
+      "epoch": 0.33466856182969373,
+      "grad_norm": 0.6947191953659058,
+      "learning_rate": 4.848033683402956e-05,
+      "loss": 1.3049,
+      "num_input_tokens_seen": 21217280,
+      "step": 2590
+    },
+    {
+      "epoch": 0.3359607184390748,
+      "grad_norm": 0.7486204504966736,
+      "learning_rate": 4.84687008484339e-05,
+      "loss": 1.0012,
+      "num_input_tokens_seen": 21299200,
+      "step": 2600
+    },
+    {
+      "epoch": 0.33725287504845586,
+      "grad_norm": 0.2647557854652405,
+      "learning_rate": 4.8457021891732866e-05,
+      "loss": 0.6511,
+      "num_input_tokens_seen": 21381120,
+      "step": 2610
+    },
+    {
+      "epoch": 0.33854503165783695,
+      "grad_norm": 0.7009825110435486,
+      "learning_rate": 4.844529998531058e-05,
+      "loss": 0.715,
+      "num_input_tokens_seen": 21463040,
+      "step": 2620
+    },
+    {
+      "epoch": 0.339837188267218,
+      "grad_norm": 0.5106743574142456,
+      "learning_rate": 4.843353515062982e-05,
+      "loss": 0.9274,
+      "num_input_tokens_seen": 21544960,
+      "step": 2630
+    },
+    {
+      "epoch": 0.341129344876599,
+      "grad_norm": 0.26822468638420105,
+      "learning_rate": 4.842172740923194e-05,
+      "loss": 0.6468,
+      "num_input_tokens_seen": 21626880,
+      "step": 2640
+    },
+    {
+      "epoch": 0.3424215014859801,
+      "grad_norm": 0.7012847065925598,
+      "learning_rate": 4.840987678273688e-05,
+      "loss": 0.9614,
+      "num_input_tokens_seen": 21708800,
+      "step": 2650
+    },
+    {
+      "epoch": 0.34371365809536114,
+      "grad_norm": 0.4316296875476837,
+      "learning_rate": 4.8397983292843095e-05,
+      "loss": 0.8697,
+      "num_input_tokens_seen": 21790720,
+      "step": 2660
+    },
+    {
+      "epoch": 0.34500581470474223,
+      "grad_norm": 0.3868614137172699,
+      "learning_rate": 4.838604696132753e-05,
+      "loss": 0.7511,
+      "num_input_tokens_seen": 21872640,
+      "step": 2670
+    },
+    {
+      "epoch": 0.34629797131412327,
+      "grad_norm": 0.3558228313922882,
+      "learning_rate": 4.837406781004554e-05,
+      "loss": 1.081,
+      "num_input_tokens_seen": 21954560,
+      "step": 2680
+    },
+    {
+      "epoch": 0.34759012792350436,
+      "grad_norm": 0.47105276584625244,
+      "learning_rate": 4.836204586093092e-05,
+      "loss": 0.9772,
+      "num_input_tokens_seen": 22036480,
+      "step": 2690
+    },
+    {
+      "epoch": 0.3488822845328854,
+      "grad_norm": 0.3401098847389221,
+      "learning_rate": 4.8349981135995826e-05,
+      "loss": 0.7102,
+      "num_input_tokens_seen": 22118400,
+      "step": 2700
+    },
+    {
+      "epoch": 0.3501744411422664,
+      "grad_norm": 0.4399671256542206,
+      "learning_rate": 4.833787365733071e-05,
+      "loss": 0.8293,
+      "num_input_tokens_seen": 22200320,
+      "step": 2710
+    },
+    {
+      "epoch": 0.3514665977516475,
+      "grad_norm": 0.27914759516716003,
+      "learning_rate": 4.832572344710433e-05,
+      "loss": 0.6742,
+      "num_input_tokens_seen": 22282240,
+      "step": 2720
+    },
+    {
+      "epoch": 0.35275875436102855,
+      "grad_norm": 0.6157001256942749,
+      "learning_rate": 4.831353052756367e-05,
+      "loss": 1.1822,
+      "num_input_tokens_seen": 22364160,
+      "step": 2730
+    },
+    {
+      "epoch": 0.35405091097040964,
+      "grad_norm": 0.45960766077041626,
+      "learning_rate": 4.830129492103392e-05,
+      "loss": 1.0566,
+      "num_input_tokens_seen": 22446080,
+      "step": 2740
+    },
+    {
+      "epoch": 0.3553430675797907,
+      "grad_norm": 0.35881564021110535,
+      "learning_rate": 4.828901664991845e-05,
+      "loss": 0.8509,
+      "num_input_tokens_seen": 22528000,
+      "step": 2750
+    },
+    {
+      "epoch": 0.3566352241891717,
+      "grad_norm": 0.5361536145210266,
+      "learning_rate": 4.8276695736698704e-05,
+      "loss": 0.9067,
+      "num_input_tokens_seen": 22609920,
+      "step": 2760
+    },
+    {
+      "epoch": 0.3579273807985528,
+      "grad_norm": 0.39405712485313416,
+      "learning_rate": 4.826433220393424e-05,
+      "loss": 0.7381,
+      "num_input_tokens_seen": 22691840,
+      "step": 2770
+    },
+    {
+      "epoch": 0.35921953740793383,
+      "grad_norm": 0.42842912673950195,
+      "learning_rate": 4.825192607426264e-05,
+      "loss": 0.9375,
+      "num_input_tokens_seen": 22773760,
+      "step": 2780
+    },
+    {
+      "epoch": 0.3605116940173149,
+      "grad_norm": 0.5455546975135803,
+      "learning_rate": 4.823947737039948e-05,
+      "loss": 1.1993,
+      "num_input_tokens_seen": 22855680,
+      "step": 2790
+    },
+    {
+      "epoch": 0.36180385062669596,
+      "grad_norm": 0.6831130981445312,
+      "learning_rate": 4.82269861151383e-05,
+      "loss": 0.7907,
+      "num_input_tokens_seen": 22937600,
+      "step": 2800
+    },
+    {
+      "epoch": 0.363096007236077,
+      "grad_norm": 0.5363448858261108,
+      "learning_rate": 4.821445233135053e-05,
+      "loss": 1.0263,
+      "num_input_tokens_seen": 23019520,
+      "step": 2810
+    },
+    {
+      "epoch": 0.3643881638454581,
+      "grad_norm": 0.48520973324775696,
+      "learning_rate": 4.8201876041985496e-05,
+      "loss": 0.8387,
+      "num_input_tokens_seen": 23101440,
+      "step": 2820
+    },
+    {
+      "epoch": 0.3656803204548391,
+      "grad_norm": 0.48314452171325684,
+      "learning_rate": 4.8189257270070335e-05,
+      "loss": 1.1519,
+      "num_input_tokens_seen": 23183360,
+      "step": 2830
+    },
+    {
+      "epoch": 0.3669724770642202,
+      "grad_norm": 0.8397269248962402,
+      "learning_rate": 4.817659603870995e-05,
+      "loss": 0.8801,
+      "num_input_tokens_seen": 23265280,
+      "step": 2840
+    },
+    {
+      "epoch": 0.36826463367360124,
+      "grad_norm": 0.5130114555358887,
+      "learning_rate": 4.8163892371087045e-05,
+      "loss": 0.8017,
+      "num_input_tokens_seen": 23347200,
+      "step": 2850
+    },
+    {
+      "epoch": 0.3695567902829823,
+      "grad_norm": 0.653531551361084,
+      "learning_rate": 4.815114629046196e-05,
+      "loss": 0.9614,
+      "num_input_tokens_seen": 23429120,
+      "step": 2860
+    },
+    {
+      "epoch": 0.37084894689236336,
+      "grad_norm": 0.39914658665657043,
+      "learning_rate": 4.813835782017274e-05,
+      "loss": 0.93,
+      "num_input_tokens_seen": 23511040,
+      "step": 2870
+    },
+    {
+      "epoch": 0.3721411035017444,
+      "grad_norm": 0.6135099530220032,
+      "learning_rate": 4.812552698363502e-05,
+      "loss": 0.9083,
+      "num_input_tokens_seen": 23592960,
+      "step": 2880
+    },
+    {
+      "epoch": 0.3734332601111255,
+      "grad_norm": 0.3109150826931,
+      "learning_rate": 4.8112653804342015e-05,
+      "loss": 0.486,
+      "num_input_tokens_seen": 23674880,
+      "step": 2890
+    },
+    {
+      "epoch": 0.3747254167205065,
+      "grad_norm": 0.4596582353115082,
+      "learning_rate": 4.809973830586446e-05,
+      "loss": 1.176,
+      "num_input_tokens_seen": 23756800,
+      "step": 2900
+    },
+    {
+      "epoch": 0.37601757332988756,
+      "grad_norm": 0.4620276391506195,
+      "learning_rate": 4.8086780511850606e-05,
+      "loss": 1.007,
+      "num_input_tokens_seen": 23838720,
+      "step": 2910
+    },
+    {
+      "epoch": 0.37730972993926865,
+      "grad_norm": 0.5108826160430908,
+      "learning_rate": 4.807378044602611e-05,
+      "loss": 1.0682,
+      "num_input_tokens_seen": 23920640,
+      "step": 2920
+    },
+    {
+      "epoch": 0.3786018865486497,
+      "grad_norm": 0.48416051268577576,
+      "learning_rate": 4.806073813219404e-05,
+      "loss": 0.9145,
+      "num_input_tokens_seen": 24002560,
+      "step": 2930
+    },
+    {
+      "epoch": 0.37989404315803077,
+      "grad_norm": 0.4877118766307831,
+      "learning_rate": 4.8047653594234855e-05,
+      "loss": 0.6768,
+      "num_input_tokens_seen": 24084480,
+      "step": 2940
+    },
+    {
+      "epoch": 0.3811861997674118,
+      "grad_norm": 0.7045567631721497,
+      "learning_rate": 4.803452685610626e-05,
+      "loss": 1.3005,
+      "num_input_tokens_seen": 24166400,
+      "step": 2950
+    },
+    {
+      "epoch": 0.3824783563767929,
+      "grad_norm": 0.6394911408424377,
+      "learning_rate": 4.802135794184329e-05,
+      "loss": 0.9298,
+      "num_input_tokens_seen": 24248320,
+      "step": 2960
+    },
+    {
+      "epoch": 0.38377051298617393,
+      "grad_norm": 0.5416907072067261,
+      "learning_rate": 4.800814687555817e-05,
+      "loss": 0.7336,
+      "num_input_tokens_seen": 24330240,
+      "step": 2970
+    },
+    {
+      "epoch": 0.38506266959555496,
+      "grad_norm": 0.47221288084983826,
+      "learning_rate": 4.799489368144031e-05,
+      "loss": 1.221,
+      "num_input_tokens_seen": 24412160,
+      "step": 2980
+    },
+    {
+      "epoch": 0.38635482620493605,
+      "grad_norm": 0.49521034955978394,
+      "learning_rate": 4.798159838375626e-05,
+      "loss": 0.9109,
+      "num_input_tokens_seen": 24494080,
+      "step": 2990
+    },
+    {
+      "epoch": 0.3876469828143171,
+      "grad_norm": 0.4463275671005249,
+      "learning_rate": 4.796826100684967e-05,
+      "loss": 0.6901,
+      "num_input_tokens_seen": 24576000,
+      "step": 3000
+    },
+    {
+      "epoch": 0.3889391394236982,
+      "grad_norm": 0.46501803398132324,
+      "learning_rate": 4.795488157514122e-05,
+      "loss": 1.0645,
+      "num_input_tokens_seen": 24657920,
+      "step": 3010
+    },
+    {
+      "epoch": 0.3902312960330792,
+      "grad_norm": 0.40190380811691284,
+      "learning_rate": 4.794146011312861e-05,
+      "loss": 1.0953,
+      "num_input_tokens_seen": 24739840,
+      "step": 3020
+    },
+    {
+      "epoch": 0.39152345264246025,
+      "grad_norm": 0.5038986206054688,
+      "learning_rate": 4.7927996645386476e-05,
+      "loss": 1.2322,
+      "num_input_tokens_seen": 24821760,
+      "step": 3030
+    },
+    {
+      "epoch": 0.39281560925184134,
+      "grad_norm": 0.4735121428966522,
+      "learning_rate": 4.791449119656638e-05,
+      "loss": 0.7708,
+      "num_input_tokens_seen": 24903680,
+      "step": 3040
+    },
+    {
+      "epoch": 0.39410776586122237,
+      "grad_norm": 0.7169963121414185,
+      "learning_rate": 4.790094379139676e-05,
+      "loss": 0.8159,
+      "num_input_tokens_seen": 24985600,
+      "step": 3050
+    },
+    {
+      "epoch": 0.39539992247060346,
+      "grad_norm": 0.9569246172904968,
+      "learning_rate": 4.7887354454682854e-05,
+      "loss": 0.7697,
+      "num_input_tokens_seen": 25067520,
+      "step": 3060
+    },
+    {
+      "epoch": 0.3966920790799845,
+      "grad_norm": 0.5772688388824463,
+      "learning_rate": 4.78737232113067e-05,
+      "loss": 1.0574,
+      "num_input_tokens_seen": 25149440,
+      "step": 3070
+    },
+    {
+      "epoch": 0.39798423568936553,
+      "grad_norm": 0.4956931173801422,
+      "learning_rate": 4.7860050086227035e-05,
+      "loss": 1.0678,
+      "num_input_tokens_seen": 25231360,
+      "step": 3080
+    },
+    {
+      "epoch": 0.3992763922987466,
+      "grad_norm": 0.39392775297164917,
+      "learning_rate": 4.784633510447932e-05,
+      "loss": 0.9143,
+      "num_input_tokens_seen": 25313280,
+      "step": 3090
+    },
+    {
+      "epoch": 0.40056854890812765,
+      "grad_norm": 0.40444672107696533,
+      "learning_rate": 4.7832578291175626e-05,
+      "loss": 0.8812,
+      "num_input_tokens_seen": 25395200,
+      "step": 3100
+    },
+    {
+      "epoch": 0.40186070551750874,
+      "grad_norm": 0.7210461497306824,
+      "learning_rate": 4.781877967150463e-05,
+      "loss": 0.6671,
+      "num_input_tokens_seen": 25477120,
+      "step": 3110
+    },
+    {
+      "epoch": 0.4031528621268898,
+      "grad_norm": 0.6195704340934753,
+      "learning_rate": 4.7804939270731564e-05,
+      "loss": 1.0019,
+      "num_input_tokens_seen": 25559040,
+      "step": 3120
+    },
+    {
+      "epoch": 0.4044450187362708,
+      "grad_norm": 0.42882412672042847,
+      "learning_rate": 4.7791057114198133e-05,
+      "loss": 0.799,
+      "num_input_tokens_seen": 25640960,
+      "step": 3130
+    },
+    {
+      "epoch": 0.4057371753456519,
+      "grad_norm": 0.3896157741546631,
+      "learning_rate": 4.7777133227322525e-05,
+      "loss": 1.0606,
+      "num_input_tokens_seen": 25722880,
+      "step": 3140
+    },
+    {
+      "epoch": 0.40702933195503294,
+      "grad_norm": 0.3924260139465332,
+      "learning_rate": 4.776316763559933e-05,
+      "loss": 0.8224,
+      "num_input_tokens_seen": 25804800,
+      "step": 3150
+    },
+    {
+      "epoch": 0.408321488564414,
+      "grad_norm": 0.44281336665153503,
+      "learning_rate": 4.774916036459949e-05,
+      "loss": 0.8995,
+      "num_input_tokens_seen": 25886720,
+      "step": 3160
+    },
+    {
+      "epoch": 0.40961364517379506,
+      "grad_norm": 0.5028753280639648,
+      "learning_rate": 4.773511143997026e-05,
+      "loss": 1.1617,
+      "num_input_tokens_seen": 25968640,
+      "step": 3170
+    },
+    {
+      "epoch": 0.4109058017831761,
+      "grad_norm": 0.6678476929664612,
+      "learning_rate": 4.7721020887435186e-05,
+      "loss": 0.8537,
+      "num_input_tokens_seen": 26050560,
+      "step": 3180
+    },
+    {
+      "epoch": 0.4121979583925572,
+      "grad_norm": 0.6059293746948242,
+      "learning_rate": 4.7706888732793996e-05,
+      "loss": 1.5184,
+      "num_input_tokens_seen": 26132480,
+      "step": 3190
+    },
+    {
+      "epoch": 0.4134901150019382,
+      "grad_norm": 0.46364808082580566,
+      "learning_rate": 4.769271500192264e-05,
+      "loss": 1.1,
+      "num_input_tokens_seen": 26214400,
+      "step": 3200
+    },
+    {
+      "epoch": 0.4147822716113193,
+      "grad_norm": 0.6045968532562256,
+      "learning_rate": 4.767849972077315e-05,
+      "loss": 1.0147,
+      "num_input_tokens_seen": 26296320,
+      "step": 3210
+    },
+    {
+      "epoch": 0.41607442822070034,
+      "grad_norm": 0.6123097538948059,
+      "learning_rate": 4.766424291537366e-05,
+      "loss": 1.0684,
+      "num_input_tokens_seen": 26378240,
+      "step": 3220
+    },
+    {
+      "epoch": 0.41736658483008143,
+      "grad_norm": 0.3833664059638977,
+      "learning_rate": 4.7649944611828316e-05,
+      "loss": 0.7629,
+      "num_input_tokens_seen": 26460160,
+      "step": 3230
+    },
+    {
+      "epoch": 0.41865874143946247,
+      "grad_norm": 0.5847949981689453,
+      "learning_rate": 4.763560483631728e-05,
+      "loss": 0.8127,
+      "num_input_tokens_seen": 26542080,
+      "step": 3240
+    },
+    {
+      "epoch": 0.4199508980488435,
+      "grad_norm": 0.5139411687850952,
+      "learning_rate": 4.762122361509662e-05,
+      "loss": 1.0787,
+      "num_input_tokens_seen": 26624000,
+      "step": 3250
+    },
+    {
+      "epoch": 0.4212430546582246,
+      "grad_norm": 0.43457046151161194,
+      "learning_rate": 4.7606800974498287e-05,
+      "loss": 1.089,
+      "num_input_tokens_seen": 26705920,
+      "step": 3260
+    },
+    {
+      "epoch": 0.4225352112676056,
+      "grad_norm": 0.5299356579780579,
+      "learning_rate": 4.75923369409301e-05,
+      "loss": 0.7698,
+      "num_input_tokens_seen": 26787840,
+      "step": 3270
+    },
+    {
+      "epoch": 0.4238273678769867,
+      "grad_norm": 0.5959427952766418,
+      "learning_rate": 4.757783154087564e-05,
+      "loss": 1.2242,
+      "num_input_tokens_seen": 26869760,
+      "step": 3280
+    },
+    {
+      "epoch": 0.42511952448636775,
+      "grad_norm": 0.5723779201507568,
+      "learning_rate": 4.756328480089425e-05,
+      "loss": 1.2536,
+      "num_input_tokens_seen": 26951680,
+      "step": 3290
+    },
+    {
+      "epoch": 0.4264116810957488,
+      "grad_norm": 0.4997044503688812,
+      "learning_rate": 4.7548696747620956e-05,
+      "loss": 1.0486,
+      "num_input_tokens_seen": 27033600,
+      "step": 3300
+    },
+    {
+      "epoch": 0.4277038377051299,
+      "grad_norm": 0.4314401149749756,
+      "learning_rate": 4.753406740776643e-05,
+      "loss": 1.1756,
+      "num_input_tokens_seen": 27115520,
+      "step": 3310
+    },
+    {
+      "epoch": 0.4289959943145109,
+      "grad_norm": 0.5475839376449585,
+      "learning_rate": 4.7519396808116933e-05,
+      "loss": 1.2511,
+      "num_input_tokens_seen": 27197440,
+      "step": 3320
+    },
+    {
+      "epoch": 0.430288150923892,
+      "grad_norm": 0.4945213794708252,
+      "learning_rate": 4.750468497553429e-05,
+      "loss": 0.8987,
+      "num_input_tokens_seen": 27279360,
+      "step": 3330
+    },
+    {
+      "epoch": 0.43158030753327303,
+      "grad_norm": 0.4221036434173584,
+      "learning_rate": 4.74899319369558e-05,
+      "loss": 1.0922,
+      "num_input_tokens_seen": 27361280,
+      "step": 3340
+    },
+    {
+      "epoch": 0.43287246414265407,
+      "grad_norm": 0.40295591950416565,
+      "learning_rate": 4.7475137719394234e-05,
+      "loss": 1.028,
+      "num_input_tokens_seen": 27443200,
+      "step": 3350
+    },
+    {
+      "epoch": 0.43416462075203516,
+      "grad_norm": 0.42078787088394165,
+      "learning_rate": 4.746030234993775e-05,
+      "loss": 0.7594,
+      "num_input_tokens_seen": 27525120,
+      "step": 3360
+    },
+    {
+      "epoch": 0.4354567773614162,
+      "grad_norm": 0.5644673109054565,
+      "learning_rate": 4.7445425855749844e-05,
+      "loss": 1.21,
+      "num_input_tokens_seen": 27607040,
+      "step": 3370
+    },
+    {
+      "epoch": 0.4367489339707973,
+      "grad_norm": 0.9208412766456604,
+      "learning_rate": 4.743050826406934e-05,
+      "loss": 0.8709,
+      "num_input_tokens_seen": 27688960,
+      "step": 3380
+    },
+    {
+      "epoch": 0.4380410905801783,
+      "grad_norm": 0.43149518966674805,
+      "learning_rate": 4.741554960221027e-05,
+      "loss": 0.8737,
+      "num_input_tokens_seen": 27770880,
+      "step": 3390
+    },
+    {
+      "epoch": 0.43933324718955935,
+      "grad_norm": 0.45487073063850403,
+      "learning_rate": 4.7400549897561914e-05,
+      "loss": 1.2981,
+      "num_input_tokens_seen": 27852800,
+      "step": 3400
+    },
+    {
+      "epoch": 0.44062540379894044,
+      "grad_norm": 0.587581217288971,
+      "learning_rate": 4.7385509177588664e-05,
+      "loss": 0.7498,
+      "num_input_tokens_seen": 27934720,
+      "step": 3410
+    },
+    {
+      "epoch": 0.4419175604083215,
+      "grad_norm": 15.474299430847168,
+      "learning_rate": 4.7370427469830016e-05,
+      "loss": 1.5585,
+      "num_input_tokens_seen": 28016640,
+      "step": 3420
+    },
+    {
+      "epoch": 0.44320971701770256,
+      "grad_norm": 0.41457799077033997,
+      "learning_rate": 4.735530480190053e-05,
+      "loss": 1.2198,
+      "num_input_tokens_seen": 28098560,
+      "step": 3430
+    },
+    {
+      "epoch": 0.4445018736270836,
+      "grad_norm": 0.41844090819358826,
+      "learning_rate": 4.734014120148976e-05,
+      "loss": 0.7208,
+      "num_input_tokens_seen": 28180480,
+      "step": 3440
+    },
+    {
+      "epoch": 0.44579403023646463,
+      "grad_norm": 0.3628556728363037,
+      "learning_rate": 4.73249366963622e-05,
+      "loss": 0.8981,
+      "num_input_tokens_seen": 28262400,
+      "step": 3450
+    },
+    {
+      "epoch": 0.4470861868458457,
+      "grad_norm": 0.6861358284950256,
+      "learning_rate": 4.730969131435724e-05,
+      "loss": 0.9347,
+      "num_input_tokens_seen": 28344320,
+      "step": 3460
+    },
+    {
+      "epoch": 0.44837834345522676,
+      "grad_norm": 0.48623812198638916,
+      "learning_rate": 4.729440508338911e-05,
+      "loss": 1.1323,
+      "num_input_tokens_seen": 28426240,
+      "step": 3470
+    },
+    {
+      "epoch": 0.44967050006460785,
+      "grad_norm": 0.41574108600616455,
+      "learning_rate": 4.727907803144686e-05,
+      "loss": 1.1095,
+      "num_input_tokens_seen": 28508160,
+      "step": 3480
+    },
+    {
+      "epoch": 0.4509626566739889,
+      "grad_norm": 0.5009231567382812,
+      "learning_rate": 4.726371018659427e-05,
+      "loss": 1.2198,
+      "num_input_tokens_seen": 28590080,
+      "step": 3490
+    },
+    {
+      "epoch": 0.45225481328336997,
+      "grad_norm": 0.48607850074768066,
+      "learning_rate": 4.724830157696979e-05,
+      "loss": 1.078,
+      "num_input_tokens_seen": 28672000,
+      "step": 3500
+    },
+    {
+      "epoch": 0.453546969892751,
+      "grad_norm": 0.8269723653793335,
+      "learning_rate": 4.723285223078653e-05,
+      "loss": 0.9332,
+      "num_input_tokens_seen": 28753920,
+      "step": 3510
+    },
+    {
+      "epoch": 0.45483912650213204,
+      "grad_norm": 0.5304045677185059,
+      "learning_rate": 4.721736217633219e-05,
+      "loss": 1.3173,
+      "num_input_tokens_seen": 28835840,
+      "step": 3520
+    },
+    {
+      "epoch": 0.45613128311151313,
+      "grad_norm": 0.4949178695678711,
+      "learning_rate": 4.7201831441969016e-05,
+      "loss": 0.9488,
+      "num_input_tokens_seen": 28917760,
+      "step": 3530
+    },
+    {
+      "epoch": 0.45742343972089416,
+      "grad_norm": 0.742077112197876,
+      "learning_rate": 4.71862600561337e-05,
+      "loss": 0.8988,
+      "num_input_tokens_seen": 28999680,
+      "step": 3540
+    },
+    {
+      "epoch": 0.45871559633027525,
+      "grad_norm": 0.6357890963554382,
+      "learning_rate": 4.7170648047337415e-05,
+      "loss": 1.401,
+      "num_input_tokens_seen": 29081600,
+      "step": 3550
+    },
+    {
+      "epoch": 0.4600077529396563,
+      "grad_norm": 0.512546956539154,
+      "learning_rate": 4.7154995444165685e-05,
+      "loss": 0.768,
+      "num_input_tokens_seen": 29163520,
+      "step": 3560
+    },
+    {
+      "epoch": 0.4612999095490373,
+      "grad_norm": 0.4622116684913635,
+      "learning_rate": 4.713930227527836e-05,
+      "loss": 0.8524,
+      "num_input_tokens_seen": 29245440,
+      "step": 3570
+    },
+    {
+      "epoch": 0.4625920661584184,
+      "grad_norm": 0.620832085609436,
+      "learning_rate": 4.712356856940958e-05,
+      "loss": 0.8993,
+      "num_input_tokens_seen": 29327360,
+      "step": 3580
+    },
+    {
+      "epoch": 0.46388422276779945,
+      "grad_norm": 0.4882233142852783,
+      "learning_rate": 4.710779435536772e-05,
+      "loss": 0.7759,
+      "num_input_tokens_seen": 29409280,
+      "step": 3590
+    },
+    {
+      "epoch": 0.46517637937718054,
+      "grad_norm": 0.39704984426498413,
+      "learning_rate": 4.709197966203528e-05,
+      "loss": 0.7109,
+      "num_input_tokens_seen": 29491200,
+      "step": 3600
+    },
+    {
+      "epoch": 0.46646853598656157,
+      "grad_norm": 0.6212629079818726,
+      "learning_rate": 4.707612451836892e-05,
+      "loss": 1.2732,
+      "num_input_tokens_seen": 29573120,
+      "step": 3610
+    },
+    {
+      "epoch": 0.4677606925959426,
+      "grad_norm": 0.4958352744579315,
+      "learning_rate": 4.706022895339936e-05,
+      "loss": 1.0464,
+      "num_input_tokens_seen": 29655040,
+      "step": 3620
+    },
+    {
+      "epoch": 0.4690528492053237,
+      "grad_norm": 0.5100958943367004,
+      "learning_rate": 4.704429299623129e-05,
+      "loss": 0.8741,
+      "num_input_tokens_seen": 29736960,
+      "step": 3630
+    },
+    {
+      "epoch": 0.47034500581470473,
+      "grad_norm": 0.4343372583389282,
+      "learning_rate": 4.7028316676043425e-05,
+      "loss": 0.9055,
+      "num_input_tokens_seen": 29818880,
+      "step": 3640
+    },
+    {
+      "epoch": 0.4716371624240858,
+      "grad_norm": 0.5606836676597595,
+      "learning_rate": 4.7012300022088326e-05,
+      "loss": 0.9934,
+      "num_input_tokens_seen": 29900800,
+      "step": 3650
+    },
+    {
+      "epoch": 0.47292931903346686,
+      "grad_norm": 0.5697008371353149,
+      "learning_rate": 4.6996243063692446e-05,
+      "loss": 1.0764,
+      "num_input_tokens_seen": 29982720,
+      "step": 3660
+    },
+    {
+      "epoch": 0.4742214756428479,
+      "grad_norm": 0.9284478425979614,
+      "learning_rate": 4.6980145830255993e-05,
+      "loss": 0.5566,
+      "num_input_tokens_seen": 30064640,
+      "step": 3670
+    },
+    {
+      "epoch": 0.475513632252229,
+      "grad_norm": 0.48587319254875183,
+      "learning_rate": 4.6964008351252964e-05,
+      "loss": 0.9627,
+      "num_input_tokens_seen": 30146560,
+      "step": 3680
+    },
+    {
+      "epoch": 0.47680578886161,
+      "grad_norm": 0.6687888503074646,
+      "learning_rate": 4.694783065623102e-05,
+      "loss": 0.8859,
+      "num_input_tokens_seen": 30228480,
+      "step": 3690
+    },
+    {
+      "epoch": 0.4780979454709911,
+      "grad_norm": 0.6682938933372498,
+      "learning_rate": 4.6931612774811445e-05,
+      "loss": 1.1395,
+      "num_input_tokens_seen": 30310400,
+      "step": 3700
+    },
+    {
+      "epoch": 0.47939010208037214,
+      "grad_norm": 0.5708907246589661,
+      "learning_rate": 4.691535473668914e-05,
+      "loss": 0.7113,
+      "num_input_tokens_seen": 30392320,
+      "step": 3710
+    },
+    {
+      "epoch": 0.4806822586897532,
+      "grad_norm": 0.5752365589141846,
+      "learning_rate": 4.68990565716325e-05,
+      "loss": 1.049,
+      "num_input_tokens_seen": 30474240,
+      "step": 3720
+    },
+    {
+      "epoch": 0.48197441529913426,
+      "grad_norm": 0.5052396059036255,
+      "learning_rate": 4.688271830948342e-05,
+      "loss": 1.1382,
+      "num_input_tokens_seen": 30556160,
+      "step": 3730
+    },
+    {
+      "epoch": 0.4832665719085153,
+      "grad_norm": 0.587308406829834,
+      "learning_rate": 4.686633998015718e-05,
+      "loss": 0.8307,
+      "num_input_tokens_seen": 30638080,
+      "step": 3740
+    },
+    {
+      "epoch": 0.4845587285178964,
+      "grad_norm": 0.6255913972854614,
+      "learning_rate": 4.6849921613642456e-05,
+      "loss": 1.0814,
+      "num_input_tokens_seen": 30720000,
+      "step": 3750
+    },
+    {
+      "epoch": 0.4858508851272774,
+      "grad_norm": 0.5788549184799194,
+      "learning_rate": 4.683346324000122e-05,
+      "loss": 1.0745,
+      "num_input_tokens_seen": 30801920,
+      "step": 3760
+    },
+    {
+      "epoch": 0.48714304173665846,
+      "grad_norm": 0.45166078209877014,
+      "learning_rate": 4.6816964889368674e-05,
+      "loss": 1.0342,
+      "num_input_tokens_seen": 30883840,
+      "step": 3770
+    },
+    {
+      "epoch": 0.48843519834603955,
+      "grad_norm": 0.38736027479171753,
+      "learning_rate": 4.680042659195325e-05,
+      "loss": 1.5249,
+      "num_input_tokens_seen": 30965760,
+      "step": 3780
+    },
+    {
+      "epoch": 0.4897273549554206,
+      "grad_norm": 0.5221673846244812,
+      "learning_rate": 4.678384837803651e-05,
+      "loss": 0.9,
+      "num_input_tokens_seen": 31047680,
+      "step": 3790
+    },
+    {
+      "epoch": 0.49101951156480167,
+      "grad_norm": 0.5739164352416992,
+      "learning_rate": 4.67672302779731e-05,
+      "loss": 0.835,
+      "num_input_tokens_seen": 31129600,
+      "step": 3800
+    },
+    {
+      "epoch": 0.4923116681741827,
+      "grad_norm": 0.6319538950920105,
+      "learning_rate": 4.6750572322190716e-05,
+      "loss": 1.2393,
+      "num_input_tokens_seen": 31211520,
+      "step": 3810
+    },
+    {
+      "epoch": 0.4936038247835638,
+      "grad_norm": 0.8037749528884888,
+      "learning_rate": 4.673387454118999e-05,
+      "loss": 0.5902,
+      "num_input_tokens_seen": 31293440,
+      "step": 3820
+    },
+    {
+      "epoch": 0.49489598139294483,
+      "grad_norm": 0.5781430006027222,
+      "learning_rate": 4.671713696554452e-05,
+      "loss": 0.8908,
+      "num_input_tokens_seen": 31375360,
+      "step": 3830
+    },
+    {
+      "epoch": 0.49618813800232586,
+      "grad_norm": 0.5275561213493347,
+      "learning_rate": 4.6700359625900724e-05,
+      "loss": 0.9977,
+      "num_input_tokens_seen": 31457280,
+      "step": 3840
+    },
+    {
+      "epoch": 0.49748029461170695,
+      "grad_norm": 0.7814369201660156,
+      "learning_rate": 4.668354255297785e-05,
+      "loss": 0.8617,
+      "num_input_tokens_seen": 31539200,
+      "step": 3850
+    },
+    {
+      "epoch": 0.498772451221088,
+      "grad_norm": 0.26794329285621643,
+      "learning_rate": 4.666668577756793e-05,
+      "loss": 0.7011,
+      "num_input_tokens_seen": 31621120,
+      "step": 3860
+    },
+    {
+      "epoch": 0.500064607830469,
+      "grad_norm": 0.4963877201080322,
+      "learning_rate": 4.664978933053562e-05,
+      "loss": 0.8713,
+      "num_input_tokens_seen": 31703040,
+      "step": 3870
+    },
+    {
+      "epoch": 0.5013567644398501,
+      "grad_norm": 0.748673141002655,
+      "learning_rate": 4.6632853242818274e-05,
+      "loss": 1.0985,
+      "num_input_tokens_seen": 31784960,
+      "step": 3880
+    },
+    {
+      "epoch": 0.5026489210492312,
+      "grad_norm": 0.2532234489917755,
+      "learning_rate": 4.66158775454258e-05,
+      "loss": 0.8102,
+      "num_input_tokens_seen": 31866880,
+      "step": 3890
+    },
+    {
+      "epoch": 0.5039410776586122,
+      "grad_norm": 0.4093916416168213,
+      "learning_rate": 4.659886226944063e-05,
+      "loss": 1.0378,
+      "num_input_tokens_seen": 31948800,
+      "step": 3900
+    },
+    {
+      "epoch": 0.5052332342679933,
+      "grad_norm": 0.656688392162323,
+      "learning_rate": 4.658180744601769e-05,
+      "loss": 0.9426,
+      "num_input_tokens_seen": 32030720,
+      "step": 3910
+    },
+    {
+      "epoch": 0.5065253908773744,
+      "grad_norm": 0.7212385535240173,
+      "learning_rate": 4.6564713106384296e-05,
+      "loss": 1.1089,
+      "num_input_tokens_seen": 32112640,
+      "step": 3920
+    },
+    {
+      "epoch": 0.5078175474867554,
+      "grad_norm": 0.5661940574645996,
+      "learning_rate": 4.65475792818401e-05,
+      "loss": 1.0339,
+      "num_input_tokens_seen": 32194560,
+      "step": 3930
+    },
+    {
+      "epoch": 0.5091097040961364,
+      "grad_norm": 0.5363628268241882,
+      "learning_rate": 4.653040600375709e-05,
+      "loss": 1.1407,
+      "num_input_tokens_seen": 32276480,
+      "step": 3940
+    },
+    {
+      "epoch": 0.5104018607055175,
+      "grad_norm": 0.29208219051361084,
+      "learning_rate": 4.6513193303579476e-05,
+      "loss": 1.1492,
+      "num_input_tokens_seen": 32358400,
+      "step": 3950
+    },
+    {
+      "epoch": 0.5116940173148986,
+      "grad_norm": 0.7336292862892151,
+      "learning_rate": 4.6495941212823644e-05,
+      "loss": 0.8435,
+      "num_input_tokens_seen": 32440320,
+      "step": 3960
+    },
+    {
+      "epoch": 0.5129861739242796,
+      "grad_norm": 0.45039644837379456,
+      "learning_rate": 4.647864976307811e-05,
+      "loss": 0.6948,
+      "num_input_tokens_seen": 32522240,
+      "step": 3970
+    },
+    {
+      "epoch": 0.5142783305336607,
+      "grad_norm": 0.4214424788951874,
+      "learning_rate": 4.646131898600345e-05,
+      "loss": 0.8107,
+      "num_input_tokens_seen": 32604160,
+      "step": 3980
+    },
+    {
+      "epoch": 0.5155704871430418,
+      "grad_norm": 0.5167589783668518,
+      "learning_rate": 4.644394891333227e-05,
+      "loss": 1.0497,
+      "num_input_tokens_seen": 32686080,
+      "step": 3990
+    },
+    {
+      "epoch": 0.5168626437524227,
+      "grad_norm": 0.5262385010719299,
+      "learning_rate": 4.64265395768691e-05,
+      "loss": 0.988,
+      "num_input_tokens_seen": 32768000,
+      "step": 4000
+    },
+    {
+      "epoch": 0.5181548003618038,
+      "grad_norm": 0.464677631855011,
+      "learning_rate": 4.6409091008490365e-05,
+      "loss": 0.9792,
+      "num_input_tokens_seen": 32849920,
+      "step": 4010
+    },
+    {
+      "epoch": 0.5194469569711849,
+      "grad_norm": 0.4901339113712311,
+      "learning_rate": 4.639160324014433e-05,
+      "loss": 0.7922,
+      "num_input_tokens_seen": 32931840,
+      "step": 4020
+    },
+    {
+      "epoch": 0.520739113580566,
+      "grad_norm": 0.47087928652763367,
+      "learning_rate": 4.637407630385104e-05,
+      "loss": 1.2068,
+      "num_input_tokens_seen": 33013760,
+      "step": 4030
+    },
+    {
+      "epoch": 0.522031270189947,
+      "grad_norm": 0.510374128818512,
+      "learning_rate": 4.6356510231702254e-05,
+      "loss": 0.9503,
+      "num_input_tokens_seen": 33095680,
+      "step": 4040
+    },
+    {
+      "epoch": 0.5233234267993281,
+      "grad_norm": 0.5987057685852051,
+      "learning_rate": 4.633890505586139e-05,
+      "loss": 0.8793,
+      "num_input_tokens_seen": 33177600,
+      "step": 4050
+    },
+    {
+      "epoch": 0.5246155834087092,
+      "grad_norm": 0.5550726652145386,
+      "learning_rate": 4.6321260808563445e-05,
+      "loss": 1.0048,
+      "num_input_tokens_seen": 33259520,
+      "step": 4060
+    },
+    {
+      "epoch": 0.5259077400180902,
+      "grad_norm": 0.727328896522522,
+      "learning_rate": 4.630357752211498e-05,
+      "loss": 0.8116,
+      "num_input_tokens_seen": 33341440,
+      "step": 4070
+    },
+    {
+      "epoch": 0.5271998966274712,
+      "grad_norm": 1.1229157447814941,
+      "learning_rate": 4.6285855228894025e-05,
+      "loss": 1.202,
+      "num_input_tokens_seen": 33423360,
+      "step": 4080
+    },
+    {
+      "epoch": 0.5284920532368523,
+      "grad_norm": 0.8320699334144592,
+      "learning_rate": 4.626809396135003e-05,
+      "loss": 1.2678,
+      "num_input_tokens_seen": 33505280,
+      "step": 4090
+    },
+    {
+      "epoch": 0.5297842098462333,
+      "grad_norm": 0.3679317533969879,
+      "learning_rate": 4.6250293752003834e-05,
+      "loss": 1.052,
+      "num_input_tokens_seen": 33587200,
+      "step": 4100
+    },
+    {
+      "epoch": 0.5310763664556144,
+      "grad_norm": 0.5202254056930542,
+      "learning_rate": 4.623245463344753e-05,
+      "loss": 0.9749,
+      "num_input_tokens_seen": 33669120,
+      "step": 4110
+    },
+    {
+      "epoch": 0.5323685230649955,
+      "grad_norm": 0.6384007930755615,
+      "learning_rate": 4.6214576638344484e-05,
+      "loss": 0.9361,
+      "num_input_tokens_seen": 33751040,
+      "step": 4120
+    },
+    {
+      "epoch": 0.5336606796743766,
+      "grad_norm": 0.39264151453971863,
+      "learning_rate": 4.619665979942924e-05,
+      "loss": 1.0865,
+      "num_input_tokens_seen": 33832960,
+      "step": 4130
+    },
+    {
+      "epoch": 0.5349528362837576,
+      "grad_norm": 0.5233849287033081,
+      "learning_rate": 4.617870414950748e-05,
+      "loss": 0.8994,
+      "num_input_tokens_seen": 33914880,
+      "step": 4140
+    },
+    {
+      "epoch": 0.5362449928931387,
+      "grad_norm": 0.4549662172794342,
+      "learning_rate": 4.616070972145591e-05,
+      "loss": 1.0843,
+      "num_input_tokens_seen": 33996800,
+      "step": 4150
+    },
+    {
+      "epoch": 0.5375371495025197,
+      "grad_norm": 0.5111933350563049,
+      "learning_rate": 4.614267654822228e-05,
+      "loss": 1.1799,
+      "num_input_tokens_seen": 34078720,
+      "step": 4160
+    },
+    {
+      "epoch": 0.5388293061119007,
+      "grad_norm": 0.49319344758987427,
+      "learning_rate": 4.612460466282525e-05,
+      "loss": 1.018,
+      "num_input_tokens_seen": 34160640,
+      "step": 4170
+    },
+    {
+      "epoch": 0.5401214627212818,
+      "grad_norm": 0.5339105725288391,
+      "learning_rate": 4.610649409835438e-05,
+      "loss": 1.2698,
+      "num_input_tokens_seen": 34242560,
+      "step": 4180
+    },
+    {
+      "epoch": 0.5414136193306629,
+      "grad_norm": 0.6269609928131104,
+      "learning_rate": 4.608834488797006e-05,
+      "loss": 1.3887,
+      "num_input_tokens_seen": 34324480,
+      "step": 4190
+    },
+    {
+      "epoch": 0.542705775940044,
+      "grad_norm": 0.5303708910942078,
+      "learning_rate": 4.607015706490341e-05,
+      "loss": 0.983,
+      "num_input_tokens_seen": 34406400,
+      "step": 4200
+    },
+    {
+      "epoch": 0.543997932549425,
+      "grad_norm": 0.3554169237613678,
+      "learning_rate": 4.6051930662456276e-05,
+      "loss": 0.8547,
+      "num_input_tokens_seen": 34488320,
+      "step": 4210
+    },
+    {
+      "epoch": 0.5452900891588061,
+      "grad_norm": 0.6087794899940491,
+      "learning_rate": 4.603366571400114e-05,
+      "loss": 1.1971,
+      "num_input_tokens_seen": 34570240,
+      "step": 4220
+    },
+    {
+      "epoch": 0.5465822457681871,
+      "grad_norm": 0.47708624601364136,
+      "learning_rate": 4.601536225298104e-05,
+      "loss": 0.927,
+      "num_input_tokens_seen": 34652160,
+      "step": 4230
+    },
+    {
+      "epoch": 0.5478744023775681,
+      "grad_norm": 0.2548501789569855,
+      "learning_rate": 4.5997020312909565e-05,
+      "loss": 0.8928,
+      "num_input_tokens_seen": 34734080,
+      "step": 4240
+    },
+    {
+      "epoch": 0.5491665589869492,
+      "grad_norm": 0.3759148418903351,
+      "learning_rate": 4.597863992737072e-05,
+      "loss": 0.9546,
+      "num_input_tokens_seen": 34816000,
+      "step": 4250
+    },
+    {
+      "epoch": 0.5504587155963303,
+      "grad_norm": 0.5236682891845703,
+      "learning_rate": 4.5960221130018946e-05,
+      "loss": 1.1029,
+      "num_input_tokens_seen": 34897920,
+      "step": 4260
+    },
+    {
+      "epoch": 0.5517508722057113,
+      "grad_norm": 0.5889397263526917,
+      "learning_rate": 4.594176395457897e-05,
+      "loss": 1.1309,
+      "num_input_tokens_seen": 34979840,
+      "step": 4270
+    },
+    {
+      "epoch": 0.5530430288150924,
+      "grad_norm": 0.5856964588165283,
+      "learning_rate": 4.592326843484583e-05,
+      "loss": 0.8424,
+      "num_input_tokens_seen": 35061760,
+      "step": 4280
+    },
+    {
+      "epoch": 0.5543351854244735,
+      "grad_norm": 0.5260830521583557,
+      "learning_rate": 4.590473460468475e-05,
+      "loss": 1.2382,
+      "num_input_tokens_seen": 35143680,
+      "step": 4290
+    },
+    {
+      "epoch": 0.5556273420338546,
+      "grad_norm": 0.5527287721633911,
+      "learning_rate": 4.58861624980311e-05,
+      "loss": 1.0987,
+      "num_input_tokens_seen": 35225600,
+      "step": 4300
+    },
+    {
+      "epoch": 0.5569194986432355,
+      "grad_norm": 0.6066960096359253,
+      "learning_rate": 4.586755214889035e-05,
+      "loss": 0.6929,
+      "num_input_tokens_seen": 35307520,
+      "step": 4310
+    },
+    {
+      "epoch": 0.5582116552526166,
+      "grad_norm": 0.8027804493904114,
+      "learning_rate": 4.584890359133797e-05,
+      "loss": 0.9671,
+      "num_input_tokens_seen": 35389440,
+      "step": 4320
+    },
+    {
+      "epoch": 0.5595038118619977,
+      "grad_norm": 0.5972751975059509,
+      "learning_rate": 4.58302168595194e-05,
+      "loss": 0.7893,
+      "num_input_tokens_seen": 35471360,
+      "step": 4330
+    },
+    {
+      "epoch": 0.5607959684713787,
+      "grad_norm": 0.36377179622650146,
+      "learning_rate": 4.5811491987649994e-05,
+      "loss": 0.8821,
+      "num_input_tokens_seen": 35553280,
+      "step": 4340
+    },
+    {
+      "epoch": 0.5620881250807598,
+      "grad_norm": 0.5296213626861572,
+      "learning_rate": 4.579272901001491e-05,
+      "loss": 0.8082,
+      "num_input_tokens_seen": 35635200,
+      "step": 4350
+    },
+    {
+      "epoch": 0.5633802816901409,
+      "grad_norm": 0.4613911509513855,
+      "learning_rate": 4.57739279609691e-05,
+      "loss": 0.9638,
+      "num_input_tokens_seen": 35717120,
+      "step": 4360
+    },
+    {
+      "epoch": 0.5646724382995219,
+      "grad_norm": 0.5832899212837219,
+      "learning_rate": 4.57550888749372e-05,
+      "loss": 1.0087,
+      "num_input_tokens_seen": 35799040,
+      "step": 4370
+    },
+    {
+      "epoch": 0.5659645949089029,
+      "grad_norm": 0.461342453956604,
+      "learning_rate": 4.5736211786413524e-05,
+      "loss": 1.0578,
+      "num_input_tokens_seen": 35880960,
+      "step": 4380
+    },
+    {
+      "epoch": 0.567256751518284,
+      "grad_norm": 0.5198622941970825,
+      "learning_rate": 4.571729672996195e-05,
+      "loss": 0.5186,
+      "num_input_tokens_seen": 35962880,
+      "step": 4390
+    },
+    {
+      "epoch": 0.5685489081276651,
+      "grad_norm": 0.4189557135105133,
+      "learning_rate": 4.5698343740215865e-05,
+      "loss": 1.0137,
+      "num_input_tokens_seen": 36044800,
+      "step": 4400
+    },
+    {
+      "epoch": 0.5698410647370461,
+      "grad_norm": 0.4138655364513397,
+      "learning_rate": 4.5679352851878135e-05,
+      "loss": 0.91,
+      "num_input_tokens_seen": 36126720,
+      "step": 4410
+    },
+    {
+      "epoch": 0.5711332213464272,
+      "grad_norm": 0.327007532119751,
+      "learning_rate": 4.5660324099721005e-05,
+      "loss": 0.8943,
+      "num_input_tokens_seen": 36208640,
+      "step": 4420
+    },
+    {
+      "epoch": 0.5724253779558083,
+      "grad_norm": 0.48592689633369446,
+      "learning_rate": 4.5641257518586044e-05,
+      "loss": 0.8402,
+      "num_input_tokens_seen": 36290560,
+      "step": 4430
+    },
+    {
+      "epoch": 0.5737175345651893,
+      "grad_norm": 0.5149102210998535,
+      "learning_rate": 4.562215314338411e-05,
+      "loss": 0.8945,
+      "num_input_tokens_seen": 36372480,
+      "step": 4440
+    },
+    {
+      "epoch": 0.5750096911745703,
+      "grad_norm": 0.47241947054862976,
+      "learning_rate": 4.560301100909522e-05,
+      "loss": 1.3013,
+      "num_input_tokens_seen": 36454400,
+      "step": 4450
+    },
+    {
+      "epoch": 0.5763018477839514,
+      "grad_norm": 0.6244943737983704,
+      "learning_rate": 4.558383115076857e-05,
+      "loss": 0.8028,
+      "num_input_tokens_seen": 36536320,
+      "step": 4460
+    },
+    {
+      "epoch": 0.5775940043933324,
+      "grad_norm": 0.42100536823272705,
+      "learning_rate": 4.556461360352241e-05,
+      "loss": 1.1116,
+      "num_input_tokens_seen": 36618240,
+      "step": 4470
+    },
+    {
+      "epoch": 0.5788861610027135,
+      "grad_norm": 0.5363230109214783,
+      "learning_rate": 4.554535840254398e-05,
+      "loss": 1.094,
+      "num_input_tokens_seen": 36700160,
+      "step": 4480
+    },
+    {
+      "epoch": 0.5801783176120946,
+      "grad_norm": 0.6321589350700378,
+      "learning_rate": 4.552606558308951e-05,
+      "loss": 0.9717,
+      "num_input_tokens_seen": 36782080,
+      "step": 4490
+    },
+    {
+      "epoch": 0.5814704742214757,
+      "grad_norm": 0.7036007046699524,
+      "learning_rate": 4.550673518048405e-05,
+      "loss": 1.0816,
+      "num_input_tokens_seen": 36864000,
+      "step": 4500
+    },
+    {
+      "epoch": 0.5827626308308567,
+      "grad_norm": 0.4853123724460602,
+      "learning_rate": 4.548736723012153e-05,
+      "loss": 0.9271,
+      "num_input_tokens_seen": 36945920,
+      "step": 4510
+    },
+    {
+      "epoch": 0.5840547874402378,
+      "grad_norm": 0.7495065331459045,
+      "learning_rate": 4.5467961767464575e-05,
+      "loss": 0.9318,
+      "num_input_tokens_seen": 37027840,
+      "step": 4520
+    },
+    {
+      "epoch": 0.5853469440496188,
+      "grad_norm": 0.43084239959716797,
+      "learning_rate": 4.5448518828044515e-05,
+      "loss": 0.7255,
+      "num_input_tokens_seen": 37109760,
+      "step": 4530
+    },
+    {
+      "epoch": 0.5866391006589998,
+      "grad_norm": 0.560870885848999,
+      "learning_rate": 4.5429038447461315e-05,
+      "loss": 1.1148,
+      "num_input_tokens_seen": 37191680,
+      "step": 4540
+    },
+    {
+      "epoch": 0.5879312572683809,
+      "grad_norm": 0.5736927390098572,
+      "learning_rate": 4.540952066138347e-05,
+      "loss": 1.1077,
+      "num_input_tokens_seen": 37273600,
+      "step": 4550
+    },
+    {
+      "epoch": 0.589223413877762,
+      "grad_norm": 0.27289411425590515,
+      "learning_rate": 4.538996550554798e-05,
+      "loss": 0.923,
+      "num_input_tokens_seen": 37355520,
+      "step": 4560
+    },
+    {
+      "epoch": 0.5905155704871431,
+      "grad_norm": 0.4996427893638611,
+      "learning_rate": 4.537037301576026e-05,
+      "loss": 0.9954,
+      "num_input_tokens_seen": 37437440,
+      "step": 4570
+    },
+    {
+      "epoch": 0.5918077270965241,
+      "grad_norm": 0.4833759367465973,
+      "learning_rate": 4.535074322789408e-05,
+      "loss": 0.9237,
+      "num_input_tokens_seen": 37519360,
+      "step": 4580
+    },
+    {
+      "epoch": 0.5930998837059052,
+      "grad_norm": 0.20252840220928192,
+      "learning_rate": 4.5331076177891527e-05,
+      "loss": 1.0009,
+      "num_input_tokens_seen": 37601280,
+      "step": 4590
+    },
+    {
+      "epoch": 0.5943920403152863,
+      "grad_norm": 0.49745073914527893,
+      "learning_rate": 4.531137190176289e-05,
+      "loss": 1.0067,
+      "num_input_tokens_seen": 37683200,
+      "step": 4600
+    },
+    {
+      "epoch": 0.5956841969246672,
+      "grad_norm": 0.5191871523857117,
+      "learning_rate": 4.529163043558662e-05,
+      "loss": 0.7292,
+      "num_input_tokens_seen": 37765120,
+      "step": 4610
+    },
+    {
+      "epoch": 0.5969763535340483,
+      "grad_norm": 0.7261500954627991,
+      "learning_rate": 4.527185181550928e-05,
+      "loss": 1.0878,
+      "num_input_tokens_seen": 37847040,
+      "step": 4620
+    },
+    {
+      "epoch": 0.5982685101434294,
+      "grad_norm": 0.6851155757904053,
+      "learning_rate": 4.525203607774544e-05,
+      "loss": 0.741,
+      "num_input_tokens_seen": 37928960,
+      "step": 4630
+    },
+    {
+      "epoch": 0.5995606667528104,
+      "grad_norm": 0.5459215044975281,
+      "learning_rate": 4.5232183258577655e-05,
+      "loss": 0.498,
+      "num_input_tokens_seen": 38010880,
+      "step": 4640
+    },
+    {
+      "epoch": 0.6008528233621915,
+      "grad_norm": 0.3943910300731659,
+      "learning_rate": 4.5212293394356356e-05,
+      "loss": 0.854,
+      "num_input_tokens_seen": 38092800,
+      "step": 4650
+    },
+    {
+      "epoch": 0.6021449799715726,
+      "grad_norm": 0.402898907661438,
+      "learning_rate": 4.519236652149981e-05,
+      "loss": 0.8904,
+      "num_input_tokens_seen": 38174720,
+      "step": 4660
+    },
+    {
+      "epoch": 0.6034371365809537,
+      "grad_norm": 0.23560667037963867,
+      "learning_rate": 4.517240267649405e-05,
+      "loss": 0.6613,
+      "num_input_tokens_seen": 38256640,
+      "step": 4670
+    },
+    {
+      "epoch": 0.6047292931903346,
+      "grad_norm": 0.3227959871292114,
+      "learning_rate": 4.515240189589282e-05,
+      "loss": 0.679,
+      "num_input_tokens_seen": 38338560,
+      "step": 4680
+    },
+    {
+      "epoch": 0.6060214497997157,
+      "grad_norm": 0.29034727811813354,
+      "learning_rate": 4.5132364216317446e-05,
+      "loss": 0.7213,
+      "num_input_tokens_seen": 38420480,
+      "step": 4690
+    },
+    {
+      "epoch": 0.6073136064090968,
+      "grad_norm": 0.5373560190200806,
+      "learning_rate": 4.5112289674456864e-05,
+      "loss": 1.0668,
+      "num_input_tokens_seen": 38502400,
+      "step": 4700
+    },
+    {
+      "epoch": 0.6086057630184778,
+      "grad_norm": 0.48361408710479736,
+      "learning_rate": 4.509217830706749e-05,
+      "loss": 0.6861,
+      "num_input_tokens_seen": 38584320,
+      "step": 4710
+    },
+    {
+      "epoch": 0.6098979196278589,
+      "grad_norm": 0.4311217963695526,
+      "learning_rate": 4.5072030150973154e-05,
+      "loss": 0.9106,
+      "num_input_tokens_seen": 38666240,
+      "step": 4720
+    },
+    {
+      "epoch": 0.61119007623724,
+      "grad_norm": 0.47615599632263184,
+      "learning_rate": 4.505184524306506e-05,
+      "loss": 0.73,
+      "num_input_tokens_seen": 38748160,
+      "step": 4730
+    },
+    {
+      "epoch": 0.612482232846621,
+      "grad_norm": 0.5375292897224426,
+      "learning_rate": 4.50316236203017e-05,
+      "loss": 1.3151,
+      "num_input_tokens_seen": 38830080,
+      "step": 4740
+    },
+    {
+      "epoch": 0.613774389456002,
+      "grad_norm": 0.42838889360427856,
+      "learning_rate": 4.5011365319708796e-05,
+      "loss": 1.0097,
+      "num_input_tokens_seen": 38912000,
+      "step": 4750
+    },
+    {
+      "epoch": 0.6150665460653831,
+      "grad_norm": 0.5048761367797852,
+      "learning_rate": 4.499107037837922e-05,
+      "loss": 1.0672,
+      "num_input_tokens_seen": 38993920,
+      "step": 4760
+    },
+    {
+      "epoch": 0.6163587026747642,
+      "grad_norm": 0.4704453647136688,
+      "learning_rate": 4.497073883347293e-05,
+      "loss": 0.8592,
+      "num_input_tokens_seen": 39075840,
+      "step": 4770
+    },
+    {
+      "epoch": 0.6176508592841452,
+      "grad_norm": 0.5054408311843872,
+      "learning_rate": 4.495037072221692e-05,
+      "loss": 0.979,
+      "num_input_tokens_seen": 39157760,
+      "step": 4780
+    },
+    {
+      "epoch": 0.6189430158935263,
+      "grad_norm": 0.4307916462421417,
+      "learning_rate": 4.49299660819051e-05,
+      "loss": 1.1349,
+      "num_input_tokens_seen": 39239680,
+      "step": 4790
+    },
+    {
+      "epoch": 0.6202351725029074,
+      "grad_norm": 0.7351888418197632,
+      "learning_rate": 4.490952494989834e-05,
+      "loss": 0.8819,
+      "num_input_tokens_seen": 39321600,
+      "step": 4800
+    },
+    {
+      "epoch": 0.6215273291122884,
+      "grad_norm": 0.7164278030395508,
+      "learning_rate": 4.4889047363624236e-05,
+      "loss": 1.0527,
+      "num_input_tokens_seen": 39403520,
+      "step": 4810
+    },
+    {
+      "epoch": 0.6228194857216695,
+      "grad_norm": 0.5169610381126404,
+      "learning_rate": 4.486853336057719e-05,
+      "loss": 1.1608,
+      "num_input_tokens_seen": 39485440,
+      "step": 4820
+    },
+    {
+      "epoch": 0.6241116423310505,
+      "grad_norm": 0.9453770518302917,
+      "learning_rate": 4.484798297831826e-05,
+      "loss": 0.9382,
+      "num_input_tokens_seen": 39567360,
+      "step": 4830
+    },
+    {
+      "epoch": 0.6254037989404316,
+      "grad_norm": 0.49842312932014465,
+      "learning_rate": 4.482739625447514e-05,
+      "loss": 0.9578,
+      "num_input_tokens_seen": 39649280,
+      "step": 4840
+    },
+    {
+      "epoch": 0.6266959555498126,
+      "grad_norm": 0.5001227855682373,
+      "learning_rate": 4.480677322674202e-05,
+      "loss": 0.5964,
+      "num_input_tokens_seen": 39731200,
+      "step": 4850
+    },
+    {
+      "epoch": 0.6279881121591937,
+      "grad_norm": 0.42781203985214233,
+      "learning_rate": 4.4786113932879605e-05,
+      "loss": 1.0508,
+      "num_input_tokens_seen": 39813120,
+      "step": 4860
+    },
+    {
+      "epoch": 0.6292802687685748,
+      "grad_norm": 0.3244008719921112,
+      "learning_rate": 4.476541841071498e-05,
+      "loss": 0.7638,
+      "num_input_tokens_seen": 39895040,
+      "step": 4870
+    },
+    {
+      "epoch": 0.6305724253779558,
+      "grad_norm": 0.3672725558280945,
+      "learning_rate": 4.4744686698141564e-05,
+      "loss": 0.5432,
+      "num_input_tokens_seen": 39976960,
+      "step": 4880
+    },
+    {
+      "epoch": 0.6318645819873369,
+      "grad_norm": 0.44159212708473206,
+      "learning_rate": 4.472391883311906e-05,
+      "loss": 1.0626,
+      "num_input_tokens_seen": 40058880,
+      "step": 4890
+    },
+    {
+      "epoch": 0.633156738596718,
+      "grad_norm": 0.5819474458694458,
+      "learning_rate": 4.470311485367335e-05,
+      "loss": 0.8067,
+      "num_input_tokens_seen": 40140800,
+      "step": 4900
+    },
+    {
+      "epoch": 0.6344488952060989,
+      "grad_norm": 0.618045449256897,
+      "learning_rate": 4.468227479789644e-05,
+      "loss": 0.8431,
+      "num_input_tokens_seen": 40222720,
+      "step": 4910
+    },
+    {
+      "epoch": 0.63574105181548,
+      "grad_norm": 0.6308985948562622,
+      "learning_rate": 4.4661398703946396e-05,
+      "loss": 0.92,
+      "num_input_tokens_seen": 40304640,
+      "step": 4920
+    },
+    {
+      "epoch": 0.6370332084248611,
+      "grad_norm": 0.49686864018440247,
+      "learning_rate": 4.464048661004727e-05,
+      "loss": 1.0994,
+      "num_input_tokens_seen": 40386560,
+      "step": 4930
+    },
+    {
+      "epoch": 0.6383253650342422,
+      "grad_norm": 0.5074322819709778,
+      "learning_rate": 4.461953855448903e-05,
+      "loss": 1.0147,
+      "num_input_tokens_seen": 40468480,
+      "step": 4940
+    },
+    {
+      "epoch": 0.6396175216436232,
+      "grad_norm": 0.7457041144371033,
+      "learning_rate": 4.4598554575627495e-05,
+      "loss": 1.1607,
+      "num_input_tokens_seen": 40550400,
+      "step": 4950
+    },
+    {
+      "epoch": 0.6409096782530043,
+      "grad_norm": 0.5672173500061035,
+      "learning_rate": 4.4577534711884244e-05,
+      "loss": 1.1159,
+      "num_input_tokens_seen": 40632320,
+      "step": 4960
+    },
+    {
+      "epoch": 0.6422018348623854,
+      "grad_norm": 0.3177810609340668,
+      "learning_rate": 4.455647900174658e-05,
+      "loss": 1.0039,
+      "num_input_tokens_seen": 40714240,
+      "step": 4970
+    },
+    {
+      "epoch": 0.6434939914717663,
+      "grad_norm": 0.6089414358139038,
+      "learning_rate": 4.453538748376742e-05,
+      "loss": 0.8848,
+      "num_input_tokens_seen": 40796160,
+      "step": 4980
+    },
+    {
+      "epoch": 0.6447861480811474,
+      "grad_norm": 0.44886818528175354,
+      "learning_rate": 4.451426019656526e-05,
+      "loss": 0.624,
+      "num_input_tokens_seen": 40878080,
+      "step": 4990
+    },
+    {
+      "epoch": 0.6460783046905285,
+      "grad_norm": 0.336834192276001,
+      "learning_rate": 4.449309717882409e-05,
+      "loss": 1.0137,
+      "num_input_tokens_seen": 40960000,
+      "step": 5000
+    },
+    {
+      "epoch": 0.6473704612999095,
+      "grad_norm": 0.8568751215934753,
+      "learning_rate": 4.4471898469293324e-05,
+      "loss": 0.9478,
+      "num_input_tokens_seen": 41041920,
+      "step": 5010
+    },
+    {
+      "epoch": 0.6486626179092906,
+      "grad_norm": 0.6555817127227783,
+      "learning_rate": 4.4450664106787706e-05,
+      "loss": 0.8829,
+      "num_input_tokens_seen": 41123840,
+      "step": 5020
+    },
+    {
+      "epoch": 0.6499547745186717,
+      "grad_norm": 0.656535267829895,
+      "learning_rate": 4.442939413018728e-05,
+      "loss": 1.0126,
+      "num_input_tokens_seen": 41205760,
+      "step": 5030
+    },
+    {
+      "epoch": 0.6512469311280528,
+      "grad_norm": 0.5044434666633606,
+      "learning_rate": 4.44080885784373e-05,
+      "loss": 0.8302,
+      "num_input_tokens_seen": 41287680,
+      "step": 5040
+    },
+    {
+      "epoch": 0.6525390877374337,
+      "grad_norm": 1.4959832429885864,
+      "learning_rate": 4.4386747490548156e-05,
+      "loss": 1.0349,
+      "num_input_tokens_seen": 41369600,
+      "step": 5050
+    },
+    {
+      "epoch": 0.6538312443468148,
+      "grad_norm": 0.623151421546936,
+      "learning_rate": 4.43653709055953e-05,
+      "loss": 1.1468,
+      "num_input_tokens_seen": 41451520,
+      "step": 5060
+    },
+    {
+      "epoch": 0.6551234009561959,
+      "grad_norm": 0.606468141078949,
+      "learning_rate": 4.434395886271917e-05,
+      "loss": 1.1647,
+      "num_input_tokens_seen": 41533440,
+      "step": 5070
+    },
+    {
+      "epoch": 0.6564155575655769,
+      "grad_norm": 0.5442026853561401,
+      "learning_rate": 4.4322511401125156e-05,
+      "loss": 1.2088,
+      "num_input_tokens_seen": 41615360,
+      "step": 5080
+    },
+    {
+      "epoch": 0.657707714174958,
+      "grad_norm": 0.5820073485374451,
+      "learning_rate": 4.430102856008347e-05,
+      "loss": 0.8394,
+      "num_input_tokens_seen": 41697280,
+      "step": 5090
+    },
+    {
+      "epoch": 0.6589998707843391,
+      "grad_norm": 0.6123668551445007,
+      "learning_rate": 4.427951037892911e-05,
+      "loss": 0.7364,
+      "num_input_tokens_seen": 41779200,
+      "step": 5100
+    },
+    {
+      "epoch": 0.6602920273937202,
+      "grad_norm": 0.4739777743816376,
+      "learning_rate": 4.4257956897061805e-05,
+      "loss": 0.8559,
+      "num_input_tokens_seen": 41861120,
+      "step": 5110
+    },
+    {
+      "epoch": 0.6615841840031012,
+      "grad_norm": 0.5354482531547546,
+      "learning_rate": 4.423636815394588e-05,
+      "loss": 0.9424,
+      "num_input_tokens_seen": 41943040,
+      "step": 5120
+    },
+    {
+      "epoch": 0.6628763406124822,
+      "grad_norm": 0.5862747430801392,
+      "learning_rate": 4.4214744189110266e-05,
+      "loss": 1.1937,
+      "num_input_tokens_seen": 42024960,
+      "step": 5130
+    },
+    {
+      "epoch": 0.6641684972218633,
+      "grad_norm": 0.4908977150917053,
+      "learning_rate": 4.4193085042148354e-05,
+      "loss": 1.0227,
+      "num_input_tokens_seen": 42106880,
+      "step": 5140
+    },
+    {
+      "epoch": 0.6654606538312443,
+      "grad_norm": 0.8534203171730042,
+      "learning_rate": 4.417139075271796e-05,
+      "loss": 1.3091,
+      "num_input_tokens_seen": 42188800,
+      "step": 5150
+    },
+    {
+      "epoch": 0.6667528104406254,
+      "grad_norm": 0.47156473994255066,
+      "learning_rate": 4.414966136054125e-05,
+      "loss": 0.6186,
+      "num_input_tokens_seen": 42270720,
+      "step": 5160
+    },
+    {
+      "epoch": 0.6680449670500065,
+      "grad_norm": 0.5299347043037415,
+      "learning_rate": 4.412789690540466e-05,
+      "loss": 0.7974,
+      "num_input_tokens_seen": 42352640,
+      "step": 5170
+    },
+    {
+      "epoch": 0.6693371236593875,
+      "grad_norm": 0.5082312226295471,
+      "learning_rate": 4.410609742715883e-05,
+      "loss": 0.8996,
+      "num_input_tokens_seen": 42434560,
+      "step": 5180
+    },
+    {
+      "epoch": 0.6706292802687686,
+      "grad_norm": 0.48898372054100037,
+      "learning_rate": 4.408426296571852e-05,
+      "loss": 1.0341,
+      "num_input_tokens_seen": 42516480,
+      "step": 5190
+    },
+    {
+      "epoch": 0.6719214368781496,
+      "grad_norm": 0.7845025658607483,
+      "learning_rate": 4.406239356106257e-05,
+      "loss": 0.9729,
+      "num_input_tokens_seen": 42598400,
+      "step": 5200
+    },
+    {
+      "epoch": 0.6732135934875307,
+      "grad_norm": 0.3541695475578308,
+      "learning_rate": 4.404048925323375e-05,
+      "loss": 0.6879,
+      "num_input_tokens_seen": 42680320,
+      "step": 5210
+    },
+    {
+      "epoch": 0.6745057500969117,
+      "grad_norm": 0.6555191874504089,
+      "learning_rate": 4.401855008233879e-05,
+      "loss": 1.2569,
+      "num_input_tokens_seen": 42762240,
+      "step": 5220
+    },
+    {
+      "epoch": 0.6757979067062928,
+      "grad_norm": 0.44755908846855164,
+      "learning_rate": 4.3996576088548214e-05,
+      "loss": 1.1061,
+      "num_input_tokens_seen": 42844160,
+      "step": 5230
+    },
+    {
+      "epoch": 0.6770900633156739,
+      "grad_norm": 0.4899725615978241,
+      "learning_rate": 4.397456731209634e-05,
+      "loss": 1.0198,
+      "num_input_tokens_seen": 42926080,
+      "step": 5240
+    },
+    {
+      "epoch": 0.6783822199250549,
+      "grad_norm": 0.30107763409614563,
+      "learning_rate": 4.395252379328115e-05,
+      "loss": 0.7051,
+      "num_input_tokens_seen": 43008000,
+      "step": 5250
+    },
+    {
+      "epoch": 0.679674376534436,
+      "grad_norm": 0.24033145606517792,
+      "learning_rate": 4.393044557246424e-05,
+      "loss": 0.853,
+      "num_input_tokens_seen": 43089920,
+      "step": 5260
+    },
+    {
+      "epoch": 0.680966533143817,
+      "grad_norm": 0.5303036570549011,
+      "learning_rate": 4.3908332690070765e-05,
+      "loss": 0.9311,
+      "num_input_tokens_seen": 43171840,
+      "step": 5270
+    },
+    {
+      "epoch": 0.682258689753198,
+      "grad_norm": 0.4759383499622345,
+      "learning_rate": 4.388618518658932e-05,
+      "loss": 1.3277,
+      "num_input_tokens_seen": 43253760,
+      "step": 5280
+    },
+    {
+      "epoch": 0.6835508463625791,
+      "grad_norm": 0.4998660385608673,
+      "learning_rate": 4.3864003102571916e-05,
+      "loss": 0.8287,
+      "num_input_tokens_seen": 43335680,
+      "step": 5290
+    },
+    {
+      "epoch": 0.6848430029719602,
+      "grad_norm": 0.47711753845214844,
+      "learning_rate": 4.384178647863385e-05,
+      "loss": 0.6284,
+      "num_input_tokens_seen": 43417600,
+      "step": 5300
+    },
+    {
+      "epoch": 0.6861351595813413,
+      "grad_norm": 0.6608704328536987,
+      "learning_rate": 4.381953535545369e-05,
+      "loss": 1.1245,
+      "num_input_tokens_seen": 43499520,
+      "step": 5310
+    },
+    {
+      "epoch": 0.6874273161907223,
+      "grad_norm": 0.6293668746948242,
+      "learning_rate": 4.3797249773773165e-05,
+      "loss": 0.8103,
+      "num_input_tokens_seen": 43581440,
+      "step": 5320
+    },
+    {
+      "epoch": 0.6887194728001034,
+      "grad_norm": 0.7585543990135193,
+      "learning_rate": 4.3774929774397086e-05,
+      "loss": 0.6222,
+      "num_input_tokens_seen": 43663360,
+      "step": 5330
+    },
+    {
+      "epoch": 0.6900116294094845,
+      "grad_norm": 0.6242161393165588,
+      "learning_rate": 4.375257539819328e-05,
+      "loss": 1.0592,
+      "num_input_tokens_seen": 43745280,
+      "step": 5340
+    },
+    {
+      "epoch": 0.6913037860188654,
+      "grad_norm": 0.47205057740211487,
+      "learning_rate": 4.373018668609256e-05,
+      "loss": 1.1099,
+      "num_input_tokens_seen": 43827200,
+      "step": 5350
+    },
+    {
+      "epoch": 0.6925959426282465,
+      "grad_norm": 0.6397084593772888,
+      "learning_rate": 4.370776367908854e-05,
+      "loss": 1.0354,
+      "num_input_tokens_seen": 43909120,
+      "step": 5360
+    },
+    {
+      "epoch": 0.6938880992376276,
+      "grad_norm": 0.5018695592880249,
+      "learning_rate": 4.368530641823769e-05,
+      "loss": 1.2438,
+      "num_input_tokens_seen": 43991040,
+      "step": 5370
+    },
+    {
+      "epoch": 0.6951802558470087,
+      "grad_norm": 0.5705529451370239,
+      "learning_rate": 4.3662814944659156e-05,
+      "loss": 0.9602,
+      "num_input_tokens_seen": 44072960,
+      "step": 5380
+    },
+    {
+      "epoch": 0.6964724124563897,
+      "grad_norm": 0.5414160490036011,
+      "learning_rate": 4.364028929953476e-05,
+      "loss": 0.9233,
+      "num_input_tokens_seen": 44154880,
+      "step": 5390
+    },
+    {
+      "epoch": 0.6977645690657708,
+      "grad_norm": 0.5052191615104675,
+      "learning_rate": 4.361772952410886e-05,
+      "loss": 1.1027,
+      "num_input_tokens_seen": 44236800,
+      "step": 5400
+    },
+    {
+      "epoch": 0.6990567256751519,
+      "grad_norm": 0.5868683457374573,
+      "learning_rate": 4.359513565968832e-05,
+      "loss": 1.0273,
+      "num_input_tokens_seen": 44318720,
+      "step": 5410
+    },
+    {
+      "epoch": 0.7003488822845328,
+      "grad_norm": 0.5551066994667053,
+      "learning_rate": 4.357250774764245e-05,
+      "loss": 0.8502,
+      "num_input_tokens_seen": 44400640,
+      "step": 5420
+    },
+    {
+      "epoch": 0.7016410388939139,
+      "grad_norm": 0.36578068137168884,
+      "learning_rate": 4.354984582940285e-05,
+      "loss": 0.7773,
+      "num_input_tokens_seen": 44482560,
+      "step": 5430
+    },
+    {
+      "epoch": 0.702933195503295,
+      "grad_norm": 0.8225210309028625,
+      "learning_rate": 4.35271499464634e-05,
+      "loss": 0.9399,
+      "num_input_tokens_seen": 44564480,
+      "step": 5440
+    },
+    {
+      "epoch": 0.704225352112676,
+      "grad_norm": 0.5141643285751343,
+      "learning_rate": 4.350442014038021e-05,
+      "loss": 0.9732,
+      "num_input_tokens_seen": 44646400,
+      "step": 5450
+    },
+    {
+      "epoch": 0.7055175087220571,
+      "grad_norm": 0.42161181569099426,
+      "learning_rate": 4.348165645277145e-05,
+      "loss": 0.9438,
+      "num_input_tokens_seen": 44728320,
+      "step": 5460
+    },
+    {
+      "epoch": 0.7068096653314382,
+      "grad_norm": 0.44062143564224243,
+      "learning_rate": 4.345885892531735e-05,
+      "loss": 1.3828,
+      "num_input_tokens_seen": 44810240,
+      "step": 5470
+    },
+    {
+      "epoch": 0.7081018219408193,
+      "grad_norm": 0.44696852564811707,
+      "learning_rate": 4.343602759976011e-05,
+      "loss": 1.1361,
+      "num_input_tokens_seen": 44892160,
+      "step": 5480
+    },
+    {
+      "epoch": 0.7093939785502003,
+      "grad_norm": 0.7456682920455933,
+      "learning_rate": 4.34131625179038e-05,
+      "loss": 0.9972,
+      "num_input_tokens_seen": 44974080,
+      "step": 5490
+    },
+    {
+      "epoch": 0.7106861351595813,
+      "grad_norm": 0.4710252285003662,
+      "learning_rate": 4.3390263721614286e-05,
+      "loss": 1.1289,
+      "num_input_tokens_seen": 45056000,
+      "step": 5500
+    },
+    {
+      "epoch": 0.7119782917689624,
+      "grad_norm": 0.4541804790496826,
+      "learning_rate": 4.33673312528192e-05,
+      "loss": 0.699,
+      "num_input_tokens_seen": 45137920,
+      "step": 5510
+    },
+    {
+      "epoch": 0.7132704483783434,
+      "grad_norm": 0.5204092860221863,
+      "learning_rate": 4.334436515350779e-05,
+      "loss": 1.1252,
+      "num_input_tokens_seen": 45219840,
+      "step": 5520
+    },
+    {
+      "epoch": 0.7145626049877245,
+      "grad_norm": 0.5446497201919556,
+      "learning_rate": 4.332136546573092e-05,
+      "loss": 1.0118,
+      "num_input_tokens_seen": 45301760,
+      "step": 5530
+    },
+    {
+      "epoch": 0.7158547615971056,
+      "grad_norm": 0.5484737753868103,
+      "learning_rate": 4.3298332231600925e-05,
+      "loss": 0.8958,
+      "num_input_tokens_seen": 45383680,
+      "step": 5540
+    },
+    {
+      "epoch": 0.7171469182064866,
+      "grad_norm": 0.4673605263233185,
+      "learning_rate": 4.327526549329157e-05,
+      "loss": 0.7937,
+      "num_input_tokens_seen": 45465600,
+      "step": 5550
+    },
+    {
+      "epoch": 0.7184390748158677,
+      "grad_norm": 0.4332837760448456,
+      "learning_rate": 4.325216529303798e-05,
+      "loss": 0.7012,
+      "num_input_tokens_seen": 45547520,
+      "step": 5560
+    },
+    {
+      "epoch": 0.7197312314252488,
+      "grad_norm": 0.5795179009437561,
+      "learning_rate": 4.3229031673136514e-05,
+      "loss": 0.7965,
+      "num_input_tokens_seen": 45629440,
+      "step": 5570
+    },
+    {
+      "epoch": 0.7210233880346298,
+      "grad_norm": 0.5822563767433167,
+      "learning_rate": 4.320586467594476e-05,
+      "loss": 0.9927,
+      "num_input_tokens_seen": 45711360,
+      "step": 5580
+    },
+    {
+      "epoch": 0.7223155446440108,
+      "grad_norm": 0.3716464340686798,
+      "learning_rate": 4.3182664343881415e-05,
+      "loss": 0.8541,
+      "num_input_tokens_seen": 45793280,
+      "step": 5590
+    },
+    {
+      "epoch": 0.7236077012533919,
+      "grad_norm": 0.6027759313583374,
+      "learning_rate": 4.315943071942619e-05,
+      "loss": 0.989,
+      "num_input_tokens_seen": 45875200,
+      "step": 5600
+    },
+    {
+      "epoch": 0.724899857862773,
+      "grad_norm": 0.4732694923877716,
+      "learning_rate": 4.313616384511976e-05,
+      "loss": 1.0818,
+      "num_input_tokens_seen": 45957120,
+      "step": 5610
+    },
+    {
+      "epoch": 0.726192014472154,
+      "grad_norm": 0.5868046283721924,
+      "learning_rate": 4.3112863763563695e-05,
+      "loss": 1.0183,
+      "num_input_tokens_seen": 46039040,
+      "step": 5620
+    },
+    {
+      "epoch": 0.7274841710815351,
+      "grad_norm": 0.5000009536743164,
+      "learning_rate": 4.308953051742036e-05,
+      "loss": 1.1861,
+      "num_input_tokens_seen": 46120960,
+      "step": 5630
+    },
+    {
+      "epoch": 0.7287763276909162,
+      "grad_norm": 0.5284093022346497,
+      "learning_rate": 4.3066164149412844e-05,
+      "loss": 0.8941,
+      "num_input_tokens_seen": 46202880,
+      "step": 5640
+    },
+    {
+      "epoch": 0.7300684843002972,
+      "grad_norm": 0.4090714454650879,
+      "learning_rate": 4.304276470232488e-05,
+      "loss": 0.8618,
+      "num_input_tokens_seen": 46284800,
+      "step": 5650
+    },
+    {
+      "epoch": 0.7313606409096782,
+      "grad_norm": 0.6819685697555542,
+      "learning_rate": 4.3019332219000766e-05,
+      "loss": 0.8772,
+      "num_input_tokens_seen": 46366720,
+      "step": 5660
+    },
+    {
+      "epoch": 0.7326527975190593,
+      "grad_norm": 0.8263905048370361,
+      "learning_rate": 4.299586674234529e-05,
+      "loss": 1.0889,
+      "num_input_tokens_seen": 46448640,
+      "step": 5670
+    },
+    {
+      "epoch": 0.7339449541284404,
+      "grad_norm": 0.5756688714027405,
+      "learning_rate": 4.2972368315323676e-05,
+      "loss": 0.7212,
+      "num_input_tokens_seen": 46530560,
+      "step": 5680
+    },
+    {
+      "epoch": 0.7352371107378214,
+      "grad_norm": 0.5566821694374084,
+      "learning_rate": 4.294883698096143e-05,
+      "loss": 0.8605,
+      "num_input_tokens_seen": 46612480,
+      "step": 5690
+    },
+    {
+      "epoch": 0.7365292673472025,
+      "grad_norm": 0.5805408358573914,
+      "learning_rate": 4.292527278234435e-05,
+      "loss": 0.9687,
+      "num_input_tokens_seen": 46694400,
+      "step": 5700
+    },
+    {
+      "epoch": 0.7378214239565836,
+      "grad_norm": 0.19487954676151276,
+      "learning_rate": 4.290167576261841e-05,
+      "loss": 0.5747,
+      "num_input_tokens_seen": 46776320,
+      "step": 5710
+    },
+    {
+      "epoch": 0.7391135805659645,
+      "grad_norm": 0.5343565940856934,
+      "learning_rate": 4.2878045964989646e-05,
+      "loss": 0.7948,
+      "num_input_tokens_seen": 46858240,
+      "step": 5720
+    },
+    {
+      "epoch": 0.7404057371753456,
+      "grad_norm": 2.30495023727417,
+      "learning_rate": 4.285438343272414e-05,
+      "loss": 1.0718,
+      "num_input_tokens_seen": 46940160,
+      "step": 5730
+    },
+    {
+      "epoch": 0.7416978937847267,
+      "grad_norm": 0.5874922871589661,
+      "learning_rate": 4.283068820914791e-05,
+      "loss": 0.9519,
+      "num_input_tokens_seen": 47022080,
+      "step": 5740
+    },
+    {
+      "epoch": 0.7429900503941078,
+      "grad_norm": 0.4819450378417969,
+      "learning_rate": 4.2806960337646804e-05,
+      "loss": 1.015,
+      "num_input_tokens_seen": 47104000,
+      "step": 5750
+    },
+    {
+      "epoch": 0.7442822070034888,
+      "grad_norm": 0.6047049164772034,
+      "learning_rate": 4.278319986166649e-05,
+      "loss": 0.8983,
+      "num_input_tokens_seen": 47185920,
+      "step": 5760
+    },
+    {
+      "epoch": 0.7455743636128699,
+      "grad_norm": 0.32134854793548584,
+      "learning_rate": 4.27594068247123e-05,
+      "loss": 0.8726,
+      "num_input_tokens_seen": 47267840,
+      "step": 5770
+    },
+    {
+      "epoch": 0.746866520222251,
+      "grad_norm": 0.7429246306419373,
+      "learning_rate": 4.27355812703492e-05,
+      "loss": 0.7588,
+      "num_input_tokens_seen": 47349760,
+      "step": 5780
+    },
+    {
+      "epoch": 0.748158676831632,
+      "grad_norm": 0.5023065805435181,
+      "learning_rate": 4.2711723242201695e-05,
+      "loss": 1.1223,
+      "num_input_tokens_seen": 47431680,
+      "step": 5790
+    },
+    {
+      "epoch": 0.749450833441013,
+      "grad_norm": 0.38930198550224304,
+      "learning_rate": 4.268783278395374e-05,
+      "loss": 0.7947,
+      "num_input_tokens_seen": 47513600,
+      "step": 5800
+    },
+    {
+      "epoch": 0.7507429900503941,
+      "grad_norm": 0.488908976316452,
+      "learning_rate": 4.2663909939348684e-05,
+      "loss": 0.9494,
+      "num_input_tokens_seen": 47595520,
+      "step": 5810
+    },
+    {
+      "epoch": 0.7520351466597751,
+      "grad_norm": 0.4929217994213104,
+      "learning_rate": 4.263995475218917e-05,
+      "loss": 0.7431,
+      "num_input_tokens_seen": 47677440,
+      "step": 5820
+    },
+    {
+      "epoch": 0.7533273032691562,
+      "grad_norm": 0.4818821847438812,
+      "learning_rate": 4.2615967266337045e-05,
+      "loss": 0.8616,
+      "num_input_tokens_seen": 47759360,
+      "step": 5830
+    },
+    {
+      "epoch": 0.7546194598785373,
+      "grad_norm": 0.5090234875679016,
+      "learning_rate": 4.2591947525713326e-05,
+      "loss": 1.1937,
+      "num_input_tokens_seen": 47841280,
+      "step": 5840
+    },
+    {
+      "epoch": 0.7559116164879184,
+      "grad_norm": 0.6463255882263184,
+      "learning_rate": 4.256789557429806e-05,
+      "loss": 0.8848,
+      "num_input_tokens_seen": 47923200,
+      "step": 5850
+    },
+    {
+      "epoch": 0.7572037730972994,
+      "grad_norm": 0.5930004119873047,
+      "learning_rate": 4.254381145613027e-05,
+      "loss": 0.9741,
+      "num_input_tokens_seen": 48005120,
+      "step": 5860
+    },
+    {
+      "epoch": 0.7584959297066804,
+      "grad_norm": 0.5355798602104187,
+      "learning_rate": 4.251969521530791e-05,
+      "loss": 1.1395,
+      "num_input_tokens_seen": 48087040,
+      "step": 5870
+    },
+    {
+      "epoch": 0.7597880863160615,
+      "grad_norm": 0.4238862693309784,
+      "learning_rate": 4.2495546895987724e-05,
+      "loss": 1.074,
+      "num_input_tokens_seen": 48168960,
+      "step": 5880
+    },
+    {
+      "epoch": 0.7610802429254425,
+      "grad_norm": 0.5069129467010498,
+      "learning_rate": 4.2471366542385196e-05,
+      "loss": 0.9192,
+      "num_input_tokens_seen": 48250880,
+      "step": 5890
+    },
+    {
+      "epoch": 0.7623723995348236,
+      "grad_norm": 0.436614066362381,
+      "learning_rate": 4.2447154198774445e-05,
+      "loss": 0.999,
+      "num_input_tokens_seen": 48332800,
+      "step": 5900
+    },
+    {
+      "epoch": 0.7636645561442047,
+      "grad_norm": 0.21864083409309387,
+      "learning_rate": 4.242290990948821e-05,
+      "loss": 1.0004,
+      "num_input_tokens_seen": 48414720,
+      "step": 5910
+    },
+    {
+      "epoch": 0.7649567127535858,
+      "grad_norm": 0.5218377113342285,
+      "learning_rate": 4.2398633718917684e-05,
+      "loss": 0.8302,
+      "num_input_tokens_seen": 48496640,
+      "step": 5920
+    },
+    {
+      "epoch": 0.7662488693629668,
+      "grad_norm": 0.6044358611106873,
+      "learning_rate": 4.237432567151248e-05,
+      "loss": 1.2325,
+      "num_input_tokens_seen": 48578560,
+      "step": 5930
+    },
+    {
+      "epoch": 0.7675410259723479,
+      "grad_norm": 1.3460071086883545,
+      "learning_rate": 4.234998581178056e-05,
+      "loss": 0.636,
+      "num_input_tokens_seen": 48660480,
+      "step": 5940
+    },
+    {
+      "epoch": 0.768833182581729,
+      "grad_norm": 0.33106595277786255,
+      "learning_rate": 4.2325614184288096e-05,
+      "loss": 0.633,
+      "num_input_tokens_seen": 48742400,
+      "step": 5950
+    },
+    {
+      "epoch": 0.7701253391911099,
+      "grad_norm": 0.6000261306762695,
+      "learning_rate": 4.2301210833659464e-05,
+      "loss": 0.8583,
+      "num_input_tokens_seen": 48824320,
+      "step": 5960
+    },
+    {
+      "epoch": 0.771417495800491,
+      "grad_norm": 0.545335054397583,
+      "learning_rate": 4.227677580457711e-05,
+      "loss": 0.9888,
+      "num_input_tokens_seen": 48906240,
+      "step": 5970
+    },
+    {
+      "epoch": 0.7727096524098721,
+      "grad_norm": 0.5507743954658508,
+      "learning_rate": 4.2252309141781464e-05,
+      "loss": 1.0736,
+      "num_input_tokens_seen": 48988160,
+      "step": 5980
+    },
+    {
+      "epoch": 0.7740018090192531,
+      "grad_norm": 0.6265859007835388,
+      "learning_rate": 4.222781089007092e-05,
+      "loss": 0.995,
+      "num_input_tokens_seen": 49070080,
+      "step": 5990
+    },
+    {
+      "epoch": 0.7752939656286342,
+      "grad_norm": 1014203.875,
+      "learning_rate": 4.220328109430167e-05,
+      "loss": 0.8812,
+      "num_input_tokens_seen": 49152000,
+      "step": 6000
+    },
+    {
+      "epoch": 0.7765861222380153,
+      "grad_norm": 0.7153500318527222,
+      "learning_rate": 4.217871979938769e-05,
+      "loss": 1.3075,
+      "num_input_tokens_seen": 49233920,
+      "step": 6010
+    },
+    {
+      "epoch": 0.7778782788473964,
+      "grad_norm": 0.7981336712837219,
+      "learning_rate": 4.215412705030063e-05,
+      "loss": 0.8665,
+      "num_input_tokens_seen": 49315840,
+      "step": 6020
+    },
+    {
+      "epoch": 0.7791704354567773,
+      "grad_norm": 0.6426622867584229,
+      "learning_rate": 4.21295028920697e-05,
+      "loss": 0.852,
+      "num_input_tokens_seen": 49397760,
+      "step": 6030
+    },
+    {
+      "epoch": 0.7804625920661584,
+      "grad_norm": 0.6837995052337646,
+      "learning_rate": 4.210484736978166e-05,
+      "loss": 0.9422,
+      "num_input_tokens_seen": 49479680,
+      "step": 6040
+    },
+    {
+      "epoch": 0.7817547486755395,
+      "grad_norm": 0.4627484679222107,
+      "learning_rate": 4.208016052858067e-05,
+      "loss": 0.8523,
+      "num_input_tokens_seen": 49561600,
+      "step": 6050
+    },
+    {
+      "epoch": 0.7830469052849205,
+      "grad_norm": 0.29264596104621887,
+      "learning_rate": 4.2055442413668264e-05,
+      "loss": 0.7916,
+      "num_input_tokens_seen": 49643520,
+      "step": 6060
+    },
+    {
+      "epoch": 0.7843390618943016,
+      "grad_norm": 0.4580781161785126,
+      "learning_rate": 4.2030693070303204e-05,
+      "loss": 1.1109,
+      "num_input_tokens_seen": 49725440,
+      "step": 6070
+    },
+    {
+      "epoch": 0.7856312185036827,
+      "grad_norm": 0.24061749875545502,
+      "learning_rate": 4.2005912543801444e-05,
+      "loss": 0.7227,
+      "num_input_tokens_seen": 49807360,
+      "step": 6080
+    },
+    {
+      "epoch": 0.7869233751130637,
+      "grad_norm": 0.4084072709083557,
+      "learning_rate": 4.198110087953606e-05,
+      "loss": 0.7407,
+      "num_input_tokens_seen": 49889280,
+      "step": 6090
+    },
+    {
+      "epoch": 0.7882155317224447,
+      "grad_norm": 0.5031906962394714,
+      "learning_rate": 4.195625812293709e-05,
+      "loss": 1.4336,
+      "num_input_tokens_seen": 49971200,
+      "step": 6100
+    },
+    {
+      "epoch": 0.7895076883318258,
+      "grad_norm": 0.21128413081169128,
+      "learning_rate": 4.193138431949155e-05,
+      "loss": 0.9719,
+      "num_input_tokens_seen": 50053120,
+      "step": 6110
+    },
+    {
+      "epoch": 0.7907998449412069,
+      "grad_norm": 0.44993460178375244,
+      "learning_rate": 4.190647951474328e-05,
+      "loss": 1.0592,
+      "num_input_tokens_seen": 50135040,
+      "step": 6120
+    },
+    {
+      "epoch": 0.7920920015505879,
+      "grad_norm": 0.5219196677207947,
+      "learning_rate": 4.188154375429288e-05,
+      "loss": 0.8778,
+      "num_input_tokens_seen": 50216960,
+      "step": 6130
+    },
+    {
+      "epoch": 0.793384158159969,
+      "grad_norm": 0.5350174307823181,
+      "learning_rate": 4.1856577083797646e-05,
+      "loss": 1.329,
+      "num_input_tokens_seen": 50298880,
+      "step": 6140
+    },
+    {
+      "epoch": 0.7946763147693501,
+      "grad_norm": 0.6473702192306519,
+      "learning_rate": 4.183157954897144e-05,
+      "loss": 1.2239,
+      "num_input_tokens_seen": 50380800,
+      "step": 6150
+    },
+    {
+      "epoch": 0.7959684713787311,
+      "grad_norm": 0.5353970527648926,
+      "learning_rate": 4.1806551195584685e-05,
+      "loss": 1.0103,
+      "num_input_tokens_seen": 50462720,
+      "step": 6160
+    },
+    {
+      "epoch": 0.7972606279881121,
+      "grad_norm": 0.6963522434234619,
+      "learning_rate": 4.178149206946419e-05,
+      "loss": 0.8958,
+      "num_input_tokens_seen": 50544640,
+      "step": 6170
+    },
+    {
+      "epoch": 0.7985527845974932,
+      "grad_norm": 0.6284576654434204,
+      "learning_rate": 4.1756402216493115e-05,
+      "loss": 1.0327,
+      "num_input_tokens_seen": 50626560,
+      "step": 6180
+    },
+    {
+      "epoch": 0.7998449412068743,
+      "grad_norm": 0.49156904220581055,
+      "learning_rate": 4.17312816826109e-05,
+      "loss": 0.9924,
+      "num_input_tokens_seen": 50708480,
+      "step": 6190
+    },
+    {
+      "epoch": 0.8011370978162553,
+      "grad_norm": 0.8656590580940247,
+      "learning_rate": 4.1706130513813146e-05,
+      "loss": 0.7853,
+      "num_input_tokens_seen": 50790400,
+      "step": 6200
+    },
+    {
+      "epoch": 0.8024292544256364,
+      "grad_norm": 0.5040183067321777,
+      "learning_rate": 4.1680948756151564e-05,
+      "loss": 0.9263,
+      "num_input_tokens_seen": 50872320,
+      "step": 6210
+    },
+    {
+      "epoch": 0.8037214110350175,
+      "grad_norm": 0.28626561164855957,
+      "learning_rate": 4.165573645573384e-05,
+      "loss": 0.9224,
+      "num_input_tokens_seen": 50954240,
+      "step": 6220
+    },
+    {
+      "epoch": 0.8050135676443985,
+      "grad_norm": 0.5348039865493774,
+      "learning_rate": 4.1630493658723606e-05,
+      "loss": 1.2382,
+      "num_input_tokens_seen": 51036160,
+      "step": 6230
+    },
+    {
+      "epoch": 0.8063057242537796,
+      "grad_norm": 0.4930530786514282,
+      "learning_rate": 4.160522041134035e-05,
+      "loss": 1.0701,
+      "num_input_tokens_seen": 51118080,
+      "step": 6240
+    },
+    {
+      "epoch": 0.8075978808631606,
+      "grad_norm": 0.27241694927215576,
+      "learning_rate": 4.1579916759859286e-05,
+      "loss": 0.9246,
+      "num_input_tokens_seen": 51200000,
+      "step": 6250
+    },
+    {
+      "epoch": 0.8088900374725416,
+      "grad_norm": 0.8106761574745178,
+      "learning_rate": 4.155458275061129e-05,
+      "loss": 1.1799,
+      "num_input_tokens_seen": 51281920,
+      "step": 6260
+    },
+    {
+      "epoch": 0.8101821940819227,
+      "grad_norm": 0.6330926418304443,
+      "learning_rate": 4.152921842998287e-05,
+      "loss": 1.2025,
+      "num_input_tokens_seen": 51363840,
+      "step": 6270
+    },
+    {
+      "epoch": 0.8114743506913038,
+      "grad_norm": 0.5390750169754028,
+      "learning_rate": 4.150382384441598e-05,
+      "loss": 0.7594,
+      "num_input_tokens_seen": 51445760,
+      "step": 6280
+    },
+    {
+      "epoch": 0.8127665073006849,
+      "grad_norm": 0.645873486995697,
+      "learning_rate": 4.147839904040803e-05,
+      "loss": 0.8714,
+      "num_input_tokens_seen": 51527680,
+      "step": 6290
+    },
+    {
+      "epoch": 0.8140586639100659,
+      "grad_norm": 0.2540639638900757,
+      "learning_rate": 4.145294406451173e-05,
+      "loss": 1.2455,
+      "num_input_tokens_seen": 51609600,
+      "step": 6300
+    },
+    {
+      "epoch": 0.815350820519447,
+      "grad_norm": 0.75770103931427,
+      "learning_rate": 4.142745896333505e-05,
+      "loss": 1.1417,
+      "num_input_tokens_seen": 51691520,
+      "step": 6310
+    },
+    {
+      "epoch": 0.816642977128828,
+      "grad_norm": 0.505504310131073,
+      "learning_rate": 4.140194378354113e-05,
+      "loss": 1.1736,
+      "num_input_tokens_seen": 51773440,
+      "step": 6320
+    },
+    {
+      "epoch": 0.817935133738209,
+      "grad_norm": 0.47906693816185,
+      "learning_rate": 4.137639857184815e-05,
+      "loss": 0.8252,
+      "num_input_tokens_seen": 51855360,
+      "step": 6330
+    },
+    {
+      "epoch": 0.8192272903475901,
+      "grad_norm": 0.7896102666854858,
+      "learning_rate": 4.1350823375029326e-05,
+      "loss": 1.2724,
+      "num_input_tokens_seen": 51937280,
+      "step": 6340
+    },
+    {
+      "epoch": 0.8205194469569712,
+      "grad_norm": 0.6090170741081238,
+      "learning_rate": 4.132521823991272e-05,
+      "loss": 0.6208,
+      "num_input_tokens_seen": 52019200,
+      "step": 6350
+    },
+    {
+      "epoch": 0.8218116035663522,
+      "grad_norm": 0.4670509099960327,
+      "learning_rate": 4.129958321338127e-05,
+      "loss": 0.8883,
+      "num_input_tokens_seen": 52101120,
+      "step": 6360
+    },
+    {
+      "epoch": 0.8231037601757333,
+      "grad_norm": 0.7980589270591736,
+      "learning_rate": 4.127391834237258e-05,
+      "loss": 0.7734,
+      "num_input_tokens_seen": 52183040,
+      "step": 6370
+    },
+    {
+      "epoch": 0.8243959167851144,
+      "grad_norm": 0.47597405314445496,
+      "learning_rate": 4.124822367387897e-05,
+      "loss": 0.8842,
+      "num_input_tokens_seen": 52264960,
+      "step": 6380
+    },
+    {
+      "epoch": 0.8256880733944955,
+      "grad_norm": 0.28550368547439575,
+      "learning_rate": 4.122249925494726e-05,
+      "loss": 0.8806,
+      "num_input_tokens_seen": 52346880,
+      "step": 6390
+    },
+    {
+      "epoch": 0.8269802300038764,
+      "grad_norm": 0.5419972538948059,
+      "learning_rate": 4.119674513267878e-05,
+      "loss": 0.9348,
+      "num_input_tokens_seen": 52428800,
+      "step": 6400
+    },
+    {
+      "epoch": 0.8282723866132575,
+      "grad_norm": 0.5108693838119507,
+      "learning_rate": 4.117096135422923e-05,
+      "loss": 0.7648,
+      "num_input_tokens_seen": 52510720,
+      "step": 6410
+    },
+    {
+      "epoch": 0.8295645432226386,
+      "grad_norm": 0.22679108381271362,
+      "learning_rate": 4.114514796680862e-05,
+      "loss": 0.6926,
+      "num_input_tokens_seen": 52592640,
+      "step": 6420
+    },
+    {
+      "epoch": 0.8308566998320196,
+      "grad_norm": 0.5265064239501953,
+      "learning_rate": 4.111930501768116e-05,
+      "loss": 1.0296,
+      "num_input_tokens_seen": 52674560,
+      "step": 6430
+    },
+    {
+      "epoch": 0.8321488564414007,
+      "grad_norm": 0.41228848695755005,
+      "learning_rate": 4.1093432554165196e-05,
+      "loss": 0.9102,
+      "num_input_tokens_seen": 52756480,
+      "step": 6440
+    },
+    {
+      "epoch": 0.8334410130507818,
+      "grad_norm": 0.35673919320106506,
+      "learning_rate": 4.106753062363311e-05,
+      "loss": 0.8159,
+      "num_input_tokens_seen": 52838400,
+      "step": 6450
+    },
+    {
+      "epoch": 0.8347331696601629,
+      "grad_norm": 0.5186113715171814,
+      "learning_rate": 4.104159927351125e-05,
+      "loss": 1.1631,
+      "num_input_tokens_seen": 52920320,
+      "step": 6460
+    },
+    {
+      "epoch": 0.8360253262695438,
+      "grad_norm": 0.45875003933906555,
+      "learning_rate": 4.1015638551279825e-05,
+      "loss": 0.8534,
+      "num_input_tokens_seen": 53002240,
+      "step": 6470
+    },
+    {
+      "epoch": 0.8373174828789249,
+      "grad_norm": 0.5119809508323669,
+      "learning_rate": 4.098964850447281e-05,
+      "loss": 1.0583,
+      "num_input_tokens_seen": 53084160,
+      "step": 6480
+    },
+    {
+      "epoch": 0.838609639488306,
+      "grad_norm": 0.35101157426834106,
+      "learning_rate": 4.0963629180677896e-05,
+      "loss": 0.8959,
+      "num_input_tokens_seen": 53166080,
+      "step": 6490
+    },
+    {
+      "epoch": 0.839901796097687,
+      "grad_norm": 0.5710439085960388,
+      "learning_rate": 4.093758062753638e-05,
+      "loss": 1.1624,
+      "num_input_tokens_seen": 53248000,
+      "step": 6500
+    },
+    {
+      "epoch": 0.8411939527070681,
+      "grad_norm": 0.34572339057922363,
+      "learning_rate": 4.0911502892743035e-05,
+      "loss": 1.2004,
+      "num_input_tokens_seen": 53329920,
+      "step": 6510
+    },
+    {
+      "epoch": 0.8424861093164492,
+      "grad_norm": 0.5202426910400391,
+      "learning_rate": 4.088539602404613e-05,
+      "loss": 1.2094,
+      "num_input_tokens_seen": 53411840,
+      "step": 6520
+    },
+    {
+      "epoch": 0.8437782659258302,
+      "grad_norm": 0.46138355135917664,
+      "learning_rate": 4.085926006924723e-05,
+      "loss": 1.0853,
+      "num_input_tokens_seen": 53493760,
+      "step": 6530
+    },
+    {
+      "epoch": 0.8450704225352113,
+      "grad_norm": 0.5127395391464233,
+      "learning_rate": 4.083309507620118e-05,
+      "loss": 0.8684,
+      "num_input_tokens_seen": 53575680,
+      "step": 6540
+    },
+    {
+      "epoch": 0.8463625791445923,
+      "grad_norm": 0.5306764245033264,
+      "learning_rate": 4.080690109281597e-05,
+      "loss": 1.1792,
+      "num_input_tokens_seen": 53657600,
+      "step": 6550
+    },
+    {
+      "epoch": 0.8476547357539734,
+      "grad_norm": 0.48974373936653137,
+      "learning_rate": 4.078067816705272e-05,
+      "loss": 0.8527,
+      "num_input_tokens_seen": 53739520,
+      "step": 6560
+    },
+    {
+      "epoch": 0.8489468923633544,
+      "grad_norm": 0.4819824993610382,
+      "learning_rate": 4.075442634692548e-05,
+      "loss": 0.9558,
+      "num_input_tokens_seen": 53821440,
+      "step": 6570
+    },
+    {
+      "epoch": 0.8502390489727355,
+      "grad_norm": 0.5945402979850769,
+      "learning_rate": 4.072814568050125e-05,
+      "loss": 1.0556,
+      "num_input_tokens_seen": 53903360,
+      "step": 6580
+    },
+    {
+      "epoch": 0.8515312055821166,
+      "grad_norm": 0.26948097348213196,
+      "learning_rate": 4.070183621589983e-05,
+      "loss": 0.9564,
+      "num_input_tokens_seen": 53985280,
+      "step": 6590
+    },
+    {
+      "epoch": 0.8528233621914976,
+      "grad_norm": 0.5950406789779663,
+      "learning_rate": 4.067549800129375e-05,
+      "loss": 1.2202,
+      "num_input_tokens_seen": 54067200,
+      "step": 6600
+    },
+    {
+      "epoch": 0.8541155188008787,
+      "grad_norm": 1.1179322004318237,
+      "learning_rate": 4.06491310849082e-05,
+      "loss": 0.9542,
+      "num_input_tokens_seen": 54149120,
+      "step": 6610
+    },
+    {
+      "epoch": 0.8554076754102597,
+      "grad_norm": 0.5142911076545715,
+      "learning_rate": 4.0622735515020896e-05,
+      "loss": 1.0358,
+      "num_input_tokens_seen": 54231040,
+      "step": 6620
+    },
+    {
+      "epoch": 0.8566998320196407,
+      "grad_norm": 0.6860795021057129,
+      "learning_rate": 4.059631133996203e-05,
+      "loss": 1.2331,
+      "num_input_tokens_seen": 54312960,
+      "step": 6630
+    },
+    {
+      "epoch": 0.8579919886290218,
+      "grad_norm": 0.47902247309684753,
+      "learning_rate": 4.0569858608114177e-05,
+      "loss": 0.9423,
+      "num_input_tokens_seen": 54394880,
+      "step": 6640
+    },
+    {
+      "epoch": 0.8592841452384029,
+      "grad_norm": 0.46857351064682007,
+      "learning_rate": 4.054337736791218e-05,
+      "loss": 1.1609,
+      "num_input_tokens_seen": 54476800,
+      "step": 6650
+    },
+    {
+      "epoch": 0.860576301847784,
+      "grad_norm": 0.5907356142997742,
+      "learning_rate": 4.05168676678431e-05,
+      "loss": 0.9658,
+      "num_input_tokens_seen": 54558720,
+      "step": 6660
+    },
+    {
+      "epoch": 0.861868458457165,
+      "grad_norm": 0.26403316855430603,
+      "learning_rate": 4.04903295564461e-05,
+      "loss": 0.7992,
+      "num_input_tokens_seen": 54640640,
+      "step": 6670
+    },
+    {
+      "epoch": 0.8631606150665461,
+      "grad_norm": 0.6622225642204285,
+      "learning_rate": 4.046376308231237e-05,
+      "loss": 0.9918,
+      "num_input_tokens_seen": 54722560,
+      "step": 6680
+    },
+    {
+      "epoch": 0.8644527716759272,
+      "grad_norm": 0.16060075163841248,
+      "learning_rate": 4.0437168294085013e-05,
+      "loss": 0.9523,
+      "num_input_tokens_seen": 54804480,
+      "step": 6690
+    },
+    {
+      "epoch": 0.8657449282853081,
+      "grad_norm": 0.5184667706489563,
+      "learning_rate": 4.0410545240459005e-05,
+      "loss": 1.1628,
+      "num_input_tokens_seen": 54886400,
+      "step": 6700
+    },
+    {
+      "epoch": 0.8670370848946892,
+      "grad_norm": 0.4965847134590149,
+      "learning_rate": 4.0383893970181054e-05,
+      "loss": 1.0468,
+      "num_input_tokens_seen": 54968320,
+      "step": 6710
+    },
+    {
+      "epoch": 0.8683292415040703,
+      "grad_norm": 6.185298442840576,
+      "learning_rate": 4.0357214532049535e-05,
+      "loss": 1.2028,
+      "num_input_tokens_seen": 55050240,
+      "step": 6720
+    },
+    {
+      "epoch": 0.8696213981134514,
+      "grad_norm": 0.43425482511520386,
+      "learning_rate": 4.03305069749144e-05,
+      "loss": 0.503,
+      "num_input_tokens_seen": 55132160,
+      "step": 6730
+    },
+    {
+      "epoch": 0.8709135547228324,
+      "grad_norm": 0.2627141773700714,
+      "learning_rate": 4.03037713476771e-05,
+      "loss": 1.0739,
+      "num_input_tokens_seen": 55214080,
+      "step": 6740
+    },
+    {
+      "epoch": 0.8722057113322135,
+      "grad_norm": 0.5449861288070679,
+      "learning_rate": 4.027700769929046e-05,
+      "loss": 0.7428,
+      "num_input_tokens_seen": 55296000,
+      "step": 6750
+    },
+    {
+      "epoch": 0.8734978679415946,
+      "grad_norm": 0.5020452737808228,
+      "learning_rate": 4.025021607875862e-05,
+      "loss": 1.0242,
+      "num_input_tokens_seen": 55377920,
+      "step": 6760
+    },
+    {
+      "epoch": 0.8747900245509755,
+      "grad_norm": 0.5593128800392151,
+      "learning_rate": 4.0223396535136945e-05,
+      "loss": 1.2703,
+      "num_input_tokens_seen": 55459840,
+      "step": 6770
+    },
+    {
+      "epoch": 0.8760821811603566,
+      "grad_norm": 0.4105585813522339,
+      "learning_rate": 4.019654911753193e-05,
+      "loss": 0.8773,
+      "num_input_tokens_seen": 55541760,
+      "step": 6780
+    },
+    {
+      "epoch": 0.8773743377697377,
+      "grad_norm": 0.37367942929267883,
+      "learning_rate": 4.016967387510108e-05,
+      "loss": 1.094,
+      "num_input_tokens_seen": 55623680,
+      "step": 6790
+    },
+    {
+      "epoch": 0.8786664943791187,
+      "grad_norm": 0.5563257932662964,
+      "learning_rate": 4.014277085705288e-05,
+      "loss": 0.9265,
+      "num_input_tokens_seen": 55705600,
+      "step": 6800
+    },
+    {
+      "epoch": 0.8799586509884998,
+      "grad_norm": 0.7399642467498779,
+      "learning_rate": 4.011584011264665e-05,
+      "loss": 0.9153,
+      "num_input_tokens_seen": 55787520,
+      "step": 6810
+    },
+    {
+      "epoch": 0.8812508075978809,
+      "grad_norm": 0.6172972321510315,
+      "learning_rate": 4.0088881691192474e-05,
+      "loss": 0.7931,
+      "num_input_tokens_seen": 55869440,
+      "step": 6820
+    },
+    {
+      "epoch": 0.882542964207262,
+      "grad_norm": 0.4733666181564331,
+      "learning_rate": 4.006189564205115e-05,
+      "loss": 0.9963,
+      "num_input_tokens_seen": 55951360,
+      "step": 6830
+    },
+    {
+      "epoch": 0.883835120816643,
+      "grad_norm": 0.6193355321884155,
+      "learning_rate": 4.0034882014634015e-05,
+      "loss": 0.7636,
+      "num_input_tokens_seen": 56033280,
+      "step": 6840
+    },
+    {
+      "epoch": 0.885127277426024,
+      "grad_norm": 0.4271906018257141,
+      "learning_rate": 4.000784085840293e-05,
+      "loss": 0.7192,
+      "num_input_tokens_seen": 56115200,
+      "step": 6850
+    },
+    {
+      "epoch": 0.8864194340354051,
+      "grad_norm": 0.5514196157455444,
+      "learning_rate": 3.9980772222870156e-05,
+      "loss": 0.9082,
+      "num_input_tokens_seen": 56197120,
+      "step": 6860
+    },
+    {
+      "epoch": 0.8877115906447861,
+      "grad_norm": 0.5211493968963623,
+      "learning_rate": 3.995367615759825e-05,
+      "loss": 1.2218,
+      "num_input_tokens_seen": 56279040,
+      "step": 6870
+    },
+    {
+      "epoch": 0.8890037472541672,
+      "grad_norm": 0.5780483484268188,
+      "learning_rate": 3.992655271220003e-05,
+      "loss": 1.0894,
+      "num_input_tokens_seen": 56360960,
+      "step": 6880
+    },
+    {
+      "epoch": 0.8902959038635483,
+      "grad_norm": 0.537337601184845,
+      "learning_rate": 3.98994019363384e-05,
+      "loss": 1.0208,
+      "num_input_tokens_seen": 56442880,
+      "step": 6890
+    },
+    {
+      "epoch": 0.8915880604729293,
+      "grad_norm": 0.4431285858154297,
+      "learning_rate": 3.9872223879726356e-05,
+      "loss": 0.7955,
+      "num_input_tokens_seen": 56524800,
+      "step": 6900
+    },
+    {
+      "epoch": 0.8928802170823104,
+      "grad_norm": 0.44692462682724,
+      "learning_rate": 3.98450185921268e-05,
+      "loss": 1.262,
+      "num_input_tokens_seen": 56606720,
+      "step": 6910
+    },
+    {
+      "epoch": 0.8941723736916914,
+      "grad_norm": 0.5244356989860535,
+      "learning_rate": 3.981778612335253e-05,
+      "loss": 1.1836,
+      "num_input_tokens_seen": 56688640,
+      "step": 6920
+    },
+    {
+      "epoch": 0.8954645303010725,
+      "grad_norm": 0.5537554025650024,
+      "learning_rate": 3.979052652326609e-05,
+      "loss": 0.7662,
+      "num_input_tokens_seen": 56770560,
+      "step": 6930
+    },
+    {
+      "epoch": 0.8967566869104535,
+      "grad_norm": 0.5707337856292725,
+      "learning_rate": 3.976323984177971e-05,
+      "loss": 0.7414,
+      "num_input_tokens_seen": 56852480,
+      "step": 6940
+    },
+    {
+      "epoch": 0.8980488435198346,
+      "grad_norm": 0.4836041331291199,
+      "learning_rate": 3.97359261288552e-05,
+      "loss": 1.0458,
+      "num_input_tokens_seen": 56934400,
+      "step": 6950
+    },
+    {
+      "epoch": 0.8993410001292157,
+      "grad_norm": 0.47009721398353577,
+      "learning_rate": 3.970858543450387e-05,
+      "loss": 1.0642,
+      "num_input_tokens_seen": 57016320,
+      "step": 6960
+    },
+    {
+      "epoch": 0.9006331567385967,
+      "grad_norm": 0.6852318644523621,
+      "learning_rate": 3.968121780878643e-05,
+      "loss": 1.3093,
+      "num_input_tokens_seen": 57098240,
+      "step": 6970
+    },
+    {
+      "epoch": 0.9019253133479778,
+      "grad_norm": 0.24057495594024658,
+      "learning_rate": 3.965382330181291e-05,
+      "loss": 0.6589,
+      "num_input_tokens_seen": 57180160,
+      "step": 6980
+    },
+    {
+      "epoch": 0.9032174699573589,
+      "grad_norm": 0.33139023184776306,
+      "learning_rate": 3.962640196374254e-05,
+      "loss": 0.6929,
+      "num_input_tokens_seen": 57262080,
+      "step": 6990
+    },
+    {
+      "epoch": 0.9045096265667399,
+      "grad_norm": 0.5860159993171692,
+      "learning_rate": 3.9598953844783705e-05,
+      "loss": 1.1313,
+      "num_input_tokens_seen": 57344000,
+      "step": 7000
+    },
+    {
+      "epoch": 0.9058017831761209,
+      "grad_norm": 0.36473211646080017,
+      "learning_rate": 3.957147899519379e-05,
+      "loss": 0.8748,
+      "num_input_tokens_seen": 57425920,
+      "step": 7010
+    },
+    {
+      "epoch": 0.907093939785502,
+      "grad_norm": 0.547675371170044,
+      "learning_rate": 3.954397746527916e-05,
+      "loss": 0.3491,
+      "num_input_tokens_seen": 57507840,
+      "step": 7020
+    },
+    {
+      "epoch": 0.9083860963948831,
+      "grad_norm": 0.3887852430343628,
+      "learning_rate": 3.951644930539502e-05,
+      "loss": 0.7414,
+      "num_input_tokens_seen": 57589760,
+      "step": 7030
+    },
+    {
+      "epoch": 0.9096782530042641,
+      "grad_norm": 0.5362502932548523,
+      "learning_rate": 3.9488894565945305e-05,
+      "loss": 0.8839,
+      "num_input_tokens_seen": 57671680,
+      "step": 7040
+    },
+    {
+      "epoch": 0.9109704096136452,
+      "grad_norm": 0.578965425491333,
+      "learning_rate": 3.9461313297382666e-05,
+      "loss": 0.5389,
+      "num_input_tokens_seen": 57753600,
+      "step": 7050
+    },
+    {
+      "epoch": 0.9122625662230263,
+      "grad_norm": 0.6477794051170349,
+      "learning_rate": 3.94337055502083e-05,
+      "loss": 1.0182,
+      "num_input_tokens_seen": 57835520,
+      "step": 7060
+    },
+    {
+      "epoch": 0.9135547228324072,
+      "grad_norm": 0.505005955696106,
+      "learning_rate": 3.9406071374971887e-05,
+      "loss": 1.0376,
+      "num_input_tokens_seen": 57917440,
+      "step": 7070
+    },
+    {
+      "epoch": 0.9148468794417883,
+      "grad_norm": 1.0744434595108032,
+      "learning_rate": 3.93784108222715e-05,
+      "loss": 0.6095,
+      "num_input_tokens_seen": 57999360,
+      "step": 7080
+    },
+    {
+      "epoch": 0.9161390360511694,
+      "grad_norm": 0.518518328666687,
+      "learning_rate": 3.935072394275352e-05,
+      "loss": 1.0714,
+      "num_input_tokens_seen": 58081280,
+      "step": 7090
+    },
+    {
+      "epoch": 0.9174311926605505,
+      "grad_norm": 0.5842453241348267,
+      "learning_rate": 3.9323010787112505e-05,
+      "loss": 1.116,
+      "num_input_tokens_seen": 58163200,
+      "step": 7100
+    },
+    {
+      "epoch": 0.9187233492699315,
+      "grad_norm": 0.43083658814430237,
+      "learning_rate": 3.929527140609115e-05,
+      "loss": 1.1167,
+      "num_input_tokens_seen": 58245120,
+      "step": 7110
+    },
+    {
+      "epoch": 0.9200155058793126,
+      "grad_norm": 2.2373194694519043,
+      "learning_rate": 3.926750585048016e-05,
+      "loss": 0.7755,
+      "num_input_tokens_seen": 58327040,
+      "step": 7120
+    },
+    {
+      "epoch": 0.9213076624886937,
+      "grad_norm": 0.5515663623809814,
+      "learning_rate": 3.9239714171118167e-05,
+      "loss": 0.7525,
+      "num_input_tokens_seen": 58408960,
+      "step": 7130
+    },
+    {
+      "epoch": 0.9225998190980746,
+      "grad_norm": 0.5348473191261292,
+      "learning_rate": 3.921189641889163e-05,
+      "loss": 0.9602,
+      "num_input_tokens_seen": 58490880,
+      "step": 7140
+    },
+    {
+      "epoch": 0.9238919757074557,
+      "grad_norm": 0.4317019581794739,
+      "learning_rate": 3.918405264473476e-05,
+      "loss": 0.7652,
+      "num_input_tokens_seen": 58572800,
+      "step": 7150
+    },
+    {
+      "epoch": 0.9251841323168368,
+      "grad_norm": 0.48419129848480225,
+      "learning_rate": 3.9156182899629404e-05,
+      "loss": 0.5639,
+      "num_input_tokens_seen": 58654720,
+      "step": 7160
+    },
+    {
+      "epoch": 0.9264762889262178,
+      "grad_norm": 0.33040711283683777,
+      "learning_rate": 3.912828723460495e-05,
+      "loss": 0.8571,
+      "num_input_tokens_seen": 58736640,
+      "step": 7170
+    },
+    {
+      "epoch": 0.9277684455355989,
+      "grad_norm": 0.46214792132377625,
+      "learning_rate": 3.9100365700738275e-05,
+      "loss": 0.747,
+      "num_input_tokens_seen": 58818560,
+      "step": 7180
+    },
+    {
+      "epoch": 0.92906060214498,
+      "grad_norm": 0.4923725426197052,
+      "learning_rate": 3.907241834915359e-05,
+      "loss": 0.8309,
+      "num_input_tokens_seen": 58900480,
+      "step": 7190
+    },
+    {
+      "epoch": 0.9303527587543611,
+      "grad_norm": 0.5313102006912231,
+      "learning_rate": 3.904444523102242e-05,
+      "loss": 0.9932,
+      "num_input_tokens_seen": 58982400,
+      "step": 7200
+    },
+    {
+      "epoch": 0.931644915363742,
+      "grad_norm": 0.5704349875450134,
+      "learning_rate": 3.901644639756342e-05,
+      "loss": 0.9913,
+      "num_input_tokens_seen": 59064320,
+      "step": 7210
+    },
+    {
+      "epoch": 0.9329370719731231,
+      "grad_norm": 0.48948806524276733,
+      "learning_rate": 3.898842190004235e-05,
+      "loss": 0.9762,
+      "num_input_tokens_seen": 59146240,
+      "step": 7220
+    },
+    {
+      "epoch": 0.9342292285825042,
+      "grad_norm": 0.6252711415290833,
+      "learning_rate": 3.896037178977196e-05,
+      "loss": 0.6812,
+      "num_input_tokens_seen": 59228160,
+      "step": 7230
+    },
+    {
+      "epoch": 0.9355213851918852,
+      "grad_norm": 0.7012498378753662,
+      "learning_rate": 3.893229611811192e-05,
+      "loss": 1.3312,
+      "num_input_tokens_seen": 59310080,
+      "step": 7240
+    },
+    {
+      "epoch": 0.9368135418012663,
+      "grad_norm": 0.5096800923347473,
+      "learning_rate": 3.8904194936468665e-05,
+      "loss": 1.0935,
+      "num_input_tokens_seen": 59392000,
+      "step": 7250
+    },
+    {
+      "epoch": 0.9381056984106474,
+      "grad_norm": 0.47949451208114624,
+      "learning_rate": 3.887606829629536e-05,
+      "loss": 0.7918,
+      "num_input_tokens_seen": 59473920,
+      "step": 7260
+    },
+    {
+      "epoch": 0.9393978550200285,
+      "grad_norm": 0.256607323884964,
+      "learning_rate": 3.884791624909178e-05,
+      "loss": 0.8013,
+      "num_input_tokens_seen": 59555840,
+      "step": 7270
+    },
+    {
+      "epoch": 0.9406900116294095,
+      "grad_norm": 0.45952385663986206,
+      "learning_rate": 3.881973884640422e-05,
+      "loss": 1.0558,
+      "num_input_tokens_seen": 59637760,
+      "step": 7280
+    },
+    {
+      "epoch": 0.9419821682387906,
+      "grad_norm": 0.8293887376785278,
+      "learning_rate": 3.87915361398254e-05,
+      "loss": 1.1583,
+      "num_input_tokens_seen": 59719680,
+      "step": 7290
+    },
+    {
+      "epoch": 0.9432743248481716,
+      "grad_norm": 0.6213748455047607,
+      "learning_rate": 3.8763308180994384e-05,
+      "loss": 0.4953,
+      "num_input_tokens_seen": 59801600,
+      "step": 7300
+    },
+    {
+      "epoch": 0.9445664814575526,
+      "grad_norm": 0.7034590244293213,
+      "learning_rate": 3.873505502159645e-05,
+      "loss": 1.1762,
+      "num_input_tokens_seen": 59883520,
+      "step": 7310
+    },
+    {
+      "epoch": 0.9458586380669337,
+      "grad_norm": 0.4893646240234375,
+      "learning_rate": 3.8706776713363025e-05,
+      "loss": 0.7593,
+      "num_input_tokens_seen": 59965440,
+      "step": 7320
+    },
+    {
+      "epoch": 0.9471507946763148,
+      "grad_norm": 0.7708596587181091,
+      "learning_rate": 3.86784733080716e-05,
+      "loss": 0.6101,
+      "num_input_tokens_seen": 60047360,
+      "step": 7330
+    },
+    {
+      "epoch": 0.9484429512856958,
+      "grad_norm": 0.541378915309906,
+      "learning_rate": 3.86501448575456e-05,
+      "loss": 1.1621,
+      "num_input_tokens_seen": 60129280,
+      "step": 7340
+    },
+    {
+      "epoch": 0.9497351078950769,
+      "grad_norm": 0.46613985300064087,
+      "learning_rate": 3.862179141365431e-05,
+      "loss": 0.6934,
+      "num_input_tokens_seen": 60211200,
+      "step": 7350
+    },
+    {
+      "epoch": 0.951027264504458,
+      "grad_norm": 0.37554696202278137,
+      "learning_rate": 3.859341302831279e-05,
+      "loss": 1.064,
+      "num_input_tokens_seen": 60293120,
+      "step": 7360
+    },
+    {
+      "epoch": 0.952319421113839,
+      "grad_norm": 0.7480294108390808,
+      "learning_rate": 3.856500975348176e-05,
+      "loss": 0.9418,
+      "num_input_tokens_seen": 60375040,
+      "step": 7370
+    },
+    {
+      "epoch": 0.95361157772322,
+      "grad_norm": 0.5717000961303711,
+      "learning_rate": 3.8536581641167506e-05,
+      "loss": 0.8951,
+      "num_input_tokens_seen": 60456960,
+      "step": 7380
+    },
+    {
+      "epoch": 0.9549037343326011,
+      "grad_norm": 0.45350784063339233,
+      "learning_rate": 3.85081287434218e-05,
+      "loss": 0.9263,
+      "num_input_tokens_seen": 60538880,
+      "step": 7390
+    },
+    {
+      "epoch": 0.9561958909419822,
+      "grad_norm": 0.839464545249939,
+      "learning_rate": 3.84796511123418e-05,
+      "loss": 1.1756,
+      "num_input_tokens_seen": 60620800,
+      "step": 7400
+    },
+    {
+      "epoch": 0.9574880475513632,
+      "grad_norm": 0.5966989994049072,
+      "learning_rate": 3.845114880006994e-05,
+      "loss": 0.9992,
+      "num_input_tokens_seen": 60702720,
+      "step": 7410
+    },
+    {
+      "epoch": 0.9587802041607443,
+      "grad_norm": 0.5134586095809937,
+      "learning_rate": 3.842262185879384e-05,
+      "loss": 1.1923,
+      "num_input_tokens_seen": 60784640,
+      "step": 7420
+    },
+    {
+      "epoch": 0.9600723607701254,
+      "grad_norm": 0.2532576024532318,
+      "learning_rate": 3.8394070340746234e-05,
+      "loss": 0.828,
+      "num_input_tokens_seen": 60866560,
+      "step": 7430
+    },
+    {
+      "epoch": 0.9613645173795063,
+      "grad_norm": 0.4352121949195862,
+      "learning_rate": 3.836549429820485e-05,
+      "loss": 0.9765,
+      "num_input_tokens_seen": 60948480,
+      "step": 7440
+    },
+    {
+      "epoch": 0.9626566739888874,
+      "grad_norm": 0.5217087864875793,
+      "learning_rate": 3.833689378349231e-05,
+      "loss": 0.9062,
+      "num_input_tokens_seen": 61030400,
+      "step": 7450
+    },
+    {
+      "epoch": 0.9639488305982685,
+      "grad_norm": 0.5556167960166931,
+      "learning_rate": 3.830826884897606e-05,
+      "loss": 1.0222,
+      "num_input_tokens_seen": 61112320,
+      "step": 7460
+    },
+    {
+      "epoch": 0.9652409872076496,
+      "grad_norm": 0.4434933662414551,
+      "learning_rate": 3.827961954706825e-05,
+      "loss": 1.118,
+      "num_input_tokens_seen": 61194240,
+      "step": 7470
+    },
+    {
+      "epoch": 0.9665331438170306,
+      "grad_norm": 0.8783986568450928,
+      "learning_rate": 3.825094593022563e-05,
+      "loss": 1.078,
+      "num_input_tokens_seen": 61276160,
+      "step": 7480
+    },
+    {
+      "epoch": 0.9678253004264117,
+      "grad_norm": 0.26476526260375977,
+      "learning_rate": 3.8222248050949505e-05,
+      "loss": 0.665,
+      "num_input_tokens_seen": 61358080,
+      "step": 7490
+    },
+    {
+      "epoch": 0.9691174570357928,
+      "grad_norm": 0.5986683368682861,
+      "learning_rate": 3.8193525961785584e-05,
+      "loss": 0.9969,
+      "num_input_tokens_seen": 61440000,
+      "step": 7500
+    },
+    {
+      "epoch": 0.9704096136451738,
+      "grad_norm": 0.2270795851945877,
+      "learning_rate": 3.8164779715323905e-05,
+      "loss": 0.9114,
+      "num_input_tokens_seen": 61521920,
+      "step": 7510
+    },
+    {
+      "epoch": 0.9717017702545548,
+      "grad_norm": 0.780129075050354,
+      "learning_rate": 3.813600936419874e-05,
+      "loss": 0.7856,
+      "num_input_tokens_seen": 61603840,
+      "step": 7520
+    },
+    {
+      "epoch": 0.9729939268639359,
+      "grad_norm": 0.5222942233085632,
+      "learning_rate": 3.81072149610885e-05,
+      "loss": 1.0216,
+      "num_input_tokens_seen": 61685760,
+      "step": 7530
+    },
+    {
+      "epoch": 0.9742860834733169,
+      "grad_norm": 0.5253265500068665,
+      "learning_rate": 3.807839655871563e-05,
+      "loss": 0.7388,
+      "num_input_tokens_seen": 61767680,
+      "step": 7540
+    },
+    {
+      "epoch": 0.975578240082698,
+      "grad_norm": 0.46493276953697205,
+      "learning_rate": 3.8049554209846514e-05,
+      "loss": 0.6208,
+      "num_input_tokens_seen": 61849600,
+      "step": 7550
+    },
+    {
+      "epoch": 0.9768703966920791,
+      "grad_norm": 0.7733160853385925,
+      "learning_rate": 3.802068796729139e-05,
+      "loss": 1.1114,
+      "num_input_tokens_seen": 61931520,
+      "step": 7560
+    },
+    {
+      "epoch": 0.9781625533014602,
+      "grad_norm": 0.21577703952789307,
+      "learning_rate": 3.7991797883904254e-05,
+      "loss": 0.9243,
+      "num_input_tokens_seen": 62013440,
+      "step": 7570
+    },
+    {
+      "epoch": 0.9794547099108412,
+      "grad_norm": 0.6388958692550659,
+      "learning_rate": 3.796288401258272e-05,
+      "loss": 0.8196,
+      "num_input_tokens_seen": 62095360,
+      "step": 7580
+    },
+    {
+      "epoch": 0.9807468665202222,
+      "grad_norm": 0.2966822683811188,
+      "learning_rate": 3.7933946406268e-05,
+      "loss": 0.8933,
+      "num_input_tokens_seen": 62177280,
+      "step": 7590
+    },
+    {
+      "epoch": 0.9820390231296033,
+      "grad_norm": 0.5425664186477661,
+      "learning_rate": 3.790498511794473e-05,
+      "loss": 0.9035,
+      "num_input_tokens_seen": 62259200,
+      "step": 7600
+    },
+    {
+      "epoch": 0.9833311797389843,
+      "grad_norm": 0.47928035259246826,
+      "learning_rate": 3.787600020064095e-05,
+      "loss": 0.9369,
+      "num_input_tokens_seen": 62341120,
+      "step": 7610
+    },
+    {
+      "epoch": 0.9846233363483654,
+      "grad_norm": 0.4411972463130951,
+      "learning_rate": 3.7846991707427905e-05,
+      "loss": 0.7782,
+      "num_input_tokens_seen": 62423040,
+      "step": 7620
+    },
+    {
+      "epoch": 0.9859154929577465,
+      "grad_norm": 0.5929402112960815,
+      "learning_rate": 3.7817959691420056e-05,
+      "loss": 0.8775,
+      "num_input_tokens_seen": 62504960,
+      "step": 7630
+    },
+    {
+      "epoch": 0.9872076495671276,
+      "grad_norm": 0.5183133482933044,
+      "learning_rate": 3.778890420577492e-05,
+      "loss": 0.7959,
+      "num_input_tokens_seen": 62586880,
+      "step": 7640
+    },
+    {
+      "epoch": 0.9884998061765086,
+      "grad_norm": 0.3291929364204407,
+      "learning_rate": 3.775982530369298e-05,
+      "loss": 1.0962,
+      "num_input_tokens_seen": 62668800,
+      "step": 7650
+    },
+    {
+      "epoch": 0.9897919627858897,
+      "grad_norm": 0.5607266426086426,
+      "learning_rate": 3.77307230384176e-05,
+      "loss": 1.1062,
+      "num_input_tokens_seen": 62750720,
+      "step": 7660
+    },
+    {
+      "epoch": 0.9910841193952707,
+      "grad_norm": 0.6233817338943481,
+      "learning_rate": 3.7701597463234916e-05,
+      "loss": 0.6531,
+      "num_input_tokens_seen": 62832640,
+      "step": 7670
+    },
+    {
+      "epoch": 0.9923762760046517,
+      "grad_norm": 0.5039672255516052,
+      "learning_rate": 3.767244863147377e-05,
+      "loss": 0.8184,
+      "num_input_tokens_seen": 62914560,
+      "step": 7680
+    },
+    {
+      "epoch": 0.9936684326140328,
+      "grad_norm": 0.5114099979400635,
+      "learning_rate": 3.764327659650553e-05,
+      "loss": 1.1191,
+      "num_input_tokens_seen": 62996480,
+      "step": 7690
+    },
+    {
+      "epoch": 0.9949605892234139,
+      "grad_norm": 0.6352154612541199,
+      "learning_rate": 3.7614081411744116e-05,
+      "loss": 1.1411,
+      "num_input_tokens_seen": 63078400,
+      "step": 7700
+    },
+    {
+      "epoch": 0.9962527458327949,
+      "grad_norm": 0.5263626575469971,
+      "learning_rate": 3.75848631306458e-05,
+      "loss": 1.0581,
+      "num_input_tokens_seen": 63160320,
+      "step": 7710
+    },
+    {
+      "epoch": 0.997544902442176,
+      "grad_norm": 0.5167150497436523,
+      "learning_rate": 3.755562180670914e-05,
+      "loss": 0.8535,
+      "num_input_tokens_seen": 63242240,
+      "step": 7720
+    },
+    {
+      "epoch": 0.9988370590515571,
+      "grad_norm": 0.5089849829673767,
+      "learning_rate": 3.75263574934749e-05,
+      "loss": 0.7162,
+      "num_input_tokens_seen": 63324160,
+      "step": 7730
+    },
+    {
+      "epoch": 1.000129215660938,
+      "grad_norm": 0.412061870098114,
+      "learning_rate": 3.7497070244525925e-05,
+      "loss": 0.6882,
+      "num_input_tokens_seen": 63406080,
+      "step": 7740
+    },
+    {
+      "epoch": 1.0014213722703191,
+      "grad_norm": 0.6777093410491943,
+      "learning_rate": 3.746776011348706e-05,
+      "loss": 1.1799,
+      "num_input_tokens_seen": 63488000,
+      "step": 7750
+    },
+    {
+      "epoch": 1.0027135288797002,
+      "grad_norm": 0.5311883091926575,
+      "learning_rate": 3.7438427154025045e-05,
+      "loss": 0.5552,
+      "num_input_tokens_seen": 63569920,
+      "step": 7760
+    },
+    {
+      "epoch": 1.0040056854890813,
+      "grad_norm": 0.5286181569099426,
+      "learning_rate": 3.7409071419848436e-05,
+      "loss": 1.1324,
+      "num_input_tokens_seen": 63651840,
+      "step": 7770
+    },
+    {
+      "epoch": 1.0052978420984624,
+      "grad_norm": 0.5085585117340088,
+      "learning_rate": 3.7379692964707456e-05,
+      "loss": 0.9001,
+      "num_input_tokens_seen": 63733760,
+      "step": 7780
+    },
+    {
+      "epoch": 1.0065899987078435,
+      "grad_norm": 0.3858306109905243,
+      "learning_rate": 3.735029184239396e-05,
+      "loss": 1.1056,
+      "num_input_tokens_seen": 63815680,
+      "step": 7790
+    },
+    {
+      "epoch": 1.0078821553172244,
+      "grad_norm": 0.5353860855102539,
+      "learning_rate": 3.73208681067413e-05,
+      "loss": 0.9637,
+      "num_input_tokens_seen": 63897600,
+      "step": 7800
+    },
+    {
+      "epoch": 1.0091743119266054,
+      "grad_norm": 0.3585314452648163,
+      "learning_rate": 3.7291421811624216e-05,
+      "loss": 0.6649,
+      "num_input_tokens_seen": 63979520,
+      "step": 7810
+    },
+    {
+      "epoch": 1.0104664685359865,
+      "grad_norm": 0.515984296798706,
+      "learning_rate": 3.726195301095877e-05,
+      "loss": 0.5021,
+      "num_input_tokens_seen": 64061440,
+      "step": 7820
+    },
+    {
+      "epoch": 1.0117586251453676,
+      "grad_norm": 0.5484461784362793,
+      "learning_rate": 3.7232461758702244e-05,
+      "loss": 0.9135,
+      "num_input_tokens_seen": 64143360,
+      "step": 7830
+    },
+    {
+      "epoch": 1.0130507817547487,
+      "grad_norm": 0.6354055404663086,
+      "learning_rate": 3.7202948108852984e-05,
+      "loss": 0.6548,
+      "num_input_tokens_seen": 64225280,
+      "step": 7840
+    },
+    {
+      "epoch": 1.0143429383641298,
+      "grad_norm": 0.6655207276344299,
+      "learning_rate": 3.717341211545039e-05,
+      "loss": 0.6679,
+      "num_input_tokens_seen": 64307200,
+      "step": 7850
+    },
+    {
+      "epoch": 1.015635094973511,
+      "grad_norm": 0.5719141364097595,
+      "learning_rate": 3.714385383257477e-05,
+      "loss": 1.115,
+      "num_input_tokens_seen": 64389120,
+      "step": 7860
+    },
+    {
+      "epoch": 1.0169272515828918,
+      "grad_norm": 0.6638296842575073,
+      "learning_rate": 3.711427331434721e-05,
+      "loss": 0.7203,
+      "num_input_tokens_seen": 64471040,
+      "step": 7870
+    },
+    {
+      "epoch": 1.0182194081922729,
+      "grad_norm": 0.5564717650413513,
+      "learning_rate": 3.7084670614929554e-05,
+      "loss": 0.7494,
+      "num_input_tokens_seen": 64552960,
+      "step": 7880
+    },
+    {
+      "epoch": 1.019511564801654,
+      "grad_norm": 0.42404186725616455,
+      "learning_rate": 3.7055045788524214e-05,
+      "loss": 0.7702,
+      "num_input_tokens_seen": 64634880,
+      "step": 7890
+    },
+    {
+      "epoch": 1.020803721411035,
+      "grad_norm": 0.5771876573562622,
+      "learning_rate": 3.702539888937414e-05,
+      "loss": 1.0498,
+      "num_input_tokens_seen": 64716800,
+      "step": 7900
+    },
+    {
+      "epoch": 1.0220958780204161,
+      "grad_norm": 0.6172158718109131,
+      "learning_rate": 3.699572997176272e-05,
+      "loss": 0.8016,
+      "num_input_tokens_seen": 64798720,
+      "step": 7910
+    },
+    {
+      "epoch": 1.0233880346297972,
+      "grad_norm": 0.6441952586174011,
+      "learning_rate": 3.696603909001361e-05,
+      "loss": 0.741,
+      "num_input_tokens_seen": 64880640,
+      "step": 7920
+    },
+    {
+      "epoch": 1.024680191239178,
+      "grad_norm": 0.5712997913360596,
+      "learning_rate": 3.69363262984907e-05,
+      "loss": 0.7687,
+      "num_input_tokens_seen": 64962560,
+      "step": 7930
+    },
+    {
+      "epoch": 1.0259723478485592,
+      "grad_norm": 0.5301061272621155,
+      "learning_rate": 3.690659165159803e-05,
+      "loss": 1.0326,
+      "num_input_tokens_seen": 65044480,
+      "step": 7940
+    },
+    {
+      "epoch": 1.0272645044579403,
+      "grad_norm": 0.4970210790634155,
+      "learning_rate": 3.6876835203779615e-05,
+      "loss": 0.961,
+      "num_input_tokens_seen": 65126400,
+      "step": 7950
+    },
+    {
+      "epoch": 1.0285566610673214,
+      "grad_norm": 0.5584379434585571,
+      "learning_rate": 3.68470570095194e-05,
+      "loss": 0.8982,
+      "num_input_tokens_seen": 65208320,
+      "step": 7960
+    },
+    {
+      "epoch": 1.0298488176767024,
+      "grad_norm": 0.6428322792053223,
+      "learning_rate": 3.681725712334115e-05,
+      "loss": 0.8534,
+      "num_input_tokens_seen": 65290240,
+      "step": 7970
+    },
+    {
+      "epoch": 1.0311409742860835,
+      "grad_norm": 0.2095443606376648,
+      "learning_rate": 3.678743559980835e-05,
+      "loss": 0.6313,
+      "num_input_tokens_seen": 65372160,
+      "step": 7980
+    },
+    {
+      "epoch": 1.0324331308954646,
+      "grad_norm": 0.6237980127334595,
+      "learning_rate": 3.67575924935241e-05,
+      "loss": 0.9277,
+      "num_input_tokens_seen": 65454080,
+      "step": 7990
+    },
+    {
+      "epoch": 1.0337252875048455,
+      "grad_norm": 0.5029511451721191,
+      "learning_rate": 3.672772785913102e-05,
+      "loss": 0.8789,
+      "num_input_tokens_seen": 65536000,
+      "step": 8000
+    },
+    {
+      "epoch": 1.0350174441142266,
+      "grad_norm": 0.2065262496471405,
+      "learning_rate": 3.669784175131115e-05,
+      "loss": 0.4837,
+      "num_input_tokens_seen": 65617920,
+      "step": 8010
+    },
+    {
+      "epoch": 1.0363096007236077,
+      "grad_norm": 0.44190657138824463,
+      "learning_rate": 3.666793422478583e-05,
+      "loss": 0.7244,
+      "num_input_tokens_seen": 65699840,
+      "step": 8020
+    },
+    {
+      "epoch": 1.0376017573329888,
+      "grad_norm": 0.49034252762794495,
+      "learning_rate": 3.663800533431564e-05,
+      "loss": 0.63,
+      "num_input_tokens_seen": 65781760,
+      "step": 8030
+    },
+    {
+      "epoch": 1.0388939139423699,
+      "grad_norm": 0.6004568934440613,
+      "learning_rate": 3.660805513470027e-05,
+      "loss": 1.0153,
+      "num_input_tokens_seen": 65863680,
+      "step": 8040
+    },
+    {
+      "epoch": 1.040186070551751,
+      "grad_norm": 0.49091774225234985,
+      "learning_rate": 3.657808368077843e-05,
+      "loss": 0.7835,
+      "num_input_tokens_seen": 65945600,
+      "step": 8050
+    },
+    {
+      "epoch": 1.041478227161132,
+      "grad_norm": 0.46571245789527893,
+      "learning_rate": 3.654809102742773e-05,
+      "loss": 0.7988,
+      "num_input_tokens_seen": 66027520,
+      "step": 8060
+    },
+    {
+      "epoch": 1.042770383770513,
+      "grad_norm": 0.7933652400970459,
+      "learning_rate": 3.651807722956462e-05,
+      "loss": 1.3038,
+      "num_input_tokens_seen": 66109440,
+      "step": 8070
+    },
+    {
+      "epoch": 1.044062540379894,
+      "grad_norm": 0.584967851638794,
+      "learning_rate": 3.648804234214425e-05,
+      "loss": 0.7774,
+      "num_input_tokens_seen": 66191360,
+      "step": 8080
+    },
+    {
+      "epoch": 1.045354696989275,
+      "grad_norm": 0.49486130475997925,
+      "learning_rate": 3.645798642016039e-05,
+      "loss": 0.7951,
+      "num_input_tokens_seen": 66273280,
+      "step": 8090
+    },
+    {
+      "epoch": 1.0466468535986562,
+      "grad_norm": 0.5906932353973389,
+      "learning_rate": 3.642790951864532e-05,
+      "loss": 0.7105,
+      "num_input_tokens_seen": 66355200,
+      "step": 8100
+    },
+    {
+      "epoch": 1.0479390102080373,
+      "grad_norm": 0.6410627961158752,
+      "learning_rate": 3.639781169266975e-05,
+      "loss": 0.644,
+      "num_input_tokens_seen": 66437120,
+      "step": 8110
+    },
+    {
+      "epoch": 1.0492311668174183,
+      "grad_norm": 0.8290396332740784,
+      "learning_rate": 3.636769299734267e-05,
+      "loss": 0.7002,
+      "num_input_tokens_seen": 66519040,
+      "step": 8120
+    },
+    {
+      "epoch": 1.0505233234267992,
+      "grad_norm": 0.5405248403549194,
+      "learning_rate": 3.63375534878113e-05,
+      "loss": 0.9435,
+      "num_input_tokens_seen": 66600960,
+      "step": 8130
+    },
+    {
+      "epoch": 1.0518154800361803,
+      "grad_norm": 0.4574602544307709,
+      "learning_rate": 3.6307393219261e-05,
+      "loss": 0.8839,
+      "num_input_tokens_seen": 66682880,
+      "step": 8140
+    },
+    {
+      "epoch": 1.0531076366455614,
+      "grad_norm": 0.5063459873199463,
+      "learning_rate": 3.627721224691507e-05,
+      "loss": 0.8676,
+      "num_input_tokens_seen": 66764800,
+      "step": 8150
+    },
+    {
+      "epoch": 1.0543997932549425,
+      "grad_norm": 0.5998566150665283,
+      "learning_rate": 3.6247010626034795e-05,
+      "loss": 0.6555,
+      "num_input_tokens_seen": 66846720,
+      "step": 8160
+    },
+    {
+      "epoch": 1.0556919498643236,
+      "grad_norm": 0.6061252951622009,
+      "learning_rate": 3.621678841191922e-05,
+      "loss": 0.8207,
+      "num_input_tokens_seen": 66928640,
+      "step": 8170
+    },
+    {
+      "epoch": 1.0569841064737047,
+      "grad_norm": 0.6531092524528503,
+      "learning_rate": 3.618654565990511e-05,
+      "loss": 1.0707,
+      "num_input_tokens_seen": 67010560,
+      "step": 8180
+    },
+    {
+      "epoch": 1.0582762630830858,
+      "grad_norm": 0.7719940543174744,
+      "learning_rate": 3.615628242536682e-05,
+      "loss": 0.7523,
+      "num_input_tokens_seen": 67092480,
+      "step": 8190
+    },
+    {
+      "epoch": 1.0595684196924666,
+      "grad_norm": 0.28535163402557373,
+      "learning_rate": 3.612599876371625e-05,
+      "loss": 0.7847,
+      "num_input_tokens_seen": 67174400,
+      "step": 8200
+    },
+    {
+      "epoch": 1.0608605763018477,
+      "grad_norm": 0.5150948166847229,
+      "learning_rate": 3.609569473040265e-05,
+      "loss": 1.0002,
+      "num_input_tokens_seen": 67256320,
+      "step": 8210
+    },
+    {
+      "epoch": 1.0621527329112288,
+      "grad_norm": 0.47803130745887756,
+      "learning_rate": 3.6065370380912587e-05,
+      "loss": 0.9216,
+      "num_input_tokens_seen": 67338240,
+      "step": 8220
+    },
+    {
+      "epoch": 1.06344488952061,
+      "grad_norm": 0.5799218416213989,
+      "learning_rate": 3.603502577076986e-05,
+      "loss": 0.9941,
+      "num_input_tokens_seen": 67420160,
+      "step": 8230
+    },
+    {
+      "epoch": 1.064737046129991,
+      "grad_norm": 0.23019935190677643,
+      "learning_rate": 3.600466095553532e-05,
+      "loss": 0.6576,
+      "num_input_tokens_seen": 67502080,
+      "step": 8240
+    },
+    {
+      "epoch": 1.066029202739372,
+      "grad_norm": 0.4924595057964325,
+      "learning_rate": 3.5974275990806846e-05,
+      "loss": 1.2263,
+      "num_input_tokens_seen": 67584000,
+      "step": 8250
+    },
+    {
+      "epoch": 1.0673213593487532,
+      "grad_norm": 0.6706606149673462,
+      "learning_rate": 3.5943870932219184e-05,
+      "loss": 0.7686,
+      "num_input_tokens_seen": 67665920,
+      "step": 8260
+    },
+    {
+      "epoch": 1.068613515958134,
+      "grad_norm": 0.4998605251312256,
+      "learning_rate": 3.59134458354439e-05,
+      "loss": 0.9912,
+      "num_input_tokens_seen": 67747840,
+      "step": 8270
+    },
+    {
+      "epoch": 1.0699056725675151,
+      "grad_norm": 0.5857017040252686,
+      "learning_rate": 3.588300075618922e-05,
+      "loss": 0.7774,
+      "num_input_tokens_seen": 67829760,
+      "step": 8280
+    },
+    {
+      "epoch": 1.0711978291768962,
+      "grad_norm": 0.23385043442249298,
+      "learning_rate": 3.5852535750199977e-05,
+      "loss": 1.002,
+      "num_input_tokens_seen": 67911680,
+      "step": 8290
+    },
+    {
+      "epoch": 1.0724899857862773,
+      "grad_norm": 0.5875013470649719,
+      "learning_rate": 3.5822050873257494e-05,
+      "loss": 0.767,
+      "num_input_tokens_seen": 67993600,
+      "step": 8300
+    },
+    {
+      "epoch": 1.0737821423956584,
+      "grad_norm": 0.5456327795982361,
+      "learning_rate": 3.579154618117946e-05,
+      "loss": 1.0006,
+      "num_input_tokens_seen": 68075520,
+      "step": 8310
+    },
+    {
+      "epoch": 1.0750742990050395,
+      "grad_norm": 0.2565229535102844,
+      "learning_rate": 3.576102172981986e-05,
+      "loss": 0.4659,
+      "num_input_tokens_seen": 68157440,
+      "step": 8320
+    },
+    {
+      "epoch": 1.0763664556144206,
+      "grad_norm": 0.510844886302948,
+      "learning_rate": 3.5730477575068845e-05,
+      "loss": 0.9332,
+      "num_input_tokens_seen": 68239360,
+      "step": 8330
+    },
+    {
+      "epoch": 1.0776586122238014,
+      "grad_norm": 0.5561206936836243,
+      "learning_rate": 3.5699913772852664e-05,
+      "loss": 0.7617,
+      "num_input_tokens_seen": 68321280,
+      "step": 8340
+    },
+    {
+      "epoch": 1.0789507688331825,
+      "grad_norm": 0.6247846484184265,
+      "learning_rate": 3.566933037913351e-05,
+      "loss": 1.1367,
+      "num_input_tokens_seen": 68403200,
+      "step": 8350
+    },
+    {
+      "epoch": 1.0802429254425636,
+      "grad_norm": 0.8798884749412537,
+      "learning_rate": 3.5638727449909473e-05,
+      "loss": 0.6604,
+      "num_input_tokens_seen": 68485120,
+      "step": 8360
+    },
+    {
+      "epoch": 1.0815350820519447,
+      "grad_norm": 0.44991663098335266,
+      "learning_rate": 3.560810504121441e-05,
+      "loss": 1.0806,
+      "num_input_tokens_seen": 68567040,
+      "step": 8370
+    },
+    {
+      "epoch": 1.0828272386613258,
+      "grad_norm": 0.3702957332134247,
+      "learning_rate": 3.5577463209117833e-05,
+      "loss": 0.7424,
+      "num_input_tokens_seen": 68648960,
+      "step": 8380
+    },
+    {
+      "epoch": 1.0841193952707069,
+      "grad_norm": 0.8632538318634033,
+      "learning_rate": 3.554680200972482e-05,
+      "loss": 0.78,
+      "num_input_tokens_seen": 68730880,
+      "step": 8390
+    },
+    {
+      "epoch": 1.085411551880088,
+      "grad_norm": 0.5223381519317627,
+      "learning_rate": 3.551612149917593e-05,
+      "loss": 1.0331,
+      "num_input_tokens_seen": 68812800,
+      "step": 8400
+    },
+    {
+      "epoch": 1.0867037084894688,
+      "grad_norm": 0.6399882435798645,
+      "learning_rate": 3.548542173364705e-05,
+      "loss": 1.3448,
+      "num_input_tokens_seen": 68894720,
+      "step": 8410
+    },
+    {
+      "epoch": 1.08799586509885,
+      "grad_norm": 0.8672426342964172,
+      "learning_rate": 3.545470276934934e-05,
+      "loss": 0.9444,
+      "num_input_tokens_seen": 68976640,
+      "step": 8420
+    },
+    {
+      "epoch": 1.089288021708231,
+      "grad_norm": 0.998540461063385,
+      "learning_rate": 3.542396466252913e-05,
+      "loss": 0.6718,
+      "num_input_tokens_seen": 69058560,
+      "step": 8430
+    },
+    {
+      "epoch": 1.0905801783176121,
+      "grad_norm": 0.8492835760116577,
+      "learning_rate": 3.539320746946775e-05,
+      "loss": 0.8492,
+      "num_input_tokens_seen": 69140480,
+      "step": 8440
+    },
+    {
+      "epoch": 1.0918723349269932,
+      "grad_norm": 0.47759363055229187,
+      "learning_rate": 3.5362431246481536e-05,
+      "loss": 0.5818,
+      "num_input_tokens_seen": 69222400,
+      "step": 8450
+    },
+    {
+      "epoch": 1.0931644915363743,
+      "grad_norm": 0.5898222327232361,
+      "learning_rate": 3.533163604992163e-05,
+      "loss": 0.8846,
+      "num_input_tokens_seen": 69304320,
+      "step": 8460
+    },
+    {
+      "epoch": 1.0944566481457552,
+      "grad_norm": 0.672615110874176,
+      "learning_rate": 3.5300821936173926e-05,
+      "loss": 0.8992,
+      "num_input_tokens_seen": 69386240,
+      "step": 8470
+    },
+    {
+      "epoch": 1.0957488047551363,
+      "grad_norm": 0.265041321516037,
+      "learning_rate": 3.526998896165894e-05,
+      "loss": 0.9501,
+      "num_input_tokens_seen": 69468160,
+      "step": 8480
+    },
+    {
+      "epoch": 1.0970409613645173,
+      "grad_norm": 0.5370000600814819,
+      "learning_rate": 3.523913718283175e-05,
+      "loss": 1.1183,
+      "num_input_tokens_seen": 69550080,
+      "step": 8490
+    },
+    {
+      "epoch": 1.0983331179738984,
+      "grad_norm": 0.6693633794784546,
+      "learning_rate": 3.520826665618184e-05,
+      "loss": 0.9322,
+      "num_input_tokens_seen": 69632000,
+      "step": 8500
+    },
+    {
+      "epoch": 1.0996252745832795,
+      "grad_norm": 0.7608528733253479,
+      "learning_rate": 3.5177377438233044e-05,
+      "loss": 0.6564,
+      "num_input_tokens_seen": 69713920,
+      "step": 8510
+    },
+    {
+      "epoch": 1.1009174311926606,
+      "grad_norm": 0.5762624740600586,
+      "learning_rate": 3.514646958554339e-05,
+      "loss": 1.0945,
+      "num_input_tokens_seen": 69795840,
+      "step": 8520
+    },
+    {
+      "epoch": 1.1022095878020417,
+      "grad_norm": 0.7134038209915161,
+      "learning_rate": 3.511554315470507e-05,
+      "loss": 0.8922,
+      "num_input_tokens_seen": 69877760,
+      "step": 8530
+    },
+    {
+      "epoch": 1.1035017444114226,
+      "grad_norm": 0.6153433322906494,
+      "learning_rate": 3.508459820234423e-05,
+      "loss": 0.9603,
+      "num_input_tokens_seen": 69959680,
+      "step": 8540
+    },
+    {
+      "epoch": 1.1047939010208037,
+      "grad_norm": 0.765910267829895,
+      "learning_rate": 3.5053634785121e-05,
+      "loss": 0.8906,
+      "num_input_tokens_seen": 70041600,
+      "step": 8550
+    },
+    {
+      "epoch": 1.1060860576301847,
+      "grad_norm": 0.4776328206062317,
+      "learning_rate": 3.5022652959729266e-05,
+      "loss": 0.7746,
+      "num_input_tokens_seen": 70123520,
+      "step": 8560
+    },
+    {
+      "epoch": 1.1073782142395658,
+      "grad_norm": 0.5070037245750427,
+      "learning_rate": 3.499165278289663e-05,
+      "loss": 0.7997,
+      "num_input_tokens_seen": 70205440,
+      "step": 8570
+    },
+    {
+      "epoch": 1.108670370848947,
+      "grad_norm": 0.759103000164032,
+      "learning_rate": 3.496063431138431e-05,
+      "loss": 0.7416,
+      "num_input_tokens_seen": 70287360,
+      "step": 8580
+    },
+    {
+      "epoch": 1.109962527458328,
+      "grad_norm": 0.6531251668930054,
+      "learning_rate": 3.492959760198702e-05,
+      "loss": 0.7489,
+      "num_input_tokens_seen": 70369280,
+      "step": 8590
+    },
+    {
+      "epoch": 1.111254684067709,
+      "grad_norm": 0.34367960691452026,
+      "learning_rate": 3.489854271153285e-05,
+      "loss": 0.8175,
+      "num_input_tokens_seen": 70451200,
+      "step": 8600
+    },
+    {
+      "epoch": 1.11254684067709,
+      "grad_norm": 0.7481246590614319,
+      "learning_rate": 3.4867469696883204e-05,
+      "loss": 0.6624,
+      "num_input_tokens_seen": 70533120,
+      "step": 8610
+    },
+    {
+      "epoch": 1.113838997286471,
+      "grad_norm": 0.6177128553390503,
+      "learning_rate": 3.483637861493264e-05,
+      "loss": 0.8943,
+      "num_input_tokens_seen": 70615040,
+      "step": 8620
+    },
+    {
+      "epoch": 1.1151311538958522,
+      "grad_norm": 0.699070930480957,
+      "learning_rate": 3.480526952260884e-05,
+      "loss": 0.9308,
+      "num_input_tokens_seen": 70696960,
+      "step": 8630
+    },
+    {
+      "epoch": 1.1164233105052332,
+      "grad_norm": 0.6072255969047546,
+      "learning_rate": 3.477414247687241e-05,
+      "loss": 0.6464,
+      "num_input_tokens_seen": 70778880,
+      "step": 8640
+    },
+    {
+      "epoch": 1.1177154671146143,
+      "grad_norm": 0.5646083950996399,
+      "learning_rate": 3.4742997534716884e-05,
+      "loss": 0.6793,
+      "num_input_tokens_seen": 70860800,
+      "step": 8650
+    },
+    {
+      "epoch": 1.1190076237239954,
+      "grad_norm": 0.571111798286438,
+      "learning_rate": 3.471183475316851e-05,
+      "loss": 0.95,
+      "num_input_tokens_seen": 70942720,
+      "step": 8660
+    },
+    {
+      "epoch": 1.1202997803333763,
+      "grad_norm": 0.2792120575904846,
+      "learning_rate": 3.468065418928625e-05,
+      "loss": 0.8991,
+      "num_input_tokens_seen": 71024640,
+      "step": 8670
+    },
+    {
+      "epoch": 1.1215919369427574,
+      "grad_norm": 0.8360409736633301,
+      "learning_rate": 3.4649455900161596e-05,
+      "loss": 1.0195,
+      "num_input_tokens_seen": 71106560,
+      "step": 8680
+    },
+    {
+      "epoch": 1.1228840935521385,
+      "grad_norm": 0.5351372957229614,
+      "learning_rate": 3.461823994291849e-05,
+      "loss": 1.0533,
+      "num_input_tokens_seen": 71188480,
+      "step": 8690
+    },
+    {
+      "epoch": 1.1241762501615196,
+      "grad_norm": 1.5099799633026123,
+      "learning_rate": 3.458700637471325e-05,
+      "loss": 0.7323,
+      "num_input_tokens_seen": 71270400,
+      "step": 8700
+    },
+    {
+      "epoch": 1.1254684067709007,
+      "grad_norm": 0.3392684757709503,
+      "learning_rate": 3.455575525273442e-05,
+      "loss": 0.4897,
+      "num_input_tokens_seen": 71352320,
+      "step": 8710
+    },
+    {
+      "epoch": 1.1267605633802817,
+      "grad_norm": 0.29004672169685364,
+      "learning_rate": 3.4524486634202685e-05,
+      "loss": 0.9862,
+      "num_input_tokens_seen": 71434240,
+      "step": 8720
+    },
+    {
+      "epoch": 1.1280527199896628,
+      "grad_norm": 0.9040658473968506,
+      "learning_rate": 3.4493200576370776e-05,
+      "loss": 0.7261,
+      "num_input_tokens_seen": 71516160,
+      "step": 8730
+    },
+    {
+      "epoch": 1.1293448765990437,
+      "grad_norm": 0.509054958820343,
+      "learning_rate": 3.4461897136523356e-05,
+      "loss": 0.8157,
+      "num_input_tokens_seen": 71598080,
+      "step": 8740
+    },
+    {
+      "epoch": 1.1306370332084248,
+      "grad_norm": 0.5921317338943481,
+      "learning_rate": 3.44305763719769e-05,
+      "loss": 1.0178,
+      "num_input_tokens_seen": 71680000,
+      "step": 8750
+    },
+    {
+      "epoch": 1.1319291898178059,
+      "grad_norm": 0.7024086117744446,
+      "learning_rate": 3.4399238340079607e-05,
+      "loss": 0.9631,
+      "num_input_tokens_seen": 71761920,
+      "step": 8760
+    },
+    {
+      "epoch": 1.133221346427187,
+      "grad_norm": 0.2952291667461395,
+      "learning_rate": 3.4367883098211316e-05,
+      "loss": 0.7918,
+      "num_input_tokens_seen": 71843840,
+      "step": 8770
+    },
+    {
+      "epoch": 1.134513503036568,
+      "grad_norm": 0.6267214417457581,
+      "learning_rate": 3.4336510703783345e-05,
+      "loss": 0.8197,
+      "num_input_tokens_seen": 71925760,
+      "step": 8780
+    },
+    {
+      "epoch": 1.1358056596459492,
+      "grad_norm": 0.2442624568939209,
+      "learning_rate": 3.4305121214238446e-05,
+      "loss": 0.6943,
+      "num_input_tokens_seen": 72007680,
+      "step": 8790
+    },
+    {
+      "epoch": 1.1370978162553302,
+      "grad_norm": 0.6830674409866333,
+      "learning_rate": 3.427371468705065e-05,
+      "loss": 0.9242,
+      "num_input_tokens_seen": 72089600,
+      "step": 8800
+    },
+    {
+      "epoch": 1.138389972864711,
+      "grad_norm": 0.41818463802337646,
+      "learning_rate": 3.42422911797252e-05,
+      "loss": 0.8938,
+      "num_input_tokens_seen": 72171520,
+      "step": 8810
+    },
+    {
+      "epoch": 1.1396821294740922,
+      "grad_norm": 0.5191003680229187,
+      "learning_rate": 3.4210850749798415e-05,
+      "loss": 1.0008,
+      "num_input_tokens_seen": 72253440,
+      "step": 8820
+    },
+    {
+      "epoch": 1.1409742860834733,
+      "grad_norm": 0.6296160817146301,
+      "learning_rate": 3.417939345483762e-05,
+      "loss": 0.6786,
+      "num_input_tokens_seen": 72335360,
+      "step": 8830
+    },
+    {
+      "epoch": 1.1422664426928544,
+      "grad_norm": 0.4805864989757538,
+      "learning_rate": 3.4147919352440995e-05,
+      "loss": 0.9551,
+      "num_input_tokens_seen": 72417280,
+      "step": 8840
+    },
+    {
+      "epoch": 1.1435585993022355,
+      "grad_norm": 0.64635169506073,
+      "learning_rate": 3.411642850023751e-05,
+      "loss": 0.7622,
+      "num_input_tokens_seen": 72499200,
+      "step": 8850
+    },
+    {
+      "epoch": 1.1448507559116166,
+      "grad_norm": 0.37850773334503174,
+      "learning_rate": 3.40849209558868e-05,
+      "loss": 0.5537,
+      "num_input_tokens_seen": 72581120,
+      "step": 8860
+    },
+    {
+      "epoch": 1.1461429125209976,
+      "grad_norm": 0.836184561252594,
+      "learning_rate": 3.405339677707906e-05,
+      "loss": 0.6828,
+      "num_input_tokens_seen": 72663040,
+      "step": 8870
+    },
+    {
+      "epoch": 1.1474350691303785,
+      "grad_norm": 0.5654054880142212,
+      "learning_rate": 3.402185602153495e-05,
+      "loss": 0.8754,
+      "num_input_tokens_seen": 72744960,
+      "step": 8880
+    },
+    {
+      "epoch": 1.1487272257397596,
+      "grad_norm": 0.6414130926132202,
+      "learning_rate": 3.3990298747005485e-05,
+      "loss": 1.1836,
+      "num_input_tokens_seen": 72826880,
+      "step": 8890
+    },
+    {
+      "epoch": 1.1500193823491407,
+      "grad_norm": 0.5052874088287354,
+      "learning_rate": 3.395872501127191e-05,
+      "loss": 0.9686,
+      "num_input_tokens_seen": 72908800,
+      "step": 8900
+    },
+    {
+      "epoch": 1.1513115389585218,
+      "grad_norm": 0.6401844620704651,
+      "learning_rate": 3.392713487214561e-05,
+      "loss": 0.9425,
+      "num_input_tokens_seen": 72990720,
+      "step": 8910
+    },
+    {
+      "epoch": 1.1526036955679029,
+      "grad_norm": 0.49744102358818054,
+      "learning_rate": 3.389552838746804e-05,
+      "loss": 0.6426,
+      "num_input_tokens_seen": 73072640,
+      "step": 8920
+    },
+    {
+      "epoch": 1.153895852177284,
+      "grad_norm": 0.6056483387947083,
+      "learning_rate": 3.386390561511055e-05,
+      "loss": 0.9311,
+      "num_input_tokens_seen": 73154560,
+      "step": 8930
+    },
+    {
+      "epoch": 1.155188008786665,
+      "grad_norm": 0.7190436720848083,
+      "learning_rate": 3.38322666129743e-05,
+      "loss": 0.94,
+      "num_input_tokens_seen": 73236480,
+      "step": 8940
+    },
+    {
+      "epoch": 1.156480165396046,
+      "grad_norm": 0.6182407736778259,
+      "learning_rate": 3.380061143899021e-05,
+      "loss": 0.8698,
+      "num_input_tokens_seen": 73318400,
+      "step": 8950
+    },
+    {
+      "epoch": 1.157772322005427,
+      "grad_norm": 0.22560502588748932,
+      "learning_rate": 3.376894015111876e-05,
+      "loss": 1.0229,
+      "num_input_tokens_seen": 73400320,
+      "step": 8960
+    },
+    {
+      "epoch": 1.159064478614808,
+      "grad_norm": 0.4778119921684265,
+      "learning_rate": 3.373725280735e-05,
+      "loss": 0.7057,
+      "num_input_tokens_seen": 73482240,
+      "step": 8970
+    },
+    {
+      "epoch": 1.1603566352241892,
+      "grad_norm": 0.5372010469436646,
+      "learning_rate": 3.3705549465703314e-05,
+      "loss": 0.8812,
+      "num_input_tokens_seen": 73564160,
+      "step": 8980
+    },
+    {
+      "epoch": 1.1616487918335703,
+      "grad_norm": 0.6138181090354919,
+      "learning_rate": 3.3673830184227414e-05,
+      "loss": 0.9767,
+      "num_input_tokens_seen": 73646080,
+      "step": 8990
+    },
+    {
+      "epoch": 1.1629409484429514,
+      "grad_norm": 0.7862790822982788,
+      "learning_rate": 3.3642095021000184e-05,
+      "loss": 1.0073,
+      "num_input_tokens_seen": 73728000,
+      "step": 9000
+    },
+    {
+      "epoch": 1.1642331050523325,
+      "grad_norm": 0.6123846769332886,
+      "learning_rate": 3.36103440341286e-05,
+      "loss": 0.5677,
+      "num_input_tokens_seen": 73809920,
+      "step": 9010
+    },
+    {
+      "epoch": 1.1655252616617133,
+      "grad_norm": 0.6137543320655823,
+      "learning_rate": 3.35785772817486e-05,
+      "loss": 0.5504,
+      "num_input_tokens_seen": 73891840,
+      "step": 9020
+    },
+    {
+      "epoch": 1.1668174182710944,
+      "grad_norm": 0.6806734204292297,
+      "learning_rate": 3.3546794822024976e-05,
+      "loss": 0.949,
+      "num_input_tokens_seen": 73973760,
+      "step": 9030
+    },
+    {
+      "epoch": 1.1681095748804755,
+      "grad_norm": 0.7158621549606323,
+      "learning_rate": 3.351499671315131e-05,
+      "loss": 0.9297,
+      "num_input_tokens_seen": 74055680,
+      "step": 9040
+    },
+    {
+      "epoch": 1.1694017314898566,
+      "grad_norm": 0.9826081395149231,
+      "learning_rate": 3.348318301334983e-05,
+      "loss": 0.993,
+      "num_input_tokens_seen": 74137600,
+      "step": 9050
+    },
+    {
+      "epoch": 1.1706938880992377,
+      "grad_norm": 0.6197934746742249,
+      "learning_rate": 3.3451353780871286e-05,
+      "loss": 0.6176,
+      "num_input_tokens_seen": 74219520,
+      "step": 9060
+    },
+    {
+      "epoch": 1.1719860447086188,
+      "grad_norm": 0.22937451303005219,
+      "learning_rate": 3.341950907399489e-05,
+      "loss": 0.5138,
+      "num_input_tokens_seen": 74301440,
+      "step": 9070
+    },
+    {
+      "epoch": 1.1732782013179996,
+      "grad_norm": 0.591966450214386,
+      "learning_rate": 3.338764895102821e-05,
+      "loss": 0.7563,
+      "num_input_tokens_seen": 74383360,
+      "step": 9080
+    },
+    {
+      "epoch": 1.1745703579273807,
+      "grad_norm": 0.6814236640930176,
+      "learning_rate": 3.335577347030697e-05,
+      "loss": 1.3017,
+      "num_input_tokens_seen": 74465280,
+      "step": 9090
+    },
+    {
+      "epoch": 1.1758625145367618,
+      "grad_norm": 0.9694181084632874,
+      "learning_rate": 3.33238826901951e-05,
+      "loss": 1.1248,
+      "num_input_tokens_seen": 74547200,
+      "step": 9100
+    },
+    {
+      "epoch": 1.177154671146143,
+      "grad_norm": 0.24693933129310608,
+      "learning_rate": 3.329197666908447e-05,
+      "loss": 0.6756,
+      "num_input_tokens_seen": 74629120,
+      "step": 9110
+    },
+    {
+      "epoch": 1.178446827755524,
+      "grad_norm": 0.5937339663505554,
+      "learning_rate": 3.32600554653949e-05,
+      "loss": 0.9454,
+      "num_input_tokens_seen": 74711040,
+      "step": 9120
+    },
+    {
+      "epoch": 1.179738984364905,
+      "grad_norm": 0.7767700552940369,
+      "learning_rate": 3.322811913757401e-05,
+      "loss": 0.863,
+      "num_input_tokens_seen": 74792960,
+      "step": 9130
+    },
+    {
+      "epoch": 1.181031140974286,
+      "grad_norm": 0.6188927292823792,
+      "learning_rate": 3.319616774409709e-05,
+      "loss": 0.8522,
+      "num_input_tokens_seen": 74874880,
+      "step": 9140
+    },
+    {
+      "epoch": 1.182323297583667,
+      "grad_norm": 0.5785146355628967,
+      "learning_rate": 3.316420134346701e-05,
+      "loss": 0.8277,
+      "num_input_tokens_seen": 74956800,
+      "step": 9150
+    },
+    {
+      "epoch": 1.1836154541930481,
+      "grad_norm": 0.6327735781669617,
+      "learning_rate": 3.313221999421415e-05,
+      "loss": 0.8846,
+      "num_input_tokens_seen": 75038720,
+      "step": 9160
+    },
+    {
+      "epoch": 1.1849076108024292,
+      "grad_norm": 0.338527113199234,
+      "learning_rate": 3.310022375489623e-05,
+      "loss": 0.6351,
+      "num_input_tokens_seen": 75120640,
+      "step": 9170
+    },
+    {
+      "epoch": 1.1861997674118103,
+      "grad_norm": 1.0302643775939941,
+      "learning_rate": 3.306821268409827e-05,
+      "loss": 1.023,
+      "num_input_tokens_seen": 75202560,
+      "step": 9180
+    },
+    {
+      "epoch": 1.1874919240211914,
+      "grad_norm": 0.4798928499221802,
+      "learning_rate": 3.30361868404324e-05,
+      "loss": 1.0803,
+      "num_input_tokens_seen": 75284480,
+      "step": 9190
+    },
+    {
+      "epoch": 1.1887840806305725,
+      "grad_norm": 0.48873379826545715,
+      "learning_rate": 3.300414628253783e-05,
+      "loss": 0.7852,
+      "num_input_tokens_seen": 75366400,
+      "step": 9200
+    },
+    {
+      "epoch": 1.1900762372399534,
+      "grad_norm": 0.8409120440483093,
+      "learning_rate": 3.297209106908072e-05,
+      "loss": 0.8063,
+      "num_input_tokens_seen": 75448320,
+      "step": 9210
+    },
+    {
+      "epoch": 1.1913683938493345,
+      "grad_norm": 0.26400497555732727,
+      "learning_rate": 3.294002125875402e-05,
+      "loss": 0.6504,
+      "num_input_tokens_seen": 75530240,
+      "step": 9220
+    },
+    {
+      "epoch": 1.1926605504587156,
+      "grad_norm": 0.6408680081367493,
+      "learning_rate": 3.290793691027746e-05,
+      "loss": 0.7654,
+      "num_input_tokens_seen": 75612160,
+      "step": 9230
+    },
+    {
+      "epoch": 1.1939527070680966,
+      "grad_norm": 0.8105805516242981,
+      "learning_rate": 3.287583808239735e-05,
+      "loss": 0.7852,
+      "num_input_tokens_seen": 75694080,
+      "step": 9240
+    },
+    {
+      "epoch": 1.1952448636774777,
+      "grad_norm": 0.5738030076026917,
+      "learning_rate": 3.284372483388652e-05,
+      "loss": 0.5354,
+      "num_input_tokens_seen": 75776000,
+      "step": 9250
+    },
+    {
+      "epoch": 1.1965370202868588,
+      "grad_norm": 0.5625064373016357,
+      "learning_rate": 3.2811597223544234e-05,
+      "loss": 0.8227,
+      "num_input_tokens_seen": 75857920,
+      "step": 9260
+    },
+    {
+      "epoch": 1.19782917689624,
+      "grad_norm": 0.5138773918151855,
+      "learning_rate": 3.277945531019601e-05,
+      "loss": 0.9902,
+      "num_input_tokens_seen": 75939840,
+      "step": 9270
+    },
+    {
+      "epoch": 1.1991213335056208,
+      "grad_norm": 0.5785451531410217,
+      "learning_rate": 3.274729915269358e-05,
+      "loss": 0.6786,
+      "num_input_tokens_seen": 76021760,
+      "step": 9280
+    },
+    {
+      "epoch": 1.2004134901150019,
+      "grad_norm": 0.3385809063911438,
+      "learning_rate": 3.271512880991476e-05,
+      "loss": 0.7933,
+      "num_input_tokens_seen": 76103680,
+      "step": 9290
+    },
+    {
+      "epoch": 1.201705646724383,
+      "grad_norm": 0.2932916581630707,
+      "learning_rate": 3.268294434076332e-05,
+      "loss": 0.8867,
+      "num_input_tokens_seen": 76185600,
+      "step": 9300
+    },
+    {
+      "epoch": 1.202997803333764,
+      "grad_norm": 0.5734339952468872,
+      "learning_rate": 3.26507458041689e-05,
+      "loss": 0.9081,
+      "num_input_tokens_seen": 76267520,
+      "step": 9310
+    },
+    {
+      "epoch": 1.2042899599431451,
+      "grad_norm": 0.6236972808837891,
+      "learning_rate": 3.261853325908691e-05,
+      "loss": 1.1582,
+      "num_input_tokens_seen": 76349440,
+      "step": 9320
+    },
+    {
+      "epoch": 1.2055821165525262,
+      "grad_norm": 0.6559344530105591,
+      "learning_rate": 3.2586306764498395e-05,
+      "loss": 1.1172,
+      "num_input_tokens_seen": 76431360,
+      "step": 9330
+    },
+    {
+      "epoch": 1.2068742731619073,
+      "grad_norm": 0.5605185627937317,
+      "learning_rate": 3.255406637940996e-05,
+      "loss": 0.8069,
+      "num_input_tokens_seen": 76513280,
+      "step": 9340
+    },
+    {
+      "epoch": 1.2081664297712882,
+      "grad_norm": 0.697979748249054,
+      "learning_rate": 3.252181216285363e-05,
+      "loss": 1.0322,
+      "num_input_tokens_seen": 76595200,
+      "step": 9350
+    },
+    {
+      "epoch": 1.2094585863806693,
+      "grad_norm": 0.7321066856384277,
+      "learning_rate": 3.2489544173886745e-05,
+      "loss": 1.1227,
+      "num_input_tokens_seen": 76677120,
+      "step": 9360
+    },
+    {
+      "epoch": 1.2107507429900504,
+      "grad_norm": 0.623622715473175,
+      "learning_rate": 3.245726247159189e-05,
+      "loss": 0.9295,
+      "num_input_tokens_seen": 76759040,
+      "step": 9370
+    },
+    {
+      "epoch": 1.2120428995994315,
+      "grad_norm": 0.35311391949653625,
+      "learning_rate": 3.242496711507673e-05,
+      "loss": 1.0272,
+      "num_input_tokens_seen": 76840960,
+      "step": 9380
+    },
+    {
+      "epoch": 1.2133350562088125,
+      "grad_norm": 0.5266684889793396,
+      "learning_rate": 3.239265816347397e-05,
+      "loss": 0.9163,
+      "num_input_tokens_seen": 76922880,
+      "step": 9390
+    },
+    {
+      "epoch": 1.2146272128181936,
+      "grad_norm": 0.4497738182544708,
+      "learning_rate": 3.236033567594115e-05,
+      "loss": 0.8623,
+      "num_input_tokens_seen": 77004800,
+      "step": 9400
+    },
+    {
+      "epoch": 1.2159193694275747,
+      "grad_norm": 0.5825258493423462,
+      "learning_rate": 3.232799971166064e-05,
+      "loss": 0.9241,
+      "num_input_tokens_seen": 77086720,
+      "step": 9410
+    },
+    {
+      "epoch": 1.2172115260369556,
+      "grad_norm": 0.9179586172103882,
+      "learning_rate": 3.2295650329839474e-05,
+      "loss": 0.75,
+      "num_input_tokens_seen": 77168640,
+      "step": 9420
+    },
+    {
+      "epoch": 1.2185036826463367,
+      "grad_norm": 0.2291688621044159,
+      "learning_rate": 3.2263287589709255e-05,
+      "loss": 0.6456,
+      "num_input_tokens_seen": 77250560,
+      "step": 9430
+    },
+    {
+      "epoch": 1.2197958392557178,
+      "grad_norm": 0.4688994288444519,
+      "learning_rate": 3.2230911550526035e-05,
+      "loss": 0.8976,
+      "num_input_tokens_seen": 77332480,
+      "step": 9440
+    },
+    {
+      "epoch": 1.2210879958650989,
+      "grad_norm": 0.5401532649993896,
+      "learning_rate": 3.219852227157022e-05,
+      "loss": 0.9984,
+      "num_input_tokens_seen": 77414400,
+      "step": 9450
+    },
+    {
+      "epoch": 1.22238015247448,
+      "grad_norm": 0.7293322086334229,
+      "learning_rate": 3.216611981214648e-05,
+      "loss": 1.0182,
+      "num_input_tokens_seen": 77496320,
+      "step": 9460
+    },
+    {
+      "epoch": 1.223672309083861,
+      "grad_norm": 0.19243964552879333,
+      "learning_rate": 3.2133704231583576e-05,
+      "loss": 0.6102,
+      "num_input_tokens_seen": 77578240,
+      "step": 9470
+    },
+    {
+      "epoch": 1.2249644656932421,
+      "grad_norm": 0.556209146976471,
+      "learning_rate": 3.210127558923434e-05,
+      "loss": 0.868,
+      "num_input_tokens_seen": 77660160,
+      "step": 9480
+    },
+    {
+      "epoch": 1.226256622302623,
+      "grad_norm": 0.29714149236679077,
+      "learning_rate": 3.206883394447547e-05,
+      "loss": 0.9574,
+      "num_input_tokens_seen": 77742080,
+      "step": 9490
+    },
+    {
+      "epoch": 1.227548778912004,
+      "grad_norm": 0.4670780599117279,
+      "learning_rate": 3.203637935670752e-05,
+      "loss": 0.852,
+      "num_input_tokens_seen": 77824000,
+      "step": 9500
+    },
+    {
+      "epoch": 1.2288409355213852,
+      "grad_norm": 0.5203260779380798,
+      "learning_rate": 3.200391188535472e-05,
+      "loss": 0.7199,
+      "num_input_tokens_seen": 77905920,
+      "step": 9510
+    },
+    {
+      "epoch": 1.2301330921307663,
+      "grad_norm": 0.28679078817367554,
+      "learning_rate": 3.197143158986489e-05,
+      "loss": 0.7419,
+      "num_input_tokens_seen": 77987840,
+      "step": 9520
+    },
+    {
+      "epoch": 1.2314252487401474,
+      "grad_norm": 0.5256805419921875,
+      "learning_rate": 3.193893852970932e-05,
+      "loss": 0.7953,
+      "num_input_tokens_seen": 78069760,
+      "step": 9530
+    },
+    {
+      "epoch": 1.2327174053495285,
+      "grad_norm": 0.8118154406547546,
+      "learning_rate": 3.1906432764382695e-05,
+      "loss": 0.8027,
+      "num_input_tokens_seen": 78151680,
+      "step": 9540
+    },
+    {
+      "epoch": 1.2340095619589095,
+      "grad_norm": 0.3833126425743103,
+      "learning_rate": 3.187391435340295e-05,
+      "loss": 0.7832,
+      "num_input_tokens_seen": 78233600,
+      "step": 9550
+    },
+    {
+      "epoch": 1.2353017185682904,
+      "grad_norm": 0.3153989315032959,
+      "learning_rate": 3.184138335631118e-05,
+      "loss": 0.7582,
+      "num_input_tokens_seen": 78315520,
+      "step": 9560
+    },
+    {
+      "epoch": 1.2365938751776715,
+      "grad_norm": 0.5784737467765808,
+      "learning_rate": 3.1808839832671523e-05,
+      "loss": 1.1442,
+      "num_input_tokens_seen": 78397440,
+      "step": 9570
+    },
+    {
+      "epoch": 1.2378860317870526,
+      "grad_norm": 0.6450271010398865,
+      "learning_rate": 3.1776283842071045e-05,
+      "loss": 0.9673,
+      "num_input_tokens_seen": 78479360,
+      "step": 9580
+    },
+    {
+      "epoch": 1.2391781883964337,
+      "grad_norm": 0.5860177874565125,
+      "learning_rate": 3.174371544411964e-05,
+      "loss": 0.8106,
+      "num_input_tokens_seen": 78561280,
+      "step": 9590
+    },
+    {
+      "epoch": 1.2404703450058148,
+      "grad_norm": 0.6213310956954956,
+      "learning_rate": 3.1711134698449946e-05,
+      "loss": 0.9658,
+      "num_input_tokens_seen": 78643200,
+      "step": 9600
+    },
+    {
+      "epoch": 1.2417625016151959,
+      "grad_norm": 0.5813376903533936,
+      "learning_rate": 3.167854166471717e-05,
+      "loss": 0.9531,
+      "num_input_tokens_seen": 78725120,
+      "step": 9610
+    },
+    {
+      "epoch": 1.2430546582245767,
+      "grad_norm": 0.3016079366207123,
+      "learning_rate": 3.164593640259904e-05,
+      "loss": 0.8195,
+      "num_input_tokens_seen": 78807040,
+      "step": 9620
+    },
+    {
+      "epoch": 1.2443468148339578,
+      "grad_norm": 0.7248777747154236,
+      "learning_rate": 3.161331897179568e-05,
+      "loss": 0.9972,
+      "num_input_tokens_seen": 78888960,
+      "step": 9630
+    },
+    {
+      "epoch": 1.245638971443339,
+      "grad_norm": 0.19406473636627197,
+      "learning_rate": 3.1580689432029484e-05,
+      "loss": 0.5308,
+      "num_input_tokens_seen": 78970880,
+      "step": 9640
+    },
+    {
+      "epoch": 1.24693112805272,
+      "grad_norm": 0.8860630989074707,
+      "learning_rate": 3.154804784304502e-05,
+      "loss": 1.0639,
+      "num_input_tokens_seen": 79052800,
+      "step": 9650
+    },
+    {
+      "epoch": 1.248223284662101,
+      "grad_norm": 0.792770504951477,
+      "learning_rate": 3.151539426460892e-05,
+      "loss": 1.2022,
+      "num_input_tokens_seen": 79134720,
+      "step": 9660
+    },
+    {
+      "epoch": 1.2495154412714822,
+      "grad_norm": 0.7376882433891296,
+      "learning_rate": 3.148272875650976e-05,
+      "loss": 0.8717,
+      "num_input_tokens_seen": 79216640,
+      "step": 9670
+    },
+    {
+      "epoch": 1.250807597880863,
+      "grad_norm": 0.5317556262016296,
+      "learning_rate": 3.145005137855796e-05,
+      "loss": 0.9134,
+      "num_input_tokens_seen": 79298560,
+      "step": 9680
+    },
+    {
+      "epoch": 1.2520997544902444,
+      "grad_norm": 0.2553159296512604,
+      "learning_rate": 3.14173621905857e-05,
+      "loss": 0.671,
+      "num_input_tokens_seen": 79380480,
+      "step": 9690
+    },
+    {
+      "epoch": 1.2533919110996252,
+      "grad_norm": 0.6660608649253845,
+      "learning_rate": 3.138466125244674e-05,
+      "loss": 1.0354,
+      "num_input_tokens_seen": 79462400,
+      "step": 9700
+    },
+    {
+      "epoch": 1.2546840677090063,
+      "grad_norm": 0.5901172757148743,
+      "learning_rate": 3.13519486240164e-05,
+      "loss": 1.3345,
+      "num_input_tokens_seen": 79544320,
+      "step": 9710
+    },
+    {
+      "epoch": 1.2559762243183874,
+      "grad_norm": 0.5691861510276794,
+      "learning_rate": 3.1319224365191366e-05,
+      "loss": 0.7637,
+      "num_input_tokens_seen": 79626240,
+      "step": 9720
+    },
+    {
+      "epoch": 1.2572683809277685,
+      "grad_norm": 0.5517555475234985,
+      "learning_rate": 3.128648853588965e-05,
+      "loss": 0.5803,
+      "num_input_tokens_seen": 79708160,
+      "step": 9730
+    },
+    {
+      "epoch": 1.2585605375371496,
+      "grad_norm": 0.9708176255226135,
+      "learning_rate": 3.1253741196050425e-05,
+      "loss": 0.4912,
+      "num_input_tokens_seen": 79790080,
+      "step": 9740
+    },
+    {
+      "epoch": 1.2598526941465304,
+      "grad_norm": 0.689449667930603,
+      "learning_rate": 3.122098240563396e-05,
+      "loss": 0.7291,
+      "num_input_tokens_seen": 79872000,
+      "step": 9750
+    },
+    {
+      "epoch": 1.2611448507559115,
+      "grad_norm": 0.8461759686470032,
+      "learning_rate": 3.118821222462147e-05,
+      "loss": 1.0072,
+      "num_input_tokens_seen": 79953920,
+      "step": 9760
+    },
+    {
+      "epoch": 1.2624370073652926,
+      "grad_norm": 0.9328972697257996,
+      "learning_rate": 3.1155430713015034e-05,
+      "loss": 0.8663,
+      "num_input_tokens_seen": 80035840,
+      "step": 9770
+    },
+    {
+      "epoch": 1.2637291639746737,
+      "grad_norm": 0.19583484530448914,
+      "learning_rate": 3.1122637930837486e-05,
+      "loss": 0.8375,
+      "num_input_tokens_seen": 80117760,
+      "step": 9780
+    },
+    {
+      "epoch": 1.2650213205840548,
+      "grad_norm": 0.22281506657600403,
+      "learning_rate": 3.10898339381323e-05,
+      "loss": 0.7423,
+      "num_input_tokens_seen": 80199680,
+      "step": 9790
+    },
+    {
+      "epoch": 1.266313477193436,
+      "grad_norm": 0.6488370299339294,
+      "learning_rate": 3.1057018794963454e-05,
+      "loss": 1.3639,
+      "num_input_tokens_seen": 80281600,
+      "step": 9800
+    },
+    {
+      "epoch": 1.267605633802817,
+      "grad_norm": 1.4011627435684204,
+      "learning_rate": 3.102419256141536e-05,
+      "loss": 0.7074,
+      "num_input_tokens_seen": 80363520,
+      "step": 9810
+    },
+    {
+      "epoch": 1.2688977904121979,
+      "grad_norm": 0.5548164248466492,
+      "learning_rate": 3.0991355297592734e-05,
+      "loss": 1.1645,
+      "num_input_tokens_seen": 80445440,
+      "step": 9820
+    },
+    {
+      "epoch": 1.270189947021579,
+      "grad_norm": 0.8125051856040955,
+      "learning_rate": 3.095850706362047e-05,
+      "loss": 0.699,
+      "num_input_tokens_seen": 80527360,
+      "step": 9830
+    },
+    {
+      "epoch": 1.27148210363096,
+      "grad_norm": 0.8188523650169373,
+      "learning_rate": 3.092564791964358e-05,
+      "loss": 1.1145,
+      "num_input_tokens_seen": 80609280,
+      "step": 9840
+    },
+    {
+      "epoch": 1.2727742602403411,
+      "grad_norm": 0.553065836429596,
+      "learning_rate": 3.089277792582704e-05,
+      "loss": 0.9243,
+      "num_input_tokens_seen": 80691200,
+      "step": 9850
+    },
+    {
+      "epoch": 1.2740664168497222,
+      "grad_norm": 0.6267945170402527,
+      "learning_rate": 3.085989714235568e-05,
+      "loss": 0.8664,
+      "num_input_tokens_seen": 80773120,
+      "step": 9860
+    },
+    {
+      "epoch": 1.2753585734591033,
+      "grad_norm": 0.551201343536377,
+      "learning_rate": 3.082700562943409e-05,
+      "loss": 0.8113,
+      "num_input_tokens_seen": 80855040,
+      "step": 9870
+    },
+    {
+      "epoch": 1.2766507300684844,
+      "grad_norm": 0.39583033323287964,
+      "learning_rate": 3.079410344728652e-05,
+      "loss": 1.0342,
+      "num_input_tokens_seen": 80936960,
+      "step": 9880
+    },
+    {
+      "epoch": 1.2779428866778653,
+      "grad_norm": 0.7506781816482544,
+      "learning_rate": 3.076119065615674e-05,
+      "loss": 0.8064,
+      "num_input_tokens_seen": 81018880,
+      "step": 9890
+    },
+    {
+      "epoch": 1.2792350432872464,
+      "grad_norm": 0.6547626852989197,
+      "learning_rate": 3.0728267316307945e-05,
+      "loss": 0.7267,
+      "num_input_tokens_seen": 81100800,
+      "step": 9900
+    },
+    {
+      "epoch": 1.2805271998966274,
+      "grad_norm": 0.8237736225128174,
+      "learning_rate": 3.069533348802266e-05,
+      "loss": 0.6364,
+      "num_input_tokens_seen": 81182720,
+      "step": 9910
+    },
+    {
+      "epoch": 1.2818193565060085,
+      "grad_norm": 0.5278318524360657,
+      "learning_rate": 3.0662389231602595e-05,
+      "loss": 0.7681,
+      "num_input_tokens_seen": 81264640,
+      "step": 9920
+    },
+    {
+      "epoch": 1.2831115131153896,
+      "grad_norm": 0.48821374773979187,
+      "learning_rate": 3.062943460736857e-05,
+      "loss": 0.6366,
+      "num_input_tokens_seen": 81346560,
+      "step": 9930
+    },
+    {
+      "epoch": 1.2844036697247707,
+      "grad_norm": 0.6118844747543335,
+      "learning_rate": 3.059646967566038e-05,
+      "loss": 0.8101,
+      "num_input_tokens_seen": 81428480,
+      "step": 9940
+    },
+    {
+      "epoch": 1.2856958263341518,
+      "grad_norm": 0.7095397710800171,
+      "learning_rate": 3.0563494496836686e-05,
+      "loss": 0.5936,
+      "num_input_tokens_seen": 81510400,
+      "step": 9950
+    },
+    {
+      "epoch": 1.2869879829435327,
+      "grad_norm": 0.8946405053138733,
+      "learning_rate": 3.0530509131274935e-05,
+      "loss": 0.8716,
+      "num_input_tokens_seen": 81592320,
+      "step": 9960
+    },
+    {
+      "epoch": 1.2882801395529138,
+      "grad_norm": 1.0983461141586304,
+      "learning_rate": 3.0497513639371195e-05,
+      "loss": 0.7882,
+      "num_input_tokens_seen": 81674240,
+      "step": 9970
+    },
+    {
+      "epoch": 1.2895722961622949,
+      "grad_norm": 0.9059731364250183,
+      "learning_rate": 3.04645080815401e-05,
+      "loss": 0.8302,
+      "num_input_tokens_seen": 81756160,
+      "step": 9980
+    },
+    {
+      "epoch": 1.290864452771676,
+      "grad_norm": 0.6056944131851196,
+      "learning_rate": 3.04314925182147e-05,
+      "loss": 0.9569,
+      "num_input_tokens_seen": 81838080,
+      "step": 9990
+    },
+    {
+      "epoch": 1.292156609381057,
+      "grad_norm": 0.536361575126648,
+      "learning_rate": 3.0398467009846375e-05,
+      "loss": 0.9173,
+      "num_input_tokens_seen": 81920000,
+      "step": 10000
+    },
+    {
+      "epoch": 1.2934487659904381,
+      "grad_norm": 0.6062700152397156,
+      "learning_rate": 3.0365431616904714e-05,
+      "loss": 0.9092,
+      "num_input_tokens_seen": 82001920,
+      "step": 10010
+    },
+    {
+      "epoch": 1.2947409225998192,
+      "grad_norm": 0.5536112189292908,
+      "learning_rate": 3.03323863998774e-05,
+      "loss": 0.8183,
+      "num_input_tokens_seen": 82083840,
+      "step": 10020
+    },
+    {
+      "epoch": 1.2960330792092,
+      "grad_norm": 0.6913787126541138,
+      "learning_rate": 3.02993314192701e-05,
+      "loss": 1.2083,
+      "num_input_tokens_seen": 82165760,
+      "step": 10030
+    },
+    {
+      "epoch": 1.2973252358185812,
+      "grad_norm": 0.8038882613182068,
+      "learning_rate": 3.0266266735606358e-05,
+      "loss": 1.1272,
+      "num_input_tokens_seen": 82247680,
+      "step": 10040
+    },
+    {
+      "epoch": 1.2986173924279623,
+      "grad_norm": 0.5819315910339355,
+      "learning_rate": 3.0233192409427492e-05,
+      "loss": 0.6936,
+      "num_input_tokens_seen": 82329600,
+      "step": 10050
+    },
+    {
+      "epoch": 1.2999095490373433,
+      "grad_norm": 0.5388988852500916,
+      "learning_rate": 3.0200108501292466e-05,
+      "loss": 1.0776,
+      "num_input_tokens_seen": 82411520,
+      "step": 10060
+    },
+    {
+      "epoch": 1.3012017056467244,
+      "grad_norm": 0.2774782180786133,
+      "learning_rate": 3.0167015071777815e-05,
+      "loss": 0.5866,
+      "num_input_tokens_seen": 82493440,
+      "step": 10070
+    },
+    {
+      "epoch": 1.3024938622561053,
+      "grad_norm": 0.20768149197101593,
+      "learning_rate": 3.0133912181477475e-05,
+      "loss": 0.903,
+      "num_input_tokens_seen": 82575360,
+      "step": 10080
+    },
+    {
+      "epoch": 1.3037860188654866,
+      "grad_norm": 0.5745561718940735,
+      "learning_rate": 3.010079989100271e-05,
+      "loss": 1.1334,
+      "num_input_tokens_seen": 82657280,
+      "step": 10090
+    },
+    {
+      "epoch": 1.3050781754748675,
+      "grad_norm": 0.5522057414054871,
+      "learning_rate": 3.0067678260982018e-05,
+      "loss": 0.708,
+      "num_input_tokens_seen": 82739200,
+      "step": 10100
+    },
+    {
+      "epoch": 1.3063703320842486,
+      "grad_norm": 0.5716668367385864,
+      "learning_rate": 3.003454735206097e-05,
+      "loss": 1.094,
+      "num_input_tokens_seen": 82821120,
+      "step": 10110
+    },
+    {
+      "epoch": 1.3076624886936297,
+      "grad_norm": 0.5987653732299805,
+      "learning_rate": 3.000140722490215e-05,
+      "loss": 0.7914,
+      "num_input_tokens_seen": 82903040,
+      "step": 10120
+    },
+    {
+      "epoch": 1.3089546453030108,
+      "grad_norm": 0.754988431930542,
+      "learning_rate": 2.9968257940184997e-05,
+      "loss": 0.882,
+      "num_input_tokens_seen": 82984960,
+      "step": 10130
+    },
+    {
+      "epoch": 1.3102468019123918,
+      "grad_norm": 0.2810363471508026,
+      "learning_rate": 2.9935099558605728e-05,
+      "loss": 0.7434,
+      "num_input_tokens_seen": 83066880,
+      "step": 10140
+    },
+    {
+      "epoch": 1.3115389585217727,
+      "grad_norm": 0.6438339352607727,
+      "learning_rate": 2.9901932140877232e-05,
+      "loss": 1.0257,
+      "num_input_tokens_seen": 83148800,
+      "step": 10150
+    },
+    {
+      "epoch": 1.312831115131154,
+      "grad_norm": 0.8908247947692871,
+      "learning_rate": 2.9868755747728927e-05,
+      "loss": 0.6914,
+      "num_input_tokens_seen": 83230720,
+      "step": 10160
+    },
+    {
+      "epoch": 1.314123271740535,
+      "grad_norm": 0.5516972541809082,
+      "learning_rate": 2.9835570439906657e-05,
+      "loss": 0.9318,
+      "num_input_tokens_seen": 83312640,
+      "step": 10170
+    },
+    {
+      "epoch": 1.315415428349916,
+      "grad_norm": 0.6137295961380005,
+      "learning_rate": 2.9802376278172612e-05,
+      "loss": 0.9633,
+      "num_input_tokens_seen": 83394560,
+      "step": 10180
+    },
+    {
+      "epoch": 1.316707584959297,
+      "grad_norm": 0.8933919668197632,
+      "learning_rate": 2.976917332330517e-05,
+      "loss": 1.1187,
+      "num_input_tokens_seen": 83476480,
+      "step": 10190
+    },
+    {
+      "epoch": 1.3179997415686782,
+      "grad_norm": 0.5550206303596497,
+      "learning_rate": 2.973596163609883e-05,
+      "loss": 0.9031,
+      "num_input_tokens_seen": 83558400,
+      "step": 10200
+    },
+    {
+      "epoch": 1.3192918981780593,
+      "grad_norm": 0.7092847228050232,
+      "learning_rate": 2.970274127736406e-05,
+      "loss": 0.8322,
+      "num_input_tokens_seen": 83640320,
+      "step": 10210
+    },
+    {
+      "epoch": 1.3205840547874401,
+      "grad_norm": 0.9514045119285583,
+      "learning_rate": 2.966951230792722e-05,
+      "loss": 0.7939,
+      "num_input_tokens_seen": 83722240,
+      "step": 10220
+    },
+    {
+      "epoch": 1.3218762113968214,
+      "grad_norm": 0.6002041697502136,
+      "learning_rate": 2.9636274788630437e-05,
+      "loss": 1.0287,
+      "num_input_tokens_seen": 83804160,
+      "step": 10230
+    },
+    {
+      "epoch": 1.3231683680062023,
+      "grad_norm": 0.3773179054260254,
+      "learning_rate": 2.9603028780331475e-05,
+      "loss": 0.8743,
+      "num_input_tokens_seen": 83886080,
+      "step": 10240
+    },
+    {
+      "epoch": 1.3244605246155834,
+      "grad_norm": 0.5780991911888123,
+      "learning_rate": 2.9569774343903662e-05,
+      "loss": 1.089,
+      "num_input_tokens_seen": 83968000,
+      "step": 10250
+    },
+    {
+      "epoch": 1.3257526812249645,
+      "grad_norm": 0.7904821634292603,
+      "learning_rate": 2.9536511540235744e-05,
+      "loss": 0.9047,
+      "num_input_tokens_seen": 84049920,
+      "step": 10260
+    },
+    {
+      "epoch": 1.3270448378343456,
+      "grad_norm": 0.6269042491912842,
+      "learning_rate": 2.9503240430231803e-05,
+      "loss": 0.7258,
+      "num_input_tokens_seen": 84131840,
+      "step": 10270
+    },
+    {
+      "epoch": 1.3283369944437267,
+      "grad_norm": 0.5256585478782654,
+      "learning_rate": 2.9469961074811103e-05,
+      "loss": 0.6985,
+      "num_input_tokens_seen": 84213760,
+      "step": 10280
+    },
+    {
+      "epoch": 1.3296291510531075,
+      "grad_norm": 0.6166103482246399,
+      "learning_rate": 2.9436673534908044e-05,
+      "loss": 0.8227,
+      "num_input_tokens_seen": 84295680,
+      "step": 10290
+    },
+    {
+      "epoch": 1.3309213076624886,
+      "grad_norm": 0.9073558449745178,
+      "learning_rate": 2.940337787147197e-05,
+      "loss": 0.883,
+      "num_input_tokens_seen": 84377600,
+      "step": 10300
+    },
+    {
+      "epoch": 1.3322134642718697,
+      "grad_norm": 0.6241338849067688,
+      "learning_rate": 2.9370074145467132e-05,
+      "loss": 1.2221,
+      "num_input_tokens_seen": 84459520,
+      "step": 10310
+    },
+    {
+      "epoch": 1.3335056208812508,
+      "grad_norm": 0.6946704983711243,
+      "learning_rate": 2.9336762417872516e-05,
+      "loss": 0.7841,
+      "num_input_tokens_seen": 84541440,
+      "step": 10320
+    },
+    {
+      "epoch": 1.3347977774906319,
+      "grad_norm": 1.0090748071670532,
+      "learning_rate": 2.9303442749681787e-05,
+      "loss": 0.8774,
+      "num_input_tokens_seen": 84623360,
+      "step": 10330
+    },
+    {
+      "epoch": 1.336089934100013,
+      "grad_norm": 0.5861837863922119,
+      "learning_rate": 2.927011520190313e-05,
+      "loss": 0.7681,
+      "num_input_tokens_seen": 84705280,
+      "step": 10340
+    },
+    {
+      "epoch": 1.337382090709394,
+      "grad_norm": 0.5563804507255554,
+      "learning_rate": 2.9236779835559165e-05,
+      "loss": 0.9598,
+      "num_input_tokens_seen": 84787200,
+      "step": 10350
+    },
+    {
+      "epoch": 1.338674247318775,
+      "grad_norm": 0.6404061317443848,
+      "learning_rate": 2.9203436711686817e-05,
+      "loss": 0.6958,
+      "num_input_tokens_seen": 84869120,
+      "step": 10360
+    },
+    {
+      "epoch": 1.339966403928156,
+      "grad_norm": 0.7478158473968506,
+      "learning_rate": 2.917008589133724e-05,
+      "loss": 1.0247,
+      "num_input_tokens_seen": 84951040,
+      "step": 10370
+    },
+    {
+      "epoch": 1.3412585605375371,
+      "grad_norm": 0.5483884811401367,
+      "learning_rate": 2.913672743557565e-05,
+      "loss": 0.5671,
+      "num_input_tokens_seen": 85032960,
+      "step": 10380
+    },
+    {
+      "epoch": 1.3425507171469182,
+      "grad_norm": 0.7392348647117615,
+      "learning_rate": 2.9103361405481272e-05,
+      "loss": 0.6446,
+      "num_input_tokens_seen": 85114880,
+      "step": 10390
+    },
+    {
+      "epoch": 1.3438428737562993,
+      "grad_norm": 1.0634355545043945,
+      "learning_rate": 2.906998786214717e-05,
+      "loss": 1.0231,
+      "num_input_tokens_seen": 85196800,
+      "step": 10400
+    },
+    {
+      "epoch": 1.3451350303656804,
+      "grad_norm": 0.5788999199867249,
+      "learning_rate": 2.9036606866680187e-05,
+      "loss": 0.8076,
+      "num_input_tokens_seen": 85278720,
+      "step": 10410
+    },
+    {
+      "epoch": 1.3464271869750615,
+      "grad_norm": 0.5242962837219238,
+      "learning_rate": 2.90032184802008e-05,
+      "loss": 1.0485,
+      "num_input_tokens_seen": 85360640,
+      "step": 10420
+    },
+    {
+      "epoch": 1.3477193435844423,
+      "grad_norm": 0.7182193994522095,
+      "learning_rate": 2.8969822763843018e-05,
+      "loss": 1.0753,
+      "num_input_tokens_seen": 85442560,
+      "step": 10430
+    },
+    {
+      "epoch": 1.3490115001938234,
+      "grad_norm": 1.0175756216049194,
+      "learning_rate": 2.8936419778754294e-05,
+      "loss": 0.9629,
+      "num_input_tokens_seen": 85524480,
+      "step": 10440
+    },
+    {
+      "epoch": 1.3503036568032045,
+      "grad_norm": 2.7674241065979004,
+      "learning_rate": 2.8903009586095353e-05,
+      "loss": 0.7803,
+      "num_input_tokens_seen": 85606400,
+      "step": 10450
+    },
+    {
+      "epoch": 1.3515958134125856,
+      "grad_norm": 0.27844467759132385,
+      "learning_rate": 2.8869592247040138e-05,
+      "loss": 0.9991,
+      "num_input_tokens_seen": 85688320,
+      "step": 10460
+    },
+    {
+      "epoch": 1.3528879700219667,
+      "grad_norm": 0.18930500745773315,
+      "learning_rate": 2.883616782277569e-05,
+      "loss": 0.7521,
+      "num_input_tokens_seen": 85770240,
+      "step": 10470
+    },
+    {
+      "epoch": 1.3541801266313478,
+      "grad_norm": 0.9251275658607483,
+      "learning_rate": 2.8802736374501994e-05,
+      "loss": 1.186,
+      "num_input_tokens_seen": 85852160,
+      "step": 10480
+    },
+    {
+      "epoch": 1.3554722832407289,
+      "grad_norm": 0.6336624026298523,
+      "learning_rate": 2.8769297963431908e-05,
+      "loss": 1.1393,
+      "num_input_tokens_seen": 85934080,
+      "step": 10490
+    },
+    {
+      "epoch": 1.3567644398501097,
+      "grad_norm": 0.6428319811820984,
+      "learning_rate": 2.8735852650791035e-05,
+      "loss": 0.6548,
+      "num_input_tokens_seen": 86016000,
+      "step": 10500
+    },
+    {
+      "epoch": 1.3580565964594908,
+      "grad_norm": 0.6062332391738892,
+      "learning_rate": 2.870240049781764e-05,
+      "loss": 0.9466,
+      "num_input_tokens_seen": 86097920,
+      "step": 10510
+    },
+    {
+      "epoch": 1.359348753068872,
+      "grad_norm": 0.6514495611190796,
+      "learning_rate": 2.8668941565762475e-05,
+      "loss": 0.9157,
+      "num_input_tokens_seen": 86179840,
+      "step": 10520
+    },
+    {
+      "epoch": 1.360640909678253,
+      "grad_norm": 0.593705952167511,
+      "learning_rate": 2.8635475915888732e-05,
+      "loss": 0.8647,
+      "num_input_tokens_seen": 86261760,
+      "step": 10530
+    },
+    {
+      "epoch": 1.361933066287634,
+      "grad_norm": 0.8103840947151184,
+      "learning_rate": 2.8602003609471888e-05,
+      "loss": 0.8976,
+      "num_input_tokens_seen": 86343680,
+      "step": 10540
+    },
+    {
+      "epoch": 1.3632252228970152,
+      "grad_norm": 0.28389742970466614,
+      "learning_rate": 2.856852470779962e-05,
+      "loss": 1.0414,
+      "num_input_tokens_seen": 86425600,
+      "step": 10550
+    },
+    {
+      "epoch": 1.3645173795063963,
+      "grad_norm": 0.7390335202217102,
+      "learning_rate": 2.853503927217167e-05,
+      "loss": 0.9523,
+      "num_input_tokens_seen": 86507520,
+      "step": 10560
+    },
+    {
+      "epoch": 1.3658095361157772,
+      "grad_norm": 0.5908385515213013,
+      "learning_rate": 2.8501547363899744e-05,
+      "loss": 0.7521,
+      "num_input_tokens_seen": 86589440,
+      "step": 10570
+    },
+    {
+      "epoch": 1.3671016927251582,
+      "grad_norm": 0.8718129992485046,
+      "learning_rate": 2.846804904430741e-05,
+      "loss": 0.9591,
+      "num_input_tokens_seen": 86671360,
+      "step": 10580
+    },
+    {
+      "epoch": 1.3683938493345393,
+      "grad_norm": 0.6367455124855042,
+      "learning_rate": 2.8434544374729965e-05,
+      "loss": 0.9854,
+      "num_input_tokens_seen": 86753280,
+      "step": 10590
+    },
+    {
+      "epoch": 1.3696860059439204,
+      "grad_norm": 0.7115527391433716,
+      "learning_rate": 2.8401033416514345e-05,
+      "loss": 1.0439,
+      "num_input_tokens_seen": 86835200,
+      "step": 10600
+    },
+    {
+      "epoch": 1.3709781625533015,
+      "grad_norm": 0.5063609480857849,
+      "learning_rate": 2.8367516231018976e-05,
+      "loss": 0.6884,
+      "num_input_tokens_seen": 86917120,
+      "step": 10610
+    },
+    {
+      "epoch": 1.3722703191626824,
+      "grad_norm": 1.5641584396362305,
+      "learning_rate": 2.8333992879613712e-05,
+      "loss": 0.9114,
+      "num_input_tokens_seen": 86999040,
+      "step": 10620
+    },
+    {
+      "epoch": 1.3735624757720637,
+      "grad_norm": 0.7738798260688782,
+      "learning_rate": 2.830046342367969e-05,
+      "loss": 0.8389,
+      "num_input_tokens_seen": 87080960,
+      "step": 10630
+    },
+    {
+      "epoch": 1.3748546323814446,
+      "grad_norm": 0.6683527827262878,
+      "learning_rate": 2.826692792460921e-05,
+      "loss": 0.5361,
+      "num_input_tokens_seen": 87162880,
+      "step": 10640
+    },
+    {
+      "epoch": 1.3761467889908257,
+      "grad_norm": 0.26369839906692505,
+      "learning_rate": 2.823338644380566e-05,
+      "loss": 1.0727,
+      "num_input_tokens_seen": 87244800,
+      "step": 10650
+    },
+    {
+      "epoch": 1.3774389456002067,
+      "grad_norm": 0.6391986012458801,
+      "learning_rate": 2.8199839042683363e-05,
+      "loss": 0.5925,
+      "num_input_tokens_seen": 87326720,
+      "step": 10660
+    },
+    {
+      "epoch": 1.3787311022095878,
+      "grad_norm": 0.6280574798583984,
+      "learning_rate": 2.8166285782667483e-05,
+      "loss": 0.9665,
+      "num_input_tokens_seen": 87408640,
+      "step": 10670
+    },
+    {
+      "epoch": 1.380023258818969,
+      "grad_norm": 0.8266412615776062,
+      "learning_rate": 2.8132726725193926e-05,
+      "loss": 0.9418,
+      "num_input_tokens_seen": 87490560,
+      "step": 10680
+    },
+    {
+      "epoch": 1.3813154154283498,
+      "grad_norm": 0.26182371377944946,
+      "learning_rate": 2.8099161931709195e-05,
+      "loss": 0.6926,
+      "num_input_tokens_seen": 87572480,
+      "step": 10690
+    },
+    {
+      "epoch": 1.382607572037731,
+      "grad_norm": 0.6281419992446899,
+      "learning_rate": 2.806559146367031e-05,
+      "loss": 1.0825,
+      "num_input_tokens_seen": 87654400,
+      "step": 10700
+    },
+    {
+      "epoch": 1.383899728647112,
+      "grad_norm": 0.5953611135482788,
+      "learning_rate": 2.803201538254467e-05,
+      "loss": 0.63,
+      "num_input_tokens_seen": 87736320,
+      "step": 10710
+    },
+    {
+      "epoch": 1.385191885256493,
+      "grad_norm": 0.5723059177398682,
+      "learning_rate": 2.799843374980996e-05,
+      "loss": 0.8605,
+      "num_input_tokens_seen": 87818240,
+      "step": 10720
+    },
+    {
+      "epoch": 1.3864840418658742,
+      "grad_norm": 0.2304965704679489,
+      "learning_rate": 2.796484662695402e-05,
+      "loss": 1.0539,
+      "num_input_tokens_seen": 87900160,
+      "step": 10730
+    },
+    {
+      "epoch": 1.3877761984752552,
+      "grad_norm": 0.6087478399276733,
+      "learning_rate": 2.7931254075474768e-05,
+      "loss": 0.7748,
+      "num_input_tokens_seen": 87982080,
+      "step": 10740
+    },
+    {
+      "epoch": 1.3890683550846363,
+      "grad_norm": 3.60947322845459,
+      "learning_rate": 2.789765615688003e-05,
+      "loss": 0.9568,
+      "num_input_tokens_seen": 88064000,
+      "step": 10750
+    },
+    {
+      "epoch": 1.3903605116940172,
+      "grad_norm": 0.5466760993003845,
+      "learning_rate": 2.786405293268747e-05,
+      "loss": 1.0008,
+      "num_input_tokens_seen": 88145920,
+      "step": 10760
+    },
+    {
+      "epoch": 1.3916526683033985,
+      "grad_norm": 0.8872881531715393,
+      "learning_rate": 2.7830444464424466e-05,
+      "loss": 0.8518,
+      "num_input_tokens_seen": 88227840,
+      "step": 10770
+    },
+    {
+      "epoch": 1.3929448249127794,
+      "grad_norm": 0.9006865620613098,
+      "learning_rate": 2.7796830813628004e-05,
+      "loss": 0.8596,
+      "num_input_tokens_seen": 88309760,
+      "step": 10780
+    },
+    {
+      "epoch": 1.3942369815221605,
+      "grad_norm": 0.2589983344078064,
+      "learning_rate": 2.776321204184456e-05,
+      "loss": 0.9577,
+      "num_input_tokens_seen": 88391680,
+      "step": 10790
+    },
+    {
+      "epoch": 1.3955291381315416,
+      "grad_norm": 0.2515423893928528,
+      "learning_rate": 2.772958821062997e-05,
+      "loss": 0.9272,
+      "num_input_tokens_seen": 88473600,
+      "step": 10800
+    },
+    {
+      "epoch": 1.3968212947409226,
+      "grad_norm": 0.8317649364471436,
+      "learning_rate": 2.7695959381549364e-05,
+      "loss": 0.8736,
+      "num_input_tokens_seen": 88555520,
+      "step": 10810
+    },
+    {
+      "epoch": 1.3981134513503037,
+      "grad_norm": 0.8371811509132385,
+      "learning_rate": 2.7662325616176993e-05,
+      "loss": 0.7469,
+      "num_input_tokens_seen": 88637440,
+      "step": 10820
+    },
+    {
+      "epoch": 1.3994056079596846,
+      "grad_norm": 0.5307440161705017,
+      "learning_rate": 2.7628686976096164e-05,
+      "loss": 0.9801,
+      "num_input_tokens_seen": 88719360,
+      "step": 10830
+    },
+    {
+      "epoch": 1.4006977645690657,
+      "grad_norm": 0.5864846110343933,
+      "learning_rate": 2.7595043522899093e-05,
+      "loss": 1.1059,
+      "num_input_tokens_seen": 88801280,
+      "step": 10840
+    },
+    {
+      "epoch": 1.4019899211784468,
+      "grad_norm": 0.8047347068786621,
+      "learning_rate": 2.756139531818684e-05,
+      "loss": 1.1236,
+      "num_input_tokens_seen": 88883200,
+      "step": 10850
+    },
+    {
+      "epoch": 1.4032820777878279,
+      "grad_norm": 0.6610074043273926,
+      "learning_rate": 2.7527742423569124e-05,
+      "loss": 0.7606,
+      "num_input_tokens_seen": 88965120,
+      "step": 10860
+    },
+    {
+      "epoch": 1.404574234397209,
+      "grad_norm": 0.7933541536331177,
+      "learning_rate": 2.7494084900664273e-05,
+      "loss": 0.7754,
+      "num_input_tokens_seen": 89047040,
+      "step": 10870
+    },
+    {
+      "epoch": 1.40586639100659,
+      "grad_norm": 0.6643183827400208,
+      "learning_rate": 2.746042281109911e-05,
+      "loss": 0.9668,
+      "num_input_tokens_seen": 89128960,
+      "step": 10880
+    },
+    {
+      "epoch": 1.4071585476159711,
+      "grad_norm": 0.6505457758903503,
+      "learning_rate": 2.7426756216508776e-05,
+      "loss": 0.9552,
+      "num_input_tokens_seen": 89210880,
+      "step": 10890
+    },
+    {
+      "epoch": 1.408450704225352,
+      "grad_norm": 0.7870205640792847,
+      "learning_rate": 2.7393085178536686e-05,
+      "loss": 0.8455,
+      "num_input_tokens_seen": 89292800,
+      "step": 10900
+    },
+    {
+      "epoch": 1.409742860834733,
+      "grad_norm": 0.5975139737129211,
+      "learning_rate": 2.7359409758834397e-05,
+      "loss": 0.515,
+      "num_input_tokens_seen": 89374720,
+      "step": 10910
+    },
+    {
+      "epoch": 1.4110350174441142,
+      "grad_norm": 0.8741236329078674,
+      "learning_rate": 2.7325730019061474e-05,
+      "loss": 0.5575,
+      "num_input_tokens_seen": 89456640,
+      "step": 10920
+    },
+    {
+      "epoch": 1.4123271740534953,
+      "grad_norm": 0.6208974719047546,
+      "learning_rate": 2.729204602088539e-05,
+      "loss": 0.7631,
+      "num_input_tokens_seen": 89538560,
+      "step": 10930
+    },
+    {
+      "epoch": 1.4136193306628764,
+      "grad_norm": 0.6551584005355835,
+      "learning_rate": 2.7258357825981433e-05,
+      "loss": 0.8117,
+      "num_input_tokens_seen": 89620480,
+      "step": 10940
+    },
+    {
+      "epoch": 1.4149114872722575,
+      "grad_norm": 1.0060181617736816,
+      "learning_rate": 2.7224665496032565e-05,
+      "loss": 0.7802,
+      "num_input_tokens_seen": 89702400,
+      "step": 10950
+    },
+    {
+      "epoch": 1.4162036438816386,
+      "grad_norm": 0.6730552315711975,
+      "learning_rate": 2.7190969092729308e-05,
+      "loss": 0.8345,
+      "num_input_tokens_seen": 89784320,
+      "step": 10960
+    },
+    {
+      "epoch": 1.4174958004910194,
+      "grad_norm": 0.5672726035118103,
+      "learning_rate": 2.7157268677769666e-05,
+      "loss": 1.1491,
+      "num_input_tokens_seen": 89866240,
+      "step": 10970
+    },
+    {
+      "epoch": 1.4187879571004005,
+      "grad_norm": 0.6032657623291016,
+      "learning_rate": 2.712356431285896e-05,
+      "loss": 1.3288,
+      "num_input_tokens_seen": 89948160,
+      "step": 10980
+    },
+    {
+      "epoch": 1.4200801137097816,
+      "grad_norm": 0.6020485758781433,
+      "learning_rate": 2.7089856059709774e-05,
+      "loss": 0.8851,
+      "num_input_tokens_seen": 90030080,
+      "step": 10990
+    },
+    {
+      "epoch": 1.4213722703191627,
+      "grad_norm": 0.7019742727279663,
+      "learning_rate": 2.7056143980041787e-05,
+      "loss": 0.7634,
+      "num_input_tokens_seen": 90112000,
+      "step": 11000
+    },
+    {
+      "epoch": 1.4226644269285438,
+      "grad_norm": 0.6318424940109253,
+      "learning_rate": 2.70224281355817e-05,
+      "loss": 0.8556,
+      "num_input_tokens_seen": 90193920,
+      "step": 11010
+    },
+    {
+      "epoch": 1.4239565835379249,
+      "grad_norm": 0.5350080132484436,
+      "learning_rate": 2.6988708588063093e-05,
+      "loss": 0.7443,
+      "num_input_tokens_seen": 90275840,
+      "step": 11020
+    },
+    {
+      "epoch": 1.425248740147306,
+      "grad_norm": 0.6152655482292175,
+      "learning_rate": 2.695498539922634e-05,
+      "loss": 0.9481,
+      "num_input_tokens_seen": 90357760,
+      "step": 11030
+    },
+    {
+      "epoch": 1.4265408967566868,
+      "grad_norm": 0.9493053555488586,
+      "learning_rate": 2.6921258630818475e-05,
+      "loss": 0.817,
+      "num_input_tokens_seen": 90439680,
+      "step": 11040
+    },
+    {
+      "epoch": 1.427833053366068,
+      "grad_norm": 8.474701881408691,
+      "learning_rate": 2.6887528344593087e-05,
+      "loss": 0.6403,
+      "num_input_tokens_seen": 90521600,
+      "step": 11050
+    },
+    {
+      "epoch": 1.429125209975449,
+      "grad_norm": 0.1764688491821289,
+      "learning_rate": 2.685379460231021e-05,
+      "loss": 0.8245,
+      "num_input_tokens_seen": 90603520,
+      "step": 11060
+    },
+    {
+      "epoch": 1.43041736658483,
+      "grad_norm": 1.4876654148101807,
+      "learning_rate": 2.6820057465736197e-05,
+      "loss": 1.2493,
+      "num_input_tokens_seen": 90685440,
+      "step": 11070
+    },
+    {
+      "epoch": 1.4317095231942112,
+      "grad_norm": 0.8256916403770447,
+      "learning_rate": 2.6786316996643623e-05,
+      "loss": 0.9927,
+      "num_input_tokens_seen": 90767360,
+      "step": 11080
+    },
+    {
+      "epoch": 1.4330016798035923,
+      "grad_norm": 0.5362711548805237,
+      "learning_rate": 2.6752573256811165e-05,
+      "loss": 0.9965,
+      "num_input_tokens_seen": 90849280,
+      "step": 11090
+    },
+    {
+      "epoch": 1.4342938364129734,
+      "grad_norm": 0.6876575350761414,
+      "learning_rate": 2.6718826308023487e-05,
+      "loss": 0.8742,
+      "num_input_tokens_seen": 90931200,
+      "step": 11100
+    },
+    {
+      "epoch": 1.4355859930223542,
+      "grad_norm": 0.5102024674415588,
+      "learning_rate": 2.668507621207113e-05,
+      "loss": 0.9033,
+      "num_input_tokens_seen": 91013120,
+      "step": 11110
+    },
+    {
+      "epoch": 1.4368781496317353,
+      "grad_norm": 1.2666608095169067,
+      "learning_rate": 2.6651323030750396e-05,
+      "loss": 0.7038,
+      "num_input_tokens_seen": 91095040,
+      "step": 11120
+    },
+    {
+      "epoch": 1.4381703062411164,
+      "grad_norm": 0.5506048202514648,
+      "learning_rate": 2.6617566825863237e-05,
+      "loss": 0.7839,
+      "num_input_tokens_seen": 91176960,
+      "step": 11130
+    },
+    {
+      "epoch": 1.4394624628504975,
+      "grad_norm": 0.5495603084564209,
+      "learning_rate": 2.6583807659217137e-05,
+      "loss": 0.7871,
+      "num_input_tokens_seen": 91258880,
+      "step": 11140
+    },
+    {
+      "epoch": 1.4407546194598786,
+      "grad_norm": 0.5707017183303833,
+      "learning_rate": 2.6550045592625007e-05,
+      "loss": 1.0343,
+      "num_input_tokens_seen": 91340800,
+      "step": 11150
+    },
+    {
+      "epoch": 1.4420467760692595,
+      "grad_norm": 0.6786594390869141,
+      "learning_rate": 2.651628068790507e-05,
+      "loss": 1.1501,
+      "num_input_tokens_seen": 91422720,
+      "step": 11160
+    },
+    {
+      "epoch": 1.4433389326786408,
+      "grad_norm": 1.807134747505188,
+      "learning_rate": 2.648251300688073e-05,
+      "loss": 0.8592,
+      "num_input_tokens_seen": 91504640,
+      "step": 11170
+    },
+    {
+      "epoch": 1.4446310892880216,
+      "grad_norm": 0.5651780366897583,
+      "learning_rate": 2.6448742611380515e-05,
+      "loss": 0.8201,
+      "num_input_tokens_seen": 91586560,
+      "step": 11180
+    },
+    {
+      "epoch": 1.4459232458974027,
+      "grad_norm": 0.630124568939209,
+      "learning_rate": 2.6414969563237874e-05,
+      "loss": 1.074,
+      "num_input_tokens_seen": 91668480,
+      "step": 11190
+    },
+    {
+      "epoch": 1.4472154025067838,
+      "grad_norm": 0.7884289026260376,
+      "learning_rate": 2.6381193924291143e-05,
+      "loss": 0.8012,
+      "num_input_tokens_seen": 91750400,
+      "step": 11200
+    },
+    {
+      "epoch": 1.448507559116165,
+      "grad_norm": 0.6598234176635742,
+      "learning_rate": 2.63474157563834e-05,
+      "loss": 1.1103,
+      "num_input_tokens_seen": 91832320,
+      "step": 11210
+    },
+    {
+      "epoch": 1.449799715725546,
+      "grad_norm": 0.7381977438926697,
+      "learning_rate": 2.6313635121362322e-05,
+      "loss": 0.6353,
+      "num_input_tokens_seen": 91914240,
+      "step": 11220
+    },
+    {
+      "epoch": 1.4510918723349269,
+      "grad_norm": 0.29279613494873047,
+      "learning_rate": 2.6279852081080153e-05,
+      "loss": 0.8131,
+      "num_input_tokens_seen": 91996160,
+      "step": 11230
+    },
+    {
+      "epoch": 1.4523840289443082,
+      "grad_norm": 0.6259168386459351,
+      "learning_rate": 2.6246066697393494e-05,
+      "loss": 0.8539,
+      "num_input_tokens_seen": 92078080,
+      "step": 11240
+    },
+    {
+      "epoch": 1.453676185553689,
+      "grad_norm": 0.2948238253593445,
+      "learning_rate": 2.6212279032163283e-05,
+      "loss": 0.799,
+      "num_input_tokens_seen": 92160000,
+      "step": 11250
+    },
+    {
+      "epoch": 1.4549683421630701,
+      "grad_norm": 0.8171486854553223,
+      "learning_rate": 2.6178489147254598e-05,
+      "loss": 0.5722,
+      "num_input_tokens_seen": 92241920,
+      "step": 11260
+    },
+    {
+      "epoch": 1.4562604987724512,
+      "grad_norm": 0.7403898239135742,
+      "learning_rate": 2.6144697104536597e-05,
+      "loss": 0.8796,
+      "num_input_tokens_seen": 92323840,
+      "step": 11270
+    },
+    {
+      "epoch": 1.4575526553818323,
+      "grad_norm": 0.6193887591362,
+      "learning_rate": 2.6110902965882383e-05,
+      "loss": 1.1459,
+      "num_input_tokens_seen": 92405760,
+      "step": 11280
+    },
+    {
+      "epoch": 1.4588448119912134,
+      "grad_norm": 0.9331411719322205,
+      "learning_rate": 2.607710679316891e-05,
+      "loss": 0.7492,
+      "num_input_tokens_seen": 92487680,
+      "step": 11290
+    },
+    {
+      "epoch": 1.4601369686005943,
+      "grad_norm": 0.6176254153251648,
+      "learning_rate": 2.6043308648276833e-05,
+      "loss": 0.9672,
+      "num_input_tokens_seen": 92569600,
+      "step": 11300
+    },
+    {
+      "epoch": 1.4614291252099754,
+      "grad_norm": 0.836600661277771,
+      "learning_rate": 2.6009508593090448e-05,
+      "loss": 0.606,
+      "num_input_tokens_seen": 92651520,
+      "step": 11310
+    },
+    {
+      "epoch": 1.4627212818193565,
+      "grad_norm": 0.6364132165908813,
+      "learning_rate": 2.5975706689497513e-05,
+      "loss": 0.7642,
+      "num_input_tokens_seen": 92733440,
+      "step": 11320
+    },
+    {
+      "epoch": 1.4640134384287375,
+      "grad_norm": 0.6128113865852356,
+      "learning_rate": 2.59419029993892e-05,
+      "loss": 0.8852,
+      "num_input_tokens_seen": 92815360,
+      "step": 11330
+    },
+    {
+      "epoch": 1.4653055950381186,
+      "grad_norm": 0.6446326375007629,
+      "learning_rate": 2.590809758465995e-05,
+      "loss": 1.1366,
+      "num_input_tokens_seen": 92897280,
+      "step": 11340
+    },
+    {
+      "epoch": 1.4665977516474997,
+      "grad_norm": 0.9294217824935913,
+      "learning_rate": 2.5874290507207337e-05,
+      "loss": 1.004,
+      "num_input_tokens_seen": 92979200,
+      "step": 11350
+    },
+    {
+      "epoch": 1.4678899082568808,
+      "grad_norm": 0.5920690298080444,
+      "learning_rate": 2.584048182893201e-05,
+      "loss": 1.1198,
+      "num_input_tokens_seen": 93061120,
+      "step": 11360
+    },
+    {
+      "epoch": 1.4691820648662617,
+      "grad_norm": 0.5159122943878174,
+      "learning_rate": 2.580667161173753e-05,
+      "loss": 0.7811,
+      "num_input_tokens_seen": 93143040,
+      "step": 11370
+    },
+    {
+      "epoch": 1.4704742214756428,
+      "grad_norm": 0.8452631235122681,
+      "learning_rate": 2.577285991753028e-05,
+      "loss": 1.0665,
+      "num_input_tokens_seen": 93224960,
+      "step": 11380
+    },
+    {
+      "epoch": 1.4717663780850239,
+      "grad_norm": 0.6538841724395752,
+      "learning_rate": 2.5739046808219348e-05,
+      "loss": 0.9232,
+      "num_input_tokens_seen": 93306880,
+      "step": 11390
+    },
+    {
+      "epoch": 1.473058534694405,
+      "grad_norm": 0.7873169779777527,
+      "learning_rate": 2.570523234571642e-05,
+      "loss": 0.8771,
+      "num_input_tokens_seen": 93388800,
+      "step": 11400
+    },
+    {
+      "epoch": 1.474350691303786,
+      "grad_norm": 0.36491021513938904,
+      "learning_rate": 2.5671416591935636e-05,
+      "loss": 0.7952,
+      "num_input_tokens_seen": 93470720,
+      "step": 11410
+    },
+    {
+      "epoch": 1.4756428479131671,
+      "grad_norm": 0.5437069535255432,
+      "learning_rate": 2.563759960879354e-05,
+      "loss": 0.8034,
+      "num_input_tokens_seen": 93552640,
+      "step": 11420
+    },
+    {
+      "epoch": 1.4769350045225482,
+      "grad_norm": 0.3457854986190796,
+      "learning_rate": 2.5603781458208885e-05,
+      "loss": 0.9861,
+      "num_input_tokens_seen": 93634560,
+      "step": 11430
+    },
+    {
+      "epoch": 1.478227161131929,
+      "grad_norm": 0.6639708280563354,
+      "learning_rate": 2.55699622021026e-05,
+      "loss": 0.9154,
+      "num_input_tokens_seen": 93716480,
+      "step": 11440
+    },
+    {
+      "epoch": 1.4795193177413102,
+      "grad_norm": 0.5724432468414307,
+      "learning_rate": 2.55361419023976e-05,
+      "loss": 0.696,
+      "num_input_tokens_seen": 93798400,
+      "step": 11450
+    },
+    {
+      "epoch": 1.4808114743506913,
+      "grad_norm": 0.8275769948959351,
+      "learning_rate": 2.5502320621018732e-05,
+      "loss": 1.0897,
+      "num_input_tokens_seen": 93880320,
+      "step": 11460
+    },
+    {
+      "epoch": 1.4821036309600724,
+      "grad_norm": 0.6131581664085388,
+      "learning_rate": 2.5468498419892656e-05,
+      "loss": 0.9171,
+      "num_input_tokens_seen": 93962240,
+      "step": 11470
+    },
+    {
+      "epoch": 1.4833957875694535,
+      "grad_norm": 1.2641193866729736,
+      "learning_rate": 2.5434675360947692e-05,
+      "loss": 0.875,
+      "num_input_tokens_seen": 94044160,
+      "step": 11480
+    },
+    {
+      "epoch": 1.4846879441788345,
+      "grad_norm": 0.8318386673927307,
+      "learning_rate": 2.5400851506113728e-05,
+      "loss": 0.7646,
+      "num_input_tokens_seen": 94126080,
+      "step": 11490
+    },
+    {
+      "epoch": 1.4859801007882156,
+      "grad_norm": 0.6216618418693542,
+      "learning_rate": 2.5367026917322117e-05,
+      "loss": 0.9129,
+      "num_input_tokens_seen": 94208000,
+      "step": 11500
+    },
+    {
+      "epoch": 1.4872722573975965,
+      "grad_norm": 0.6390239000320435,
+      "learning_rate": 2.5333201656505567e-05,
+      "loss": 1.2751,
+      "num_input_tokens_seen": 94289920,
+      "step": 11510
+    },
+    {
+      "epoch": 1.4885644140069776,
+      "grad_norm": 0.8883216381072998,
+      "learning_rate": 2.5299375785598005e-05,
+      "loss": 0.8298,
+      "num_input_tokens_seen": 94371840,
+      "step": 11520
+    },
+    {
+      "epoch": 1.4898565706163587,
+      "grad_norm": 0.4036071300506592,
+      "learning_rate": 2.5265549366534475e-05,
+      "loss": 1.0023,
+      "num_input_tokens_seen": 94453760,
+      "step": 11530
+    },
+    {
+      "epoch": 1.4911487272257398,
+      "grad_norm": 0.9503811597824097,
+      "learning_rate": 2.5231722461251017e-05,
+      "loss": 0.8267,
+      "num_input_tokens_seen": 94535680,
+      "step": 11540
+    },
+    {
+      "epoch": 1.4924408838351209,
+      "grad_norm": 0.9378893971443176,
+      "learning_rate": 2.519789513168459e-05,
+      "loss": 0.6509,
+      "num_input_tokens_seen": 94617600,
+      "step": 11550
+    },
+    {
+      "epoch": 1.493733040444502,
+      "grad_norm": 0.6294723153114319,
+      "learning_rate": 2.5164067439772898e-05,
+      "loss": 0.7988,
+      "num_input_tokens_seen": 94699520,
+      "step": 11560
+    },
+    {
+      "epoch": 1.495025197053883,
+      "grad_norm": 2.2070484161376953,
+      "learning_rate": 2.5130239447454328e-05,
+      "loss": 0.7822,
+      "num_input_tokens_seen": 94781440,
+      "step": 11570
+    },
+    {
+      "epoch": 1.496317353663264,
+      "grad_norm": 0.5767642259597778,
+      "learning_rate": 2.509641121666781e-05,
+      "loss": 0.9319,
+      "num_input_tokens_seen": 94863360,
+      "step": 11580
+    },
+    {
+      "epoch": 1.497609510272645,
+      "grad_norm": 0.5216675400733948,
+      "learning_rate": 2.5062582809352704e-05,
+      "loss": 0.8723,
+      "num_input_tokens_seen": 94945280,
+      "step": 11590
+    },
+    {
+      "epoch": 1.498901666882026,
+      "grad_norm": 0.5917890667915344,
+      "learning_rate": 2.5028754287448695e-05,
+      "loss": 1.0985,
+      "num_input_tokens_seen": 95027200,
+      "step": 11600
+    },
+    {
+      "epoch": 1.5001938234914072,
+      "grad_norm": 0.5695708394050598,
+      "learning_rate": 2.4994925712895697e-05,
+      "loss": 0.898,
+      "num_input_tokens_seen": 95109120,
+      "step": 11610
+    },
+    {
+      "epoch": 1.5014859801007883,
+      "grad_norm": 0.6653333902359009,
+      "learning_rate": 2.4961097147633698e-05,
+      "loss": 1.2631,
+      "num_input_tokens_seen": 95191040,
+      "step": 11620
+    },
+    {
+      "epoch": 1.5027781367101691,
+      "grad_norm": 0.7980930209159851,
+      "learning_rate": 2.4927268653602684e-05,
+      "loss": 0.9865,
+      "num_input_tokens_seen": 95272960,
+      "step": 11630
+    },
+    {
+      "epoch": 1.5040702933195504,
+      "grad_norm": 0.27931126952171326,
+      "learning_rate": 2.489344029274249e-05,
+      "loss": 1.063,
+      "num_input_tokens_seen": 95354880,
+      "step": 11640
+    },
+    {
+      "epoch": 1.5053624499289313,
+      "grad_norm": 0.4017459452152252,
+      "learning_rate": 2.4859612126992737e-05,
+      "loss": 0.4969,
+      "num_input_tokens_seen": 95436800,
+      "step": 11650
+    },
+    {
+      "epoch": 1.5066546065383124,
+      "grad_norm": 0.9155080318450928,
+      "learning_rate": 2.4825784218292664e-05,
+      "loss": 0.6354,
+      "num_input_tokens_seen": 95518720,
+      "step": 11660
+    },
+    {
+      "epoch": 1.5079467631476935,
+      "grad_norm": 0.7411128282546997,
+      "learning_rate": 2.479195662858105e-05,
+      "loss": 0.9183,
+      "num_input_tokens_seen": 95600640,
+      "step": 11670
+    },
+    {
+      "epoch": 1.5092389197570746,
+      "grad_norm": 1.0255645513534546,
+      "learning_rate": 2.4758129419796094e-05,
+      "loss": 0.8669,
+      "num_input_tokens_seen": 95682560,
+      "step": 11680
+    },
+    {
+      "epoch": 1.5105310763664557,
+      "grad_norm": 0.612771213054657,
+      "learning_rate": 2.4724302653875275e-05,
+      "loss": 1.1856,
+      "num_input_tokens_seen": 95764480,
+      "step": 11690
+    },
+    {
+      "epoch": 1.5118232329758365,
+      "grad_norm": 0.7555944919586182,
+      "learning_rate": 2.4690476392755298e-05,
+      "loss": 0.7345,
+      "num_input_tokens_seen": 95846400,
+      "step": 11700
+    },
+    {
+      "epoch": 1.5131153895852179,
+      "grad_norm": 0.7884678244590759,
+      "learning_rate": 2.4656650698371903e-05,
+      "loss": 0.5009,
+      "num_input_tokens_seen": 95928320,
+      "step": 11710
+    },
+    {
+      "epoch": 1.5144075461945987,
+      "grad_norm": 0.5397049188613892,
+      "learning_rate": 2.462282563265982e-05,
+      "loss": 0.7891,
+      "num_input_tokens_seen": 96010240,
+      "step": 11720
+    },
+    {
+      "epoch": 1.5156997028039798,
+      "grad_norm": 0.6234694123268127,
+      "learning_rate": 2.4589001257552637e-05,
+      "loss": 0.6393,
+      "num_input_tokens_seen": 96092160,
+      "step": 11730
+    },
+    {
+      "epoch": 1.516991859413361,
+      "grad_norm": 0.6878401637077332,
+      "learning_rate": 2.455517763498264e-05,
+      "loss": 0.8309,
+      "num_input_tokens_seen": 96174080,
+      "step": 11740
+    },
+    {
+      "epoch": 1.518284016022742,
+      "grad_norm": 0.23810110986232758,
+      "learning_rate": 2.452135482688077e-05,
+      "loss": 0.7155,
+      "num_input_tokens_seen": 96256000,
+      "step": 11750
+    },
+    {
+      "epoch": 1.519576172632123,
+      "grad_norm": 0.5660243630409241,
+      "learning_rate": 2.4487532895176457e-05,
+      "loss": 0.8101,
+      "num_input_tokens_seen": 96337920,
+      "step": 11760
+    },
+    {
+      "epoch": 1.520868329241504,
+      "grad_norm": 0.9258086085319519,
+      "learning_rate": 2.4453711901797543e-05,
+      "loss": 0.5187,
+      "num_input_tokens_seen": 96419840,
+      "step": 11770
+    },
+    {
+      "epoch": 1.5221604858508853,
+      "grad_norm": 0.5753411054611206,
+      "learning_rate": 2.4419891908670127e-05,
+      "loss": 1.1635,
+      "num_input_tokens_seen": 96501760,
+      "step": 11780
+    },
+    {
+      "epoch": 1.5234526424602661,
+      "grad_norm": 0.6174433827400208,
+      "learning_rate": 2.4386072977718503e-05,
+      "loss": 0.8433,
+      "num_input_tokens_seen": 96583680,
+      "step": 11790
+    },
+    {
+      "epoch": 1.5247447990696472,
+      "grad_norm": 0.5415393710136414,
+      "learning_rate": 2.4352255170865025e-05,
+      "loss": 0.9885,
+      "num_input_tokens_seen": 96665600,
+      "step": 11800
+    },
+    {
+      "epoch": 1.5260369556790283,
+      "grad_norm": 0.5386336445808411,
+      "learning_rate": 2.4318438550029946e-05,
+      "loss": 1.0425,
+      "num_input_tokens_seen": 96747520,
+      "step": 11810
+    },
+    {
+      "epoch": 1.5273291122884094,
+      "grad_norm": 0.721867561340332,
+      "learning_rate": 2.4284623177131395e-05,
+      "loss": 0.6342,
+      "num_input_tokens_seen": 96829440,
+      "step": 11820
+    },
+    {
+      "epoch": 1.5286212688977905,
+      "grad_norm": 0.5259259343147278,
+      "learning_rate": 2.4250809114085183e-05,
+      "loss": 1.1709,
+      "num_input_tokens_seen": 96911360,
+      "step": 11830
+    },
+    {
+      "epoch": 1.5299134255071714,
+      "grad_norm": 0.7975210547447205,
+      "learning_rate": 2.421699642280475e-05,
+      "loss": 1.0094,
+      "num_input_tokens_seen": 96993280,
+      "step": 11840
+    },
+    {
+      "epoch": 1.5312055821165527,
+      "grad_norm": 0.5395787954330444,
+      "learning_rate": 2.4183185165200998e-05,
+      "loss": 0.7204,
+      "num_input_tokens_seen": 97075200,
+      "step": 11850
+    },
+    {
+      "epoch": 1.5324977387259335,
+      "grad_norm": 0.6701288223266602,
+      "learning_rate": 2.4149375403182216e-05,
+      "loss": 0.7901,
+      "num_input_tokens_seen": 97157120,
+      "step": 11860
+    },
+    {
+      "epoch": 1.5337898953353146,
+      "grad_norm": 0.7676398158073425,
+      "learning_rate": 2.4115567198653963e-05,
+      "loss": 1.0571,
+      "num_input_tokens_seen": 97239040,
+      "step": 11870
+    },
+    {
+      "epoch": 1.5350820519446957,
+      "grad_norm": 0.38611501455307007,
+      "learning_rate": 2.4081760613518924e-05,
+      "loss": 0.5656,
+      "num_input_tokens_seen": 97320960,
+      "step": 11880
+    },
+    {
+      "epoch": 1.5363742085540768,
+      "grad_norm": 0.5478717088699341,
+      "learning_rate": 2.4047955709676852e-05,
+      "loss": 1.2245,
+      "num_input_tokens_seen": 97402880,
+      "step": 11890
+    },
+    {
+      "epoch": 1.537666365163458,
+      "grad_norm": 0.5873267650604248,
+      "learning_rate": 2.401415254902438e-05,
+      "loss": 0.9675,
+      "num_input_tokens_seen": 97484800,
+      "step": 11900
+    },
+    {
+      "epoch": 1.5389585217728388,
+      "grad_norm": 0.5424181222915649,
+      "learning_rate": 2.3980351193455e-05,
+      "loss": 1.0065,
+      "num_input_tokens_seen": 97566720,
+      "step": 11910
+    },
+    {
+      "epoch": 1.54025067838222,
+      "grad_norm": 0.47638383507728577,
+      "learning_rate": 2.3946551704858838e-05,
+      "loss": 0.8522,
+      "num_input_tokens_seen": 97648640,
+      "step": 11920
+    },
+    {
+      "epoch": 1.541542834991601,
+      "grad_norm": 0.6472703218460083,
+      "learning_rate": 2.3912754145122663e-05,
+      "loss": 0.9128,
+      "num_input_tokens_seen": 97730560,
+      "step": 11930
+    },
+    {
+      "epoch": 1.542834991600982,
+      "grad_norm": 0.6344046592712402,
+      "learning_rate": 2.3878958576129664e-05,
+      "loss": 0.8387,
+      "num_input_tokens_seen": 97812480,
+      "step": 11940
+    },
+    {
+      "epoch": 1.5441271482103631,
+      "grad_norm": 0.8018627762794495,
+      "learning_rate": 2.3845165059759402e-05,
+      "loss": 0.9149,
+      "num_input_tokens_seen": 97894400,
+      "step": 11950
+    },
+    {
+      "epoch": 1.545419304819744,
+      "grad_norm": 0.398837685585022,
+      "learning_rate": 2.3811373657887705e-05,
+      "loss": 0.8989,
+      "num_input_tokens_seen": 97976320,
+      "step": 11960
+    },
+    {
+      "epoch": 1.5467114614291253,
+      "grad_norm": 0.6712442636489868,
+      "learning_rate": 2.3777584432386474e-05,
+      "loss": 0.8954,
+      "num_input_tokens_seen": 98058240,
+      "step": 11970
+    },
+    {
+      "epoch": 1.5480036180385062,
+      "grad_norm": 0.685006856918335,
+      "learning_rate": 2.3743797445123688e-05,
+      "loss": 0.9203,
+      "num_input_tokens_seen": 98140160,
+      "step": 11980
+    },
+    {
+      "epoch": 1.5492957746478875,
+      "grad_norm": 1.1985323429107666,
+      "learning_rate": 2.3710012757963175e-05,
+      "loss": 1.0436,
+      "num_input_tokens_seen": 98222080,
+      "step": 11990
+    },
+    {
+      "epoch": 1.5505879312572683,
+      "grad_norm": 0.6150766015052795,
+      "learning_rate": 2.367623043276459e-05,
+      "loss": 1.0798,
+      "num_input_tokens_seen": 98304000,
+      "step": 12000
+    },
+    {
+      "epoch": 1.5518800878666494,
+      "grad_norm": 0.5276364684104919,
+      "learning_rate": 2.364245053138323e-05,
+      "loss": 1.0669,
+      "num_input_tokens_seen": 98385920,
+      "step": 12010
+    },
+    {
+      "epoch": 1.5531722444760305,
+      "grad_norm": 0.291031152009964,
+      "learning_rate": 2.3608673115669978e-05,
+      "loss": 0.8337,
+      "num_input_tokens_seen": 98467840,
+      "step": 12020
+    },
+    {
+      "epoch": 1.5544644010854114,
+      "grad_norm": 0.4851202964782715,
+      "learning_rate": 2.3574898247471167e-05,
+      "loss": 1.2425,
+      "num_input_tokens_seen": 98549760,
+      "step": 12030
+    },
+    {
+      "epoch": 1.5557565576947927,
+      "grad_norm": 0.6985998153686523,
+      "learning_rate": 2.354112598862845e-05,
+      "loss": 0.9847,
+      "num_input_tokens_seen": 98631680,
+      "step": 12040
+    },
+    {
+      "epoch": 1.5570487143041736,
+      "grad_norm": 0.26096710562705994,
+      "learning_rate": 2.350735640097871e-05,
+      "loss": 0.79,
+      "num_input_tokens_seen": 98713600,
+      "step": 12050
+    },
+    {
+      "epoch": 1.5583408709135549,
+      "grad_norm": 0.830028772354126,
+      "learning_rate": 2.347358954635393e-05,
+      "loss": 0.8655,
+      "num_input_tokens_seen": 98795520,
+      "step": 12060
+    },
+    {
+      "epoch": 1.5596330275229358,
+      "grad_norm": 0.9304326772689819,
+      "learning_rate": 2.3439825486581116e-05,
+      "loss": 0.7855,
+      "num_input_tokens_seen": 98877440,
+      "step": 12070
+    },
+    {
+      "epoch": 1.5609251841323168,
+      "grad_norm": 0.505251407623291,
+      "learning_rate": 2.3406064283482115e-05,
+      "loss": 0.911,
+      "num_input_tokens_seen": 98959360,
+      "step": 12080
+    },
+    {
+      "epoch": 1.562217340741698,
+      "grad_norm": 0.6220946907997131,
+      "learning_rate": 2.337230599887358e-05,
+      "loss": 0.6987,
+      "num_input_tokens_seen": 99041280,
+      "step": 12090
+    },
+    {
+      "epoch": 1.5635094973510788,
+      "grad_norm": 1.1672903299331665,
+      "learning_rate": 2.3338550694566817e-05,
+      "loss": 0.7693,
+      "num_input_tokens_seen": 99123200,
+      "step": 12100
+    },
+    {
+      "epoch": 1.5648016539604601,
+      "grad_norm": 0.5260895490646362,
+      "learning_rate": 2.3304798432367645e-05,
+      "loss": 0.5875,
+      "num_input_tokens_seen": 99205120,
+      "step": 12110
+    },
+    {
+      "epoch": 1.566093810569841,
+      "grad_norm": 0.7243925333023071,
+      "learning_rate": 2.327104927407634e-05,
+      "loss": 1.1252,
+      "num_input_tokens_seen": 99287040,
+      "step": 12120
+    },
+    {
+      "epoch": 1.567385967179222,
+      "grad_norm": 0.5413958430290222,
+      "learning_rate": 2.3237303281487487e-05,
+      "loss": 0.8912,
+      "num_input_tokens_seen": 99368960,
+      "step": 12130
+    },
+    {
+      "epoch": 1.5686781237886032,
+      "grad_norm": 0.6078057885169983,
+      "learning_rate": 2.3203560516389882e-05,
+      "loss": 0.6356,
+      "num_input_tokens_seen": 99450880,
+      "step": 12140
+    },
+    {
+      "epoch": 1.5699702803979843,
+      "grad_norm": 0.9383816719055176,
+      "learning_rate": 2.3169821040566387e-05,
+      "loss": 1.1298,
+      "num_input_tokens_seen": 99532800,
+      "step": 12150
+    },
+    {
+      "epoch": 1.5712624370073653,
+      "grad_norm": 0.6377266049385071,
+      "learning_rate": 2.313608491579387e-05,
+      "loss": 1.0096,
+      "num_input_tokens_seen": 99614720,
+      "step": 12160
+    },
+    {
+      "epoch": 1.5725545936167462,
+      "grad_norm": 0.5839458703994751,
+      "learning_rate": 2.3102352203843063e-05,
+      "loss": 0.7456,
+      "num_input_tokens_seen": 99696640,
+      "step": 12170
+    },
+    {
+      "epoch": 1.5738467502261275,
+      "grad_norm": 0.5719941854476929,
+      "learning_rate": 2.306862296647841e-05,
+      "loss": 0.8698,
+      "num_input_tokens_seen": 99778560,
+      "step": 12180
+    },
+    {
+      "epoch": 1.5751389068355084,
+      "grad_norm": 0.7049387693405151,
+      "learning_rate": 2.3034897265458056e-05,
+      "loss": 0.6883,
+      "num_input_tokens_seen": 99860480,
+      "step": 12190
+    },
+    {
+      "epoch": 1.5764310634448895,
+      "grad_norm": 0.5691668391227722,
+      "learning_rate": 2.3001175162533606e-05,
+      "loss": 0.8952,
+      "num_input_tokens_seen": 99942400,
+      "step": 12200
+    },
+    {
+      "epoch": 1.5777232200542706,
+      "grad_norm": 0.9318535327911377,
+      "learning_rate": 2.2967456719450127e-05,
+      "loss": 0.7274,
+      "num_input_tokens_seen": 100024320,
+      "step": 12210
+    },
+    {
+      "epoch": 1.5790153766636517,
+      "grad_norm": 0.4153442680835724,
+      "learning_rate": 2.2933741997945954e-05,
+      "loss": 0.8773,
+      "num_input_tokens_seen": 100106240,
+      "step": 12220
+    },
+    {
+      "epoch": 1.5803075332730327,
+      "grad_norm": 0.7407699227333069,
+      "learning_rate": 2.290003105975262e-05,
+      "loss": 0.9009,
+      "num_input_tokens_seen": 100188160,
+      "step": 12230
+    },
+    {
+      "epoch": 1.5815996898824136,
+      "grad_norm": 0.6650556325912476,
+      "learning_rate": 2.2866323966594736e-05,
+      "loss": 0.6566,
+      "num_input_tokens_seen": 100270080,
+      "step": 12240
+    },
+    {
+      "epoch": 1.582891846491795,
+      "grad_norm": 0.2326425462961197,
+      "learning_rate": 2.283262078018985e-05,
+      "loss": 0.7216,
+      "num_input_tokens_seen": 100352000,
+      "step": 12250
+    },
+    {
+      "epoch": 1.5841840031011758,
+      "grad_norm": 0.9894224405288696,
+      "learning_rate": 2.27989215622484e-05,
+      "loss": 0.9039,
+      "num_input_tokens_seen": 100433920,
+      "step": 12260
+    },
+    {
+      "epoch": 1.5854761597105569,
+      "grad_norm": 0.6276862621307373,
+      "learning_rate": 2.2765226374473504e-05,
+      "loss": 0.9027,
+      "num_input_tokens_seen": 100515840,
+      "step": 12270
+    },
+    {
+      "epoch": 1.586768316319938,
+      "grad_norm": 0.33717378973960876,
+      "learning_rate": 2.2731535278560944e-05,
+      "loss": 0.7404,
+      "num_input_tokens_seen": 100597760,
+      "step": 12280
+    },
+    {
+      "epoch": 1.588060472929319,
+      "grad_norm": 0.8220341205596924,
+      "learning_rate": 2.269784833619898e-05,
+      "loss": 0.8567,
+      "num_input_tokens_seen": 100679680,
+      "step": 12290
+    },
+    {
+      "epoch": 1.5893526295387002,
+      "grad_norm": 0.7151164412498474,
+      "learning_rate": 2.2664165609068304e-05,
+      "loss": 0.9177,
+      "num_input_tokens_seen": 100761600,
+      "step": 12300
+    },
+    {
+      "epoch": 1.590644786148081,
+      "grad_norm": 0.5890063047409058,
+      "learning_rate": 2.263048715884184e-05,
+      "loss": 0.8511,
+      "num_input_tokens_seen": 100843520,
+      "step": 12310
+    },
+    {
+      "epoch": 1.5919369427574623,
+      "grad_norm": 0.919540286064148,
+      "learning_rate": 2.2596813047184715e-05,
+      "loss": 0.8876,
+      "num_input_tokens_seen": 100925440,
+      "step": 12320
+    },
+    {
+      "epoch": 1.5932290993668432,
+      "grad_norm": 0.8143252730369568,
+      "learning_rate": 2.2563143335754118e-05,
+      "loss": 0.9139,
+      "num_input_tokens_seen": 101007360,
+      "step": 12330
+    },
+    {
+      "epoch": 1.5945212559762243,
+      "grad_norm": 0.48788896203041077,
+      "learning_rate": 2.252947808619914e-05,
+      "loss": 0.5727,
+      "num_input_tokens_seen": 101089280,
+      "step": 12340
+    },
+    {
+      "epoch": 1.5958134125856054,
+      "grad_norm": 0.5264637470245361,
+      "learning_rate": 2.249581736016076e-05,
+      "loss": 0.7672,
+      "num_input_tokens_seen": 101171200,
+      "step": 12350
+    },
+    {
+      "epoch": 1.5971055691949865,
+      "grad_norm": 0.5975572466850281,
+      "learning_rate": 2.2462161219271622e-05,
+      "loss": 0.8705,
+      "num_input_tokens_seen": 101253120,
+      "step": 12360
+    },
+    {
+      "epoch": 1.5983977258043676,
+      "grad_norm": 0.2425682097673416,
+      "learning_rate": 2.242850972515601e-05,
+      "loss": 0.8957,
+      "num_input_tokens_seen": 101335040,
+      "step": 12370
+    },
+    {
+      "epoch": 1.5996898824137484,
+      "grad_norm": 0.5979612469673157,
+      "learning_rate": 2.2394862939429677e-05,
+      "loss": 1.0392,
+      "num_input_tokens_seen": 101416960,
+      "step": 12380
+    },
+    {
+      "epoch": 1.6009820390231297,
+      "grad_norm": 0.6197012066841125,
+      "learning_rate": 2.236122092369977e-05,
+      "loss": 0.7616,
+      "num_input_tokens_seen": 101498880,
+      "step": 12390
+    },
+    {
+      "epoch": 1.6022741956325106,
+      "grad_norm": 0.734671413898468,
+      "learning_rate": 2.2327583739564696e-05,
+      "loss": 1.1416,
+      "num_input_tokens_seen": 101580800,
+      "step": 12400
+    },
+    {
+      "epoch": 1.6035663522418917,
+      "grad_norm": 0.4598730504512787,
+      "learning_rate": 2.229395144861402e-05,
+      "loss": 0.8276,
+      "num_input_tokens_seen": 101662720,
+      "step": 12410
+    },
+    {
+      "epoch": 1.6048585088512728,
+      "grad_norm": 0.7918890714645386,
+      "learning_rate": 2.2260324112428336e-05,
+      "loss": 1.044,
+      "num_input_tokens_seen": 101744640,
+      "step": 12420
+    },
+    {
+      "epoch": 1.6061506654606539,
+      "grad_norm": 1.4990414381027222,
+      "learning_rate": 2.2226701792579176e-05,
+      "loss": 0.5725,
+      "num_input_tokens_seen": 101826560,
+      "step": 12430
+    },
+    {
+      "epoch": 1.607442822070035,
+      "grad_norm": 0.7049417495727539,
+      "learning_rate": 2.219308455062889e-05,
+      "loss": 0.8819,
+      "num_input_tokens_seen": 101908480,
+      "step": 12440
+    },
+    {
+      "epoch": 1.6087349786794158,
+      "grad_norm": 1.0287914276123047,
+      "learning_rate": 2.2159472448130513e-05,
+      "loss": 0.6064,
+      "num_input_tokens_seen": 101990400,
+      "step": 12450
+    },
+    {
+      "epoch": 1.6100271352887972,
+      "grad_norm": 0.6095555424690247,
+      "learning_rate": 2.212586554662769e-05,
+      "loss": 0.7968,
+      "num_input_tokens_seen": 102072320,
+      "step": 12460
+    },
+    {
+      "epoch": 1.611319291898178,
+      "grad_norm": 0.7661923170089722,
+      "learning_rate": 2.2092263907654544e-05,
+      "loss": 0.6481,
+      "num_input_tokens_seen": 102154240,
+      "step": 12470
+    },
+    {
+      "epoch": 1.612611448507559,
+      "grad_norm": 0.6737602353096008,
+      "learning_rate": 2.2058667592735532e-05,
+      "loss": 0.722,
+      "num_input_tokens_seen": 102236160,
+      "step": 12480
+    },
+    {
+      "epoch": 1.6139036051169402,
+      "grad_norm": 0.5054346919059753,
+      "learning_rate": 2.20250766633854e-05,
+      "loss": 1.321,
+      "num_input_tokens_seen": 102318080,
+      "step": 12490
+    },
+    {
+      "epoch": 1.615195761726321,
+      "grad_norm": 0.5641369819641113,
+      "learning_rate": 2.199149118110901e-05,
+      "loss": 0.6964,
+      "num_input_tokens_seen": 102400000,
+      "step": 12500
+    },
+    {
+      "epoch": 1.6164879183357024,
+      "grad_norm": 0.27639254927635193,
+      "learning_rate": 2.1957911207401267e-05,
+      "loss": 0.965,
+      "num_input_tokens_seen": 102481920,
+      "step": 12510
+    },
+    {
+      "epoch": 1.6177800749450832,
+      "grad_norm": 0.8085689544677734,
+      "learning_rate": 2.192433680374696e-05,
+      "loss": 0.6768,
+      "num_input_tokens_seen": 102563840,
+      "step": 12520
+    },
+    {
+      "epoch": 1.6190722315544646,
+      "grad_norm": 0.7601416110992432,
+      "learning_rate": 2.1890768031620705e-05,
+      "loss": 0.975,
+      "num_input_tokens_seen": 102645760,
+      "step": 12530
+    },
+    {
+      "epoch": 1.6203643881638454,
+      "grad_norm": 0.615585207939148,
+      "learning_rate": 2.1857204952486824e-05,
+      "loss": 1.0603,
+      "num_input_tokens_seen": 102727680,
+      "step": 12540
+    },
+    {
+      "epoch": 1.6216565447732265,
+      "grad_norm": 0.7139294147491455,
+      "learning_rate": 2.182364762779916e-05,
+      "loss": 0.737,
+      "num_input_tokens_seen": 102809600,
+      "step": 12550
+    },
+    {
+      "epoch": 1.6229487013826076,
+      "grad_norm": 0.38794225454330444,
+      "learning_rate": 2.1790096119001077e-05,
+      "loss": 0.8256,
+      "num_input_tokens_seen": 102891520,
+      "step": 12560
+    },
+    {
+      "epoch": 1.6242408579919885,
+      "grad_norm": 0.8240430951118469,
+      "learning_rate": 2.1756550487525247e-05,
+      "loss": 1.0425,
+      "num_input_tokens_seen": 102973440,
+      "step": 12570
+    },
+    {
+      "epoch": 1.6255330146013698,
+      "grad_norm": 0.7485008239746094,
+      "learning_rate": 2.1723010794793612e-05,
+      "loss": 0.6268,
+      "num_input_tokens_seen": 103055360,
+      "step": 12580
+    },
+    {
+      "epoch": 1.6268251712107507,
+      "grad_norm": 0.5244811177253723,
+      "learning_rate": 2.168947710221722e-05,
+      "loss": 0.9757,
+      "num_input_tokens_seen": 103137280,
+      "step": 12590
+    },
+    {
+      "epoch": 1.628117327820132,
+      "grad_norm": 0.6735374927520752,
+      "learning_rate": 2.165594947119613e-05,
+      "loss": 0.7146,
+      "num_input_tokens_seen": 103219200,
+      "step": 12600
+    },
+    {
+      "epoch": 1.6294094844295128,
+      "grad_norm": 0.7091389894485474,
+      "learning_rate": 2.1622427963119337e-05,
+      "loss": 1.4099,
+      "num_input_tokens_seen": 103301120,
+      "step": 12610
+    },
+    {
+      "epoch": 1.630701641038894,
+      "grad_norm": 0.21646621823310852,
+      "learning_rate": 2.1588912639364567e-05,
+      "loss": 1.0245,
+      "num_input_tokens_seen": 103383040,
+      "step": 12620
+    },
+    {
+      "epoch": 1.631993797648275,
+      "grad_norm": 0.4036228358745575,
+      "learning_rate": 2.1555403561298287e-05,
+      "loss": 1.1632,
+      "num_input_tokens_seen": 103464960,
+      "step": 12630
+    },
+    {
+      "epoch": 1.6332859542576559,
+      "grad_norm": 0.5068750977516174,
+      "learning_rate": 2.152190079027547e-05,
+      "loss": 0.747,
+      "num_input_tokens_seen": 103546880,
+      "step": 12640
+    },
+    {
+      "epoch": 1.6345781108670372,
+      "grad_norm": 0.657037079334259,
+      "learning_rate": 2.148840438763959e-05,
+      "loss": 1.056,
+      "num_input_tokens_seen": 103628800,
+      "step": 12650
+    },
+    {
+      "epoch": 1.635870267476418,
+      "grad_norm": 0.7527886033058167,
+      "learning_rate": 2.1454914414722417e-05,
+      "loss": 0.9465,
+      "num_input_tokens_seen": 103710720,
+      "step": 12660
+    },
+    {
+      "epoch": 1.6371624240857992,
+      "grad_norm": 0.504432201385498,
+      "learning_rate": 2.1421430932843988e-05,
+      "loss": 1.1174,
+      "num_input_tokens_seen": 103792640,
+      "step": 12670
+    },
+    {
+      "epoch": 1.6384545806951802,
+      "grad_norm": 0.23135700821876526,
+      "learning_rate": 2.138795400331242e-05,
+      "loss": 0.6262,
+      "num_input_tokens_seen": 103874560,
+      "step": 12680
+    },
+    {
+      "epoch": 1.6397467373045613,
+      "grad_norm": 0.5520825982093811,
+      "learning_rate": 2.135448368742385e-05,
+      "loss": 1.066,
+      "num_input_tokens_seen": 103956480,
+      "step": 12690
+    },
+    {
+      "epoch": 1.6410388939139424,
+      "grad_norm": 0.7319164276123047,
+      "learning_rate": 2.1321020046462318e-05,
+      "loss": 0.9554,
+      "num_input_tokens_seen": 104038400,
+      "step": 12700
+    },
+    {
+      "epoch": 1.6423310505233233,
+      "grad_norm": 0.40483906865119934,
+      "learning_rate": 2.128756314169961e-05,
+      "loss": 1.0322,
+      "num_input_tokens_seen": 104120320,
+      "step": 12710
+    },
+    {
+      "epoch": 1.6436232071327046,
+      "grad_norm": 0.683261513710022,
+      "learning_rate": 2.1254113034395212e-05,
+      "loss": 0.6685,
+      "num_input_tokens_seen": 104202240,
+      "step": 12720
+    },
+    {
+      "epoch": 1.6449153637420855,
+      "grad_norm": 0.6625344157218933,
+      "learning_rate": 2.122066978579613e-05,
+      "loss": 0.7232,
+      "num_input_tokens_seen": 104284160,
+      "step": 12730
+    },
+    {
+      "epoch": 1.6462075203514666,
+      "grad_norm": 0.5251320600509644,
+      "learning_rate": 2.1187233457136858e-05,
+      "loss": 0.8379,
+      "num_input_tokens_seen": 104366080,
+      "step": 12740
+    },
+    {
+      "epoch": 1.6474996769608476,
+      "grad_norm": 0.37905430793762207,
+      "learning_rate": 2.1153804109639157e-05,
+      "loss": 0.7044,
+      "num_input_tokens_seen": 104448000,
+      "step": 12750
+    },
+    {
+      "epoch": 1.6487918335702287,
+      "grad_norm": 0.7667315006256104,
+      "learning_rate": 2.1120381804512066e-05,
+      "loss": 0.7293,
+      "num_input_tokens_seen": 104529920,
+      "step": 12760
+    },
+    {
+      "epoch": 1.6500839901796098,
+      "grad_norm": 0.8801140785217285,
+      "learning_rate": 2.1086966602951696e-05,
+      "loss": 0.9354,
+      "num_input_tokens_seen": 104611840,
+      "step": 12770
+    },
+    {
+      "epoch": 1.6513761467889907,
+      "grad_norm": 0.6519145369529724,
+      "learning_rate": 2.105355856614115e-05,
+      "loss": 0.699,
+      "num_input_tokens_seen": 104693760,
+      "step": 12780
+    },
+    {
+      "epoch": 1.652668303398372,
+      "grad_norm": 0.5816752910614014,
+      "learning_rate": 2.1020157755250437e-05,
+      "loss": 0.6137,
+      "num_input_tokens_seen": 104775680,
+      "step": 12790
+    },
+    {
+      "epoch": 1.6539604600077529,
+      "grad_norm": 0.8572224378585815,
+      "learning_rate": 2.09867642314363e-05,
+      "loss": 0.8616,
+      "num_input_tokens_seen": 104857600,
+      "step": 12800
+    },
+    {
+      "epoch": 1.655252616617134,
+      "grad_norm": 0.367939293384552,
+      "learning_rate": 2.0953378055842183e-05,
+      "loss": 0.668,
+      "num_input_tokens_seen": 104939520,
+      "step": 12810
+    },
+    {
+      "epoch": 1.656544773226515,
+      "grad_norm": 0.5176908373832703,
+      "learning_rate": 2.0919999289598027e-05,
+      "loss": 0.961,
+      "num_input_tokens_seen": 105021440,
+      "step": 12820
+    },
+    {
+      "epoch": 1.6578369298358961,
+      "grad_norm": 0.21786247193813324,
+      "learning_rate": 2.088662799382024e-05,
+      "loss": 0.6605,
+      "num_input_tokens_seen": 105103360,
+      "step": 12830
+    },
+    {
+      "epoch": 1.6591290864452772,
+      "grad_norm": 0.7281939387321472,
+      "learning_rate": 2.0853264229611557e-05,
+      "loss": 1.176,
+      "num_input_tokens_seen": 105185280,
+      "step": 12840
+    },
+    {
+      "epoch": 1.660421243054658,
+      "grad_norm": 0.6157118678092957,
+      "learning_rate": 2.081990805806089e-05,
+      "loss": 0.9159,
+      "num_input_tokens_seen": 105267200,
+      "step": 12850
+    },
+    {
+      "epoch": 1.6617133996640394,
+      "grad_norm": 0.3668000102043152,
+      "learning_rate": 2.078655954024327e-05,
+      "loss": 0.6293,
+      "num_input_tokens_seen": 105349120,
+      "step": 12860
+    },
+    {
+      "epoch": 1.6630055562734203,
+      "grad_norm": 0.6818766593933105,
+      "learning_rate": 2.075321873721972e-05,
+      "loss": 0.7573,
+      "num_input_tokens_seen": 105431040,
+      "step": 12870
+    },
+    {
+      "epoch": 1.6642977128828014,
+      "grad_norm": 0.6922422647476196,
+      "learning_rate": 2.0719885710037122e-05,
+      "loss": 0.4979,
+      "num_input_tokens_seen": 105512960,
+      "step": 12880
+    },
+    {
+      "epoch": 1.6655898694921825,
+      "grad_norm": 0.6830812692642212,
+      "learning_rate": 2.0686560519728117e-05,
+      "loss": 1.026,
+      "num_input_tokens_seen": 105594880,
+      "step": 12890
+    },
+    {
+      "epoch": 1.6668820261015636,
+      "grad_norm": 0.574186384677887,
+      "learning_rate": 2.0653243227311014e-05,
+      "loss": 0.5754,
+      "num_input_tokens_seen": 105676800,
+      "step": 12900
+    },
+    {
+      "epoch": 1.6681741827109446,
+      "grad_norm": 1.0729289054870605,
+      "learning_rate": 2.0619933893789673e-05,
+      "loss": 0.8647,
+      "num_input_tokens_seen": 105758720,
+      "step": 12910
+    },
+    {
+      "epoch": 1.6694663393203255,
+      "grad_norm": 0.6160525679588318,
+      "learning_rate": 2.0586632580153328e-05,
+      "loss": 1.2784,
+      "num_input_tokens_seen": 105840640,
+      "step": 12920
+    },
+    {
+      "epoch": 1.6707584959297068,
+      "grad_norm": 0.5736418962478638,
+      "learning_rate": 2.0553339347376592e-05,
+      "loss": 0.8836,
+      "num_input_tokens_seen": 105922560,
+      "step": 12930
+    },
+    {
+      "epoch": 1.6720506525390877,
+      "grad_norm": 0.743789553642273,
+      "learning_rate": 2.0520054256419236e-05,
+      "loss": 0.909,
+      "num_input_tokens_seen": 106004480,
+      "step": 12940
+    },
+    {
+      "epoch": 1.6733428091484688,
+      "grad_norm": 0.6692883372306824,
+      "learning_rate": 2.0486777368226143e-05,
+      "loss": 0.7516,
+      "num_input_tokens_seen": 106086400,
+      "step": 12950
+    },
+    {
+      "epoch": 1.6746349657578499,
+      "grad_norm": 1.0037450790405273,
+      "learning_rate": 2.045350874372717e-05,
+      "loss": 0.977,
+      "num_input_tokens_seen": 106168320,
+      "step": 12960
+    },
+    {
+      "epoch": 1.675927122367231,
+      "grad_norm": 0.8744035363197327,
+      "learning_rate": 2.0420248443837048e-05,
+      "loss": 0.8461,
+      "num_input_tokens_seen": 106250240,
+      "step": 12970
+    },
+    {
+      "epoch": 1.677219278976612,
+      "grad_norm": 0.7862047553062439,
+      "learning_rate": 2.0386996529455276e-05,
+      "loss": 0.7468,
+      "num_input_tokens_seen": 106332160,
+      "step": 12980
+    },
+    {
+      "epoch": 1.678511435585993,
+      "grad_norm": 0.8968716263771057,
+      "learning_rate": 2.0353753061465972e-05,
+      "loss": 0.7406,
+      "num_input_tokens_seen": 106414080,
+      "step": 12990
+    },
+    {
+      "epoch": 1.6798035921953742,
+      "grad_norm": 0.9072631001472473,
+      "learning_rate": 2.0320518100737817e-05,
+      "loss": 0.9977,
+      "num_input_tokens_seen": 106496000,
+      "step": 13000
+    },
+    {
+      "epoch": 1.681095748804755,
+      "grad_norm": 0.5240357518196106,
+      "learning_rate": 2.0287291708123888e-05,
+      "loss": 0.867,
+      "num_input_tokens_seen": 106577920,
+      "step": 13010
+    },
+    {
+      "epoch": 1.6823879054141362,
+      "grad_norm": 0.6482962369918823,
+      "learning_rate": 2.0254073944461603e-05,
+      "loss": 0.86,
+      "num_input_tokens_seen": 106659840,
+      "step": 13020
+    },
+    {
+      "epoch": 1.6836800620235173,
+      "grad_norm": 0.7693918347358704,
+      "learning_rate": 2.0220864870572555e-05,
+      "loss": 0.9676,
+      "num_input_tokens_seen": 106741760,
+      "step": 13030
+    },
+    {
+      "epoch": 1.6849722186328981,
+      "grad_norm": 0.6665438413619995,
+      "learning_rate": 2.0187664547262446e-05,
+      "loss": 0.7239,
+      "num_input_tokens_seen": 106823680,
+      "step": 13040
+    },
+    {
+      "epoch": 1.6862643752422795,
+      "grad_norm": 0.2982095777988434,
+      "learning_rate": 2.0154473035320936e-05,
+      "loss": 0.9287,
+      "num_input_tokens_seen": 106905600,
+      "step": 13050
+    },
+    {
+      "epoch": 1.6875565318516603,
+      "grad_norm": 14.455750465393066,
+      "learning_rate": 2.0121290395521566e-05,
+      "loss": 0.7107,
+      "num_input_tokens_seen": 106987520,
+      "step": 13060
+    },
+    {
+      "epoch": 1.6888486884610416,
+      "grad_norm": 0.7520779371261597,
+      "learning_rate": 2.008811668862164e-05,
+      "loss": 0.7824,
+      "num_input_tokens_seen": 107069440,
+      "step": 13070
+    },
+    {
+      "epoch": 1.6901408450704225,
+      "grad_norm": 0.6052248477935791,
+      "learning_rate": 2.0054951975362067e-05,
+      "loss": 0.6419,
+      "num_input_tokens_seen": 107151360,
+      "step": 13080
+    },
+    {
+      "epoch": 1.6914330016798036,
+      "grad_norm": 0.2903338372707367,
+      "learning_rate": 2.0021796316467346e-05,
+      "loss": 0.5254,
+      "num_input_tokens_seen": 107233280,
+      "step": 13090
+    },
+    {
+      "epoch": 1.6927251582891847,
+      "grad_norm": 0.5044610500335693,
+      "learning_rate": 1.9988649772645346e-05,
+      "loss": 0.6578,
+      "num_input_tokens_seen": 107315200,
+      "step": 13100
+    },
+    {
+      "epoch": 1.6940173148985656,
+      "grad_norm": 0.6834781169891357,
+      "learning_rate": 1.995551240458728e-05,
+      "loss": 0.7578,
+      "num_input_tokens_seen": 107397120,
+      "step": 13110
+    },
+    {
+      "epoch": 1.6953094715079469,
+      "grad_norm": 0.47251078486442566,
+      "learning_rate": 1.9922384272967535e-05,
+      "loss": 0.6271,
+      "num_input_tokens_seen": 107479040,
+      "step": 13120
+    },
+    {
+      "epoch": 1.6966016281173277,
+      "grad_norm": 0.27154985070228577,
+      "learning_rate": 1.9889265438443607e-05,
+      "loss": 0.6214,
+      "num_input_tokens_seen": 107560960,
+      "step": 13130
+    },
+    {
+      "epoch": 1.697893784726709,
+      "grad_norm": 0.7411037683486938,
+      "learning_rate": 1.985615596165597e-05,
+      "loss": 1.0424,
+      "num_input_tokens_seen": 107642880,
+      "step": 13140
+    },
+    {
+      "epoch": 1.69918594133609,
+      "grad_norm": 0.7935945391654968,
+      "learning_rate": 1.982305590322793e-05,
+      "loss": 1.0063,
+      "num_input_tokens_seen": 107724800,
+      "step": 13150
+    },
+    {
+      "epoch": 1.700478097945471,
+      "grad_norm": 0.5809916257858276,
+      "learning_rate": 1.97899653237656e-05,
+      "loss": 0.796,
+      "num_input_tokens_seen": 107806720,
+      "step": 13160
+    },
+    {
+      "epoch": 1.701770254554852,
+      "grad_norm": 0.7981992363929749,
+      "learning_rate": 1.9756884283857685e-05,
+      "loss": 0.8084,
+      "num_input_tokens_seen": 107888640,
+      "step": 13170
+    },
+    {
+      "epoch": 1.703062411164233,
+      "grad_norm": 0.6108459234237671,
+      "learning_rate": 1.9723812844075473e-05,
+      "loss": 1.0913,
+      "num_input_tokens_seen": 107970560,
+      "step": 13180
+    },
+    {
+      "epoch": 1.7043545677736143,
+      "grad_norm": 0.6195822954177856,
+      "learning_rate": 1.9690751064972625e-05,
+      "loss": 1.1137,
+      "num_input_tokens_seen": 108052480,
+      "step": 13190
+    },
+    {
+      "epoch": 1.7056467243829951,
+      "grad_norm": 0.6808088421821594,
+      "learning_rate": 1.965769900708515e-05,
+      "loss": 0.9466,
+      "num_input_tokens_seen": 108134400,
+      "step": 13200
+    },
+    {
+      "epoch": 1.7069388809923762,
+      "grad_norm": 0.6025552749633789,
+      "learning_rate": 1.9624656730931258e-05,
+      "loss": 1.0154,
+      "num_input_tokens_seen": 108216320,
+      "step": 13210
+    },
+    {
+      "epoch": 1.7082310376017573,
+      "grad_norm": 0.6096053123474121,
+      "learning_rate": 1.959162429701121e-05,
+      "loss": 0.9086,
+      "num_input_tokens_seen": 108298240,
+      "step": 13220
+    },
+    {
+      "epoch": 1.7095231942111384,
+      "grad_norm": 0.7634369730949402,
+      "learning_rate": 1.955860176580729e-05,
+      "loss": 0.9669,
+      "num_input_tokens_seen": 108380160,
+      "step": 13230
+    },
+    {
+      "epoch": 1.7108153508205195,
+      "grad_norm": 5.748688220977783,
+      "learning_rate": 1.9525589197783618e-05,
+      "loss": 0.8869,
+      "num_input_tokens_seen": 108462080,
+      "step": 13240
+    },
+    {
+      "epoch": 1.7121075074299004,
+      "grad_norm": 0.9097843170166016,
+      "learning_rate": 1.9492586653386103e-05,
+      "loss": 1.3718,
+      "num_input_tokens_seen": 108544000,
+      "step": 13250
+    },
+    {
+      "epoch": 1.7133996640392817,
+      "grad_norm": 0.6823452711105347,
+      "learning_rate": 1.945959419304226e-05,
+      "loss": 1.1025,
+      "num_input_tokens_seen": 108625920,
+      "step": 13260
+    },
+    {
+      "epoch": 1.7146918206486625,
+      "grad_norm": 0.8677635192871094,
+      "learning_rate": 1.942661187716118e-05,
+      "loss": 0.7662,
+      "num_input_tokens_seen": 108707840,
+      "step": 13270
+    },
+    {
+      "epoch": 1.7159839772580436,
+      "grad_norm": 0.8804795742034912,
+      "learning_rate": 1.9393639766133363e-05,
+      "loss": 0.9356,
+      "num_input_tokens_seen": 108789760,
+      "step": 13280
+    },
+    {
+      "epoch": 1.7172761338674247,
+      "grad_norm": 0.6428921222686768,
+      "learning_rate": 1.936067792033061e-05,
+      "loss": 0.608,
+      "num_input_tokens_seen": 108871680,
+      "step": 13290
+    },
+    {
+      "epoch": 1.7185682904768058,
+      "grad_norm": 0.5811804533004761,
+      "learning_rate": 1.9327726400105963e-05,
+      "loss": 1.0139,
+      "num_input_tokens_seen": 108953600,
+      "step": 13300
+    },
+    {
+      "epoch": 1.719860447086187,
+      "grad_norm": 1.0158241987228394,
+      "learning_rate": 1.9294785265793514e-05,
+      "loss": 0.7744,
+      "num_input_tokens_seen": 109035520,
+      "step": 13310
+    },
+    {
+      "epoch": 1.7211526036955678,
+      "grad_norm": 1.1166173219680786,
+      "learning_rate": 1.9261854577708366e-05,
+      "loss": 0.7847,
+      "num_input_tokens_seen": 109117440,
+      "step": 13320
+    },
+    {
+      "epoch": 1.722444760304949,
+      "grad_norm": 0.5615746974945068,
+      "learning_rate": 1.9228934396146486e-05,
+      "loss": 0.9334,
+      "num_input_tokens_seen": 109199360,
+      "step": 13330
+    },
+    {
+      "epoch": 1.72373691691433,
+      "grad_norm": 0.474103182554245,
+      "learning_rate": 1.9196024781384607e-05,
+      "loss": 0.8011,
+      "num_input_tokens_seen": 109281280,
+      "step": 13340
+    },
+    {
+      "epoch": 1.725029073523711,
+      "grad_norm": 0.5429681539535522,
+      "learning_rate": 1.9163125793680125e-05,
+      "loss": 0.8737,
+      "num_input_tokens_seen": 109363200,
+      "step": 13350
+    },
+    {
+      "epoch": 1.7263212301330921,
+      "grad_norm": 0.22766174376010895,
+      "learning_rate": 1.9130237493270948e-05,
+      "loss": 0.5359,
+      "num_input_tokens_seen": 109445120,
+      "step": 13360
+    },
+    {
+      "epoch": 1.7276133867424732,
+      "grad_norm": 0.9762837290763855,
+      "learning_rate": 1.9097359940375452e-05,
+      "loss": 0.6703,
+      "num_input_tokens_seen": 109527040,
+      "step": 13370
+    },
+    {
+      "epoch": 1.7289055433518543,
+      "grad_norm": 0.3081243932247162,
+      "learning_rate": 1.9064493195192293e-05,
+      "loss": 0.8231,
+      "num_input_tokens_seen": 109608960,
+      "step": 13380
+    },
+    {
+      "epoch": 1.7301976999612352,
+      "grad_norm": 0.8469387292861938,
+      "learning_rate": 1.9031637317900386e-05,
+      "loss": 0.7302,
+      "num_input_tokens_seen": 109690880,
+      "step": 13390
+    },
+    {
+      "epoch": 1.7314898565706165,
+      "grad_norm": 0.8543357253074646,
+      "learning_rate": 1.8998792368658703e-05,
+      "loss": 0.7282,
+      "num_input_tokens_seen": 109772800,
+      "step": 13400
+    },
+    {
+      "epoch": 1.7327820131799974,
+      "grad_norm": 0.5329307317733765,
+      "learning_rate": 1.8965958407606236e-05,
+      "loss": 0.8997,
+      "num_input_tokens_seen": 109854720,
+      "step": 13410
+    },
+    {
+      "epoch": 1.7340741697893785,
+      "grad_norm": 0.7571797966957092,
+      "learning_rate": 1.893313549486184e-05,
+      "loss": 0.8703,
+      "num_input_tokens_seen": 109936640,
+      "step": 13420
+    },
+    {
+      "epoch": 1.7353663263987595,
+      "grad_norm": 0.6506255865097046,
+      "learning_rate": 1.890032369052415e-05,
+      "loss": 0.5685,
+      "num_input_tokens_seen": 110018560,
+      "step": 13430
+    },
+    {
+      "epoch": 1.7366584830081406,
+      "grad_norm": 0.2349756509065628,
+      "learning_rate": 1.8867523054671475e-05,
+      "loss": 0.9621,
+      "num_input_tokens_seen": 110100480,
+      "step": 13440
+    },
+    {
+      "epoch": 1.7379506396175217,
+      "grad_norm": 0.5002673864364624,
+      "learning_rate": 1.8834733647361635e-05,
+      "loss": 0.8279,
+      "num_input_tokens_seen": 110182400,
+      "step": 13450
+    },
+    {
+      "epoch": 1.7392427962269026,
+      "grad_norm": 0.4068397581577301,
+      "learning_rate": 1.880195552863194e-05,
+      "loss": 0.6358,
+      "num_input_tokens_seen": 110264320,
+      "step": 13460
+    },
+    {
+      "epoch": 1.740534952836284,
+      "grad_norm": 0.6473715901374817,
+      "learning_rate": 1.8769188758498973e-05,
+      "loss": 0.8022,
+      "num_input_tokens_seen": 110346240,
+      "step": 13470
+    },
+    {
+      "epoch": 1.7418271094456648,
+      "grad_norm": 0.29056867957115173,
+      "learning_rate": 1.8736433396958605e-05,
+      "loss": 0.6169,
+      "num_input_tokens_seen": 110428160,
+      "step": 13480
+    },
+    {
+      "epoch": 1.7431192660550459,
+      "grad_norm": 0.49378782510757446,
+      "learning_rate": 1.8703689503985754e-05,
+      "loss": 0.6387,
+      "num_input_tokens_seen": 110510080,
+      "step": 13490
+    },
+    {
+      "epoch": 1.744411422664427,
+      "grad_norm": 0.570563554763794,
+      "learning_rate": 1.867095713953439e-05,
+      "loss": 1.0567,
+      "num_input_tokens_seen": 110592000,
+      "step": 13500
+    },
+    {
+      "epoch": 1.745703579273808,
+      "grad_norm": 0.7375537753105164,
+      "learning_rate": 1.8638236363537348e-05,
+      "loss": 0.5526,
+      "num_input_tokens_seen": 110673920,
+      "step": 13510
+    },
+    {
+      "epoch": 1.7469957358831891,
+      "grad_norm": 0.6009176969528198,
+      "learning_rate": 1.8605527235906235e-05,
+      "loss": 0.987,
+      "num_input_tokens_seen": 110755840,
+      "step": 13520
+    },
+    {
+      "epoch": 1.74828789249257,
+      "grad_norm": 0.41143083572387695,
+      "learning_rate": 1.8572829816531364e-05,
+      "loss": 0.604,
+      "num_input_tokens_seen": 110837760,
+      "step": 13530
+    },
+    {
+      "epoch": 1.7495800491019513,
+      "grad_norm": 1.1467337608337402,
+      "learning_rate": 1.854014416528157e-05,
+      "loss": 0.6048,
+      "num_input_tokens_seen": 110919680,
+      "step": 13540
+    },
+    {
+      "epoch": 1.7508722057113322,
+      "grad_norm": 0.6407749652862549,
+      "learning_rate": 1.8507470342004182e-05,
+      "loss": 0.8796,
+      "num_input_tokens_seen": 111001600,
+      "step": 13550
+    },
+    {
+      "epoch": 1.7521643623207133,
+      "grad_norm": 0.8775169253349304,
+      "learning_rate": 1.847480840652483e-05,
+      "loss": 0.8039,
+      "num_input_tokens_seen": 111083520,
+      "step": 13560
+    },
+    {
+      "epoch": 1.7534565189300944,
+      "grad_norm": 0.9216033220291138,
+      "learning_rate": 1.844215841864741e-05,
+      "loss": 0.6875,
+      "num_input_tokens_seen": 111165440,
+      "step": 13570
+    },
+    {
+      "epoch": 1.7547486755394752,
+      "grad_norm": 0.7694590091705322,
+      "learning_rate": 1.8409520438153933e-05,
+      "loss": 0.9024,
+      "num_input_tokens_seen": 111247360,
+      "step": 13580
+    },
+    {
+      "epoch": 1.7560408321488565,
+      "grad_norm": 0.25773632526397705,
+      "learning_rate": 1.8376894524804416e-05,
+      "loss": 0.6588,
+      "num_input_tokens_seen": 111329280,
+      "step": 13590
+    },
+    {
+      "epoch": 1.7573329887582374,
+      "grad_norm": 0.8692495226860046,
+      "learning_rate": 1.8344280738336796e-05,
+      "loss": 0.9931,
+      "num_input_tokens_seen": 111411200,
+      "step": 13600
+    },
+    {
+      "epoch": 1.7586251453676187,
+      "grad_norm": 0.17592206597328186,
+      "learning_rate": 1.8311679138466772e-05,
+      "loss": 0.9949,
+      "num_input_tokens_seen": 111493120,
+      "step": 13610
+    },
+    {
+      "epoch": 1.7599173019769996,
+      "grad_norm": 0.8434010148048401,
+      "learning_rate": 1.827908978488779e-05,
+      "loss": 1.0308,
+      "num_input_tokens_seen": 111575040,
+      "step": 13620
+    },
+    {
+      "epoch": 1.7612094585863807,
+      "grad_norm": 1.8556253910064697,
+      "learning_rate": 1.8246512737270798e-05,
+      "loss": 0.996,
+      "num_input_tokens_seen": 111656960,
+      "step": 13630
+    },
+    {
+      "epoch": 1.7625016151957618,
+      "grad_norm": 4.145997047424316,
+      "learning_rate": 1.8213948055264278e-05,
+      "loss": 1.0672,
+      "num_input_tokens_seen": 111738880,
+      "step": 13640
+    },
+    {
+      "epoch": 1.7637937718051426,
+      "grad_norm": 0.3059723377227783,
+      "learning_rate": 1.8181395798494048e-05,
+      "loss": 0.5107,
+      "num_input_tokens_seen": 111820800,
+      "step": 13650
+    },
+    {
+      "epoch": 1.765085928414524,
+      "grad_norm": 0.9706294536590576,
+      "learning_rate": 1.8148856026563148e-05,
+      "loss": 1.0283,
+      "num_input_tokens_seen": 111902720,
+      "step": 13660
+    },
+    {
+      "epoch": 1.7663780850239048,
+      "grad_norm": 0.6483558416366577,
+      "learning_rate": 1.81163287990518e-05,
+      "loss": 0.9913,
+      "num_input_tokens_seen": 111984640,
+      "step": 13670
+    },
+    {
+      "epoch": 1.767670241633286,
+      "grad_norm": 0.5920456647872925,
+      "learning_rate": 1.8083814175517234e-05,
+      "loss": 0.5184,
+      "num_input_tokens_seen": 112066560,
+      "step": 13680
+    },
+    {
+      "epoch": 1.768962398242667,
+      "grad_norm": 0.6596719622612,
+      "learning_rate": 1.80513122154936e-05,
+      "loss": 0.962,
+      "num_input_tokens_seen": 112148480,
+      "step": 13690
+    },
+    {
+      "epoch": 1.770254554852048,
+      "grad_norm": 0.7937291264533997,
+      "learning_rate": 1.8018822978491872e-05,
+      "loss": 0.6034,
+      "num_input_tokens_seen": 112230400,
+      "step": 13700
+    },
+    {
+      "epoch": 1.7715467114614292,
+      "grad_norm": 0.5990145206451416,
+      "learning_rate": 1.798634652399972e-05,
+      "loss": 0.8974,
+      "num_input_tokens_seen": 112312320,
+      "step": 13710
+    },
+    {
+      "epoch": 1.77283886807081,
+      "grad_norm": 0.6401026844978333,
+      "learning_rate": 1.795388291148143e-05,
+      "loss": 1.0047,
+      "num_input_tokens_seen": 112394240,
+      "step": 13720
+    },
+    {
+      "epoch": 1.7741310246801913,
+      "grad_norm": 0.6857889294624329,
+      "learning_rate": 1.7921432200377734e-05,
+      "loss": 0.608,
+      "num_input_tokens_seen": 112476160,
+      "step": 13730
+    },
+    {
+      "epoch": 1.7754231812895722,
+      "grad_norm": 0.6856977939605713,
+      "learning_rate": 1.7888994450105788e-05,
+      "loss": 1.2077,
+      "num_input_tokens_seen": 112558080,
+      "step": 13740
+    },
+    {
+      "epoch": 1.7767153378989533,
+      "grad_norm": 0.4340798258781433,
+      "learning_rate": 1.785656972005897e-05,
+      "loss": 1.0667,
+      "num_input_tokens_seen": 112640000,
+      "step": 13750
+    },
+    {
+      "epoch": 1.7780074945083344,
+      "grad_norm": 0.6757582426071167,
+      "learning_rate": 1.7824158069606867e-05,
+      "loss": 0.8964,
+      "num_input_tokens_seen": 112721920,
+      "step": 13760
+    },
+    {
+      "epoch": 1.7792996511177155,
+      "grad_norm": 0.8289663791656494,
+      "learning_rate": 1.7791759558095077e-05,
+      "loss": 0.6691,
+      "num_input_tokens_seen": 112803840,
+      "step": 13770
+    },
+    {
+      "epoch": 1.7805918077270966,
+      "grad_norm": 0.620309054851532,
+      "learning_rate": 1.775937424484515e-05,
+      "loss": 0.9059,
+      "num_input_tokens_seen": 112885760,
+      "step": 13780
+    },
+    {
+      "epoch": 1.7818839643364774,
+      "grad_norm": 0.7046340703964233,
+      "learning_rate": 1.7727002189154502e-05,
+      "loss": 0.724,
+      "num_input_tokens_seen": 112967680,
+      "step": 13790
+    },
+    {
+      "epoch": 1.7831761209458588,
+      "grad_norm": 0.7934147715568542,
+      "learning_rate": 1.7694643450296216e-05,
+      "loss": 1.064,
+      "num_input_tokens_seen": 113049600,
+      "step": 13800
+    },
+    {
+      "epoch": 1.7844682775552396,
+      "grad_norm": 0.5995201468467712,
+      "learning_rate": 1.7662298087519052e-05,
+      "loss": 0.9433,
+      "num_input_tokens_seen": 113131520,
+      "step": 13810
+    },
+    {
+      "epoch": 1.7857604341646207,
+      "grad_norm": 0.8406663537025452,
+      "learning_rate": 1.762996616004723e-05,
+      "loss": 0.8747,
+      "num_input_tokens_seen": 113213440,
+      "step": 13820
+    },
+    {
+      "epoch": 1.7870525907740018,
+      "grad_norm": 0.5735936760902405,
+      "learning_rate": 1.7597647727080408e-05,
+      "loss": 0.9011,
+      "num_input_tokens_seen": 113295360,
+      "step": 13830
+    },
+    {
+      "epoch": 1.788344747383383,
+      "grad_norm": 0.7899253368377686,
+      "learning_rate": 1.7565342847793502e-05,
+      "loss": 0.9851,
+      "num_input_tokens_seen": 113377280,
+      "step": 13840
+    },
+    {
+      "epoch": 1.789636903992764,
+      "grad_norm": 0.9319769740104675,
+      "learning_rate": 1.7533051581336644e-05,
+      "loss": 0.9077,
+      "num_input_tokens_seen": 113459200,
+      "step": 13850
+    },
+    {
+      "epoch": 1.7909290606021449,
+      "grad_norm": 0.6628711819648743,
+      "learning_rate": 1.7500773986835013e-05,
+      "loss": 0.9925,
+      "num_input_tokens_seen": 113541120,
+      "step": 13860
+    },
+    {
+      "epoch": 1.7922212172115262,
+      "grad_norm": 0.373602032661438,
+      "learning_rate": 1.7468510123388775e-05,
+      "loss": 0.7818,
+      "num_input_tokens_seen": 113623040,
+      "step": 13870
+    },
+    {
+      "epoch": 1.793513373820907,
+      "grad_norm": 0.574948787689209,
+      "learning_rate": 1.743626005007294e-05,
+      "loss": 0.9479,
+      "num_input_tokens_seen": 113704960,
+      "step": 13880
+    },
+    {
+      "epoch": 1.7948055304302881,
+      "grad_norm": 0.7120422124862671,
+      "learning_rate": 1.740402382593727e-05,
+      "loss": 0.9624,
+      "num_input_tokens_seen": 113786880,
+      "step": 13890
+    },
+    {
+      "epoch": 1.7960976870396692,
+      "grad_norm": 0.5288504958152771,
+      "learning_rate": 1.7371801510006193e-05,
+      "loss": 0.9844,
+      "num_input_tokens_seen": 113868800,
+      "step": 13900
+    },
+    {
+      "epoch": 1.7973898436490503,
+      "grad_norm": 0.5549502372741699,
+      "learning_rate": 1.733959316127862e-05,
+      "loss": 0.8112,
+      "num_input_tokens_seen": 113950720,
+      "step": 13910
+    },
+    {
+      "epoch": 1.7986820002584314,
+      "grad_norm": 0.6821995377540588,
+      "learning_rate": 1.730739883872795e-05,
+      "loss": 0.9197,
+      "num_input_tokens_seen": 114032640,
+      "step": 13920
+    },
+    {
+      "epoch": 1.7999741568678123,
+      "grad_norm": 0.44002169370651245,
+      "learning_rate": 1.7275218601301848e-05,
+      "loss": 0.6208,
+      "num_input_tokens_seen": 114114560,
+      "step": 13930
+    },
+    {
+      "epoch": 1.8012663134771936,
+      "grad_norm": 0.7106473445892334,
+      "learning_rate": 1.7243052507922226e-05,
+      "loss": 0.9374,
+      "num_input_tokens_seen": 114196480,
+      "step": 13940
+    },
+    {
+      "epoch": 1.8025584700865744,
+      "grad_norm": 0.6388192772865295,
+      "learning_rate": 1.7210900617485075e-05,
+      "loss": 1.0343,
+      "num_input_tokens_seen": 114278400,
+      "step": 13950
+    },
+    {
+      "epoch": 1.8038506266959555,
+      "grad_norm": 0.644347071647644,
+      "learning_rate": 1.7178762988860393e-05,
+      "loss": 0.8684,
+      "num_input_tokens_seen": 114360320,
+      "step": 13960
+    },
+    {
+      "epoch": 1.8051427833053366,
+      "grad_norm": 0.5687951445579529,
+      "learning_rate": 1.7146639680892062e-05,
+      "loss": 1.0918,
+      "num_input_tokens_seen": 114442240,
+      "step": 13970
+    },
+    {
+      "epoch": 1.8064349399147177,
+      "grad_norm": 0.775614321231842,
+      "learning_rate": 1.711453075239773e-05,
+      "loss": 0.7429,
+      "num_input_tokens_seen": 114524160,
+      "step": 13980
+    },
+    {
+      "epoch": 1.8077270965240988,
+      "grad_norm": 0.25670918822288513,
+      "learning_rate": 1.7082436262168745e-05,
+      "loss": 0.7727,
+      "num_input_tokens_seen": 114606080,
+      "step": 13990
+    },
+    {
+      "epoch": 1.8090192531334797,
+      "grad_norm": 0.9331316351890564,
+      "learning_rate": 1.705035626896998e-05,
+      "loss": 1.0647,
+      "num_input_tokens_seen": 114688000,
+      "step": 14000
+    },
+    {
+      "epoch": 1.810311409742861,
+      "grad_norm": 1.6310906410217285,
+      "learning_rate": 1.7018290831539795e-05,
+      "loss": 0.7367,
+      "num_input_tokens_seen": 114769920,
+      "step": 14010
+    },
+    {
+      "epoch": 1.8116035663522418,
+      "grad_norm": 0.9030694961547852,
+      "learning_rate": 1.6986240008589903e-05,
+      "loss": 0.9311,
+      "num_input_tokens_seen": 114851840,
+      "step": 14020
+    },
+    {
+      "epoch": 1.812895722961623,
+      "grad_norm": 0.6895557045936584,
+      "learning_rate": 1.695420385880522e-05,
+      "loss": 0.9413,
+      "num_input_tokens_seen": 114933760,
+      "step": 14030
+    },
+    {
+      "epoch": 1.814187879571004,
+      "grad_norm": 0.7710975408554077,
+      "learning_rate": 1.6922182440843843e-05,
+      "loss": 0.7935,
+      "num_input_tokens_seen": 115015680,
+      "step": 14040
+    },
+    {
+      "epoch": 1.8154800361803851,
+      "grad_norm": 0.979817807674408,
+      "learning_rate": 1.689017581333685e-05,
+      "loss": 0.9507,
+      "num_input_tokens_seen": 115097600,
+      "step": 14050
+    },
+    {
+      "epoch": 1.8167721927897662,
+      "grad_norm": 0.2595027983188629,
+      "learning_rate": 1.685818403488827e-05,
+      "loss": 1.1626,
+      "num_input_tokens_seen": 115179520,
+      "step": 14060
+    },
+    {
+      "epoch": 1.818064349399147,
+      "grad_norm": 0.20249100029468536,
+      "learning_rate": 1.6826207164074924e-05,
+      "loss": 0.4731,
+      "num_input_tokens_seen": 115261440,
+      "step": 14070
+    },
+    {
+      "epoch": 1.8193565060085284,
+      "grad_norm": 0.5475636124610901,
+      "learning_rate": 1.6794245259446347e-05,
+      "loss": 0.8077,
+      "num_input_tokens_seen": 115343360,
+      "step": 14080
+    },
+    {
+      "epoch": 1.8206486626179093,
+      "grad_norm": 0.6575671434402466,
+      "learning_rate": 1.6762298379524684e-05,
+      "loss": 0.8489,
+      "num_input_tokens_seen": 115425280,
+      "step": 14090
+    },
+    {
+      "epoch": 1.8219408192272903,
+      "grad_norm": 0.5988595485687256,
+      "learning_rate": 1.6730366582804535e-05,
+      "loss": 0.5274,
+      "num_input_tokens_seen": 115507200,
+      "step": 14100
+    },
+    {
+      "epoch": 1.8232329758366714,
+      "grad_norm": 0.2947266399860382,
+      "learning_rate": 1.6698449927752924e-05,
+      "loss": 0.659,
+      "num_input_tokens_seen": 115589120,
+      "step": 14110
+    },
+    {
+      "epoch": 1.8245251324460523,
+      "grad_norm": 0.7184574604034424,
+      "learning_rate": 1.6666548472809104e-05,
+      "loss": 1.1145,
+      "num_input_tokens_seen": 115671040,
+      "step": 14120
+    },
+    {
+      "epoch": 1.8258172890554336,
+      "grad_norm": 0.8786848783493042,
+      "learning_rate": 1.6634662276384548e-05,
+      "loss": 0.8391,
+      "num_input_tokens_seen": 115752960,
+      "step": 14130
+    },
+    {
+      "epoch": 1.8271094456648145,
+      "grad_norm": 0.6880916357040405,
+      "learning_rate": 1.660279139686275e-05,
+      "loss": 0.802,
+      "num_input_tokens_seen": 115834880,
+      "step": 14140
+    },
+    {
+      "epoch": 1.8284016022741958,
+      "grad_norm": 0.8588114380836487,
+      "learning_rate": 1.657093589259917e-05,
+      "loss": 0.7398,
+      "num_input_tokens_seen": 115916800,
+      "step": 14150
+    },
+    {
+      "epoch": 1.8296937588835767,
+      "grad_norm": 0.6294015049934387,
+      "learning_rate": 1.6539095821921136e-05,
+      "loss": 0.8751,
+      "num_input_tokens_seen": 115998720,
+      "step": 14160
+    },
+    {
+      "epoch": 1.8309859154929577,
+      "grad_norm": 0.8277295231819153,
+      "learning_rate": 1.650727124312768e-05,
+      "loss": 0.7839,
+      "num_input_tokens_seen": 116080640,
+      "step": 14170
+    },
+    {
+      "epoch": 1.8322780721023388,
+      "grad_norm": 0.5361806154251099,
+      "learning_rate": 1.6475462214489513e-05,
+      "loss": 0.8393,
+      "num_input_tokens_seen": 116162560,
+      "step": 14180
+    },
+    {
+      "epoch": 1.8335702287117197,
+      "grad_norm": 0.5568910241127014,
+      "learning_rate": 1.6443668794248828e-05,
+      "loss": 0.6922,
+      "num_input_tokens_seen": 116244480,
+      "step": 14190
+    },
+    {
+      "epoch": 1.834862385321101,
+      "grad_norm": 0.615774393081665,
+      "learning_rate": 1.641189104061928e-05,
+      "loss": 0.9365,
+      "num_input_tokens_seen": 116326400,
+      "step": 14200
+    },
+    {
+      "epoch": 1.8361545419304819,
+      "grad_norm": 0.5470173358917236,
+      "learning_rate": 1.63801290117858e-05,
+      "loss": 0.8715,
+      "num_input_tokens_seen": 116408320,
+      "step": 14210
+    },
+    {
+      "epoch": 1.837446698539863,
+      "grad_norm": 0.7241597771644592,
+      "learning_rate": 1.6348382765904567e-05,
+      "loss": 0.8787,
+      "num_input_tokens_seen": 116490240,
+      "step": 14220
+    },
+    {
+      "epoch": 1.838738855149244,
+      "grad_norm": 0.7052979469299316,
+      "learning_rate": 1.631665236110283e-05,
+      "loss": 0.8164,
+      "num_input_tokens_seen": 116572160,
+      "step": 14230
+    },
+    {
+      "epoch": 1.8400310117586252,
+      "grad_norm": 0.6551203727722168,
+      "learning_rate": 1.6284937855478837e-05,
+      "loss": 1.0631,
+      "num_input_tokens_seen": 116654080,
+      "step": 14240
+    },
+    {
+      "epoch": 1.8413231683680062,
+      "grad_norm": 0.6758942604064941,
+      "learning_rate": 1.6253239307101748e-05,
+      "loss": 1.008,
+      "num_input_tokens_seen": 116736000,
+      "step": 14250
+    },
+    {
+      "epoch": 1.8426153249773871,
+      "grad_norm": 0.588173508644104,
+      "learning_rate": 1.6221556774011474e-05,
+      "loss": 1.006,
+      "num_input_tokens_seen": 116817920,
+      "step": 14260
+    },
+    {
+      "epoch": 1.8439074815867684,
+      "grad_norm": 0.2741510272026062,
+      "learning_rate": 1.6189890314218634e-05,
+      "loss": 0.8024,
+      "num_input_tokens_seen": 116899840,
+      "step": 14270
+    },
+    {
+      "epoch": 1.8451996381961493,
+      "grad_norm": 0.38996413350105286,
+      "learning_rate": 1.6158239985704378e-05,
+      "loss": 0.9059,
+      "num_input_tokens_seen": 116981760,
+      "step": 14280
+    },
+    {
+      "epoch": 1.8464917948055304,
+      "grad_norm": 0.6522138118743896,
+      "learning_rate": 1.6126605846420366e-05,
+      "loss": 0.7783,
+      "num_input_tokens_seen": 117063680,
+      "step": 14290
+    },
+    {
+      "epoch": 1.8477839514149115,
+      "grad_norm": 0.5174360275268555,
+      "learning_rate": 1.609498795428857e-05,
+      "loss": 0.8708,
+      "num_input_tokens_seen": 117145600,
+      "step": 14300
+    },
+    {
+      "epoch": 1.8490761080242926,
+      "grad_norm": 0.47565585374832153,
+      "learning_rate": 1.606338636720125e-05,
+      "loss": 0.4722,
+      "num_input_tokens_seen": 117227520,
+      "step": 14310
+    },
+    {
+      "epoch": 1.8503682646336737,
+      "grad_norm": 0.5831138491630554,
+      "learning_rate": 1.6031801143020785e-05,
+      "loss": 0.9373,
+      "num_input_tokens_seen": 117309440,
+      "step": 14320
+    },
+    {
+      "epoch": 1.8516604212430545,
+      "grad_norm": 0.5426406860351562,
+      "learning_rate": 1.6000232339579616e-05,
+      "loss": 0.9453,
+      "num_input_tokens_seen": 117391360,
+      "step": 14330
+    },
+    {
+      "epoch": 1.8529525778524358,
+      "grad_norm": 0.899085283279419,
+      "learning_rate": 1.5968680014680105e-05,
+      "loss": 0.8859,
+      "num_input_tokens_seen": 117473280,
+      "step": 14340
+    },
+    {
+      "epoch": 1.8542447344618167,
+      "grad_norm": 0.5543519258499146,
+      "learning_rate": 1.5937144226094426e-05,
+      "loss": 1.0594,
+      "num_input_tokens_seen": 117555200,
+      "step": 14350
+    },
+    {
+      "epoch": 1.8555368910711978,
+      "grad_norm": 0.8025951385498047,
+      "learning_rate": 1.590562503156452e-05,
+      "loss": 0.8882,
+      "num_input_tokens_seen": 117637120,
+      "step": 14360
+    },
+    {
+      "epoch": 1.8568290476805789,
+      "grad_norm": 0.6252597570419312,
+      "learning_rate": 1.5874122488801888e-05,
+      "loss": 0.6613,
+      "num_input_tokens_seen": 117719040,
+      "step": 14370
+    },
+    {
+      "epoch": 1.85812120428996,
+      "grad_norm": 0.5692026615142822,
+      "learning_rate": 1.5842636655487585e-05,
+      "loss": 0.6339,
+      "num_input_tokens_seen": 117800960,
+      "step": 14380
+    },
+    {
+      "epoch": 1.859413360899341,
+      "grad_norm": 0.651740550994873,
+      "learning_rate": 1.5811167589272068e-05,
+      "loss": 1.0055,
+      "num_input_tokens_seen": 117882880,
+      "step": 14390
+    },
+    {
+      "epoch": 1.860705517508722,
+      "grad_norm": 0.6472187638282776,
+      "learning_rate": 1.577971534777507e-05,
+      "loss": 0.7919,
+      "num_input_tokens_seen": 117964800,
+      "step": 14400
+    },
+    {
+      "epoch": 1.8619976741181032,
+      "grad_norm": 0.5527770519256592,
+      "learning_rate": 1.5748279988585528e-05,
+      "loss": 0.7019,
+      "num_input_tokens_seen": 118046720,
+      "step": 14410
+    },
+    {
+      "epoch": 1.863289830727484,
+      "grad_norm": 0.3149762451648712,
+      "learning_rate": 1.571686156926147e-05,
+      "loss": 0.7455,
+      "num_input_tokens_seen": 118128640,
+      "step": 14420
+    },
+    {
+      "epoch": 1.8645819873368652,
+      "grad_norm": 0.6503821015357971,
+      "learning_rate": 1.5685460147329917e-05,
+      "loss": 0.68,
+      "num_input_tokens_seen": 118210560,
+      "step": 14430
+    },
+    {
+      "epoch": 1.8658741439462463,
+      "grad_norm": 0.6401498913764954,
+      "learning_rate": 1.5654075780286742e-05,
+      "loss": 0.9481,
+      "num_input_tokens_seen": 118292480,
+      "step": 14440
+    },
+    {
+      "epoch": 1.8671663005556274,
+      "grad_norm": 0.5408946871757507,
+      "learning_rate": 1.562270852559661e-05,
+      "loss": 1.0229,
+      "num_input_tokens_seen": 118374400,
+      "step": 14450
+    },
+    {
+      "epoch": 1.8684584571650085,
+      "grad_norm": 0.7171612977981567,
+      "learning_rate": 1.5591358440692865e-05,
+      "loss": 1.0381,
+      "num_input_tokens_seen": 118456320,
+      "step": 14460
+    },
+    {
+      "epoch": 1.8697506137743893,
+      "grad_norm": 0.7857020497322083,
+      "learning_rate": 1.5560025582977377e-05,
+      "loss": 0.988,
+      "num_input_tokens_seen": 118538240,
+      "step": 14470
+    },
+    {
+      "epoch": 1.8710427703837706,
+      "grad_norm": 0.3020257353782654,
+      "learning_rate": 1.5528710009820513e-05,
+      "loss": 1.0819,
+      "num_input_tokens_seen": 118620160,
+      "step": 14480
+    },
+    {
+      "epoch": 1.8723349269931515,
+      "grad_norm": 0.6323096752166748,
+      "learning_rate": 1.5497411778560954e-05,
+      "loss": 1.2201,
+      "num_input_tokens_seen": 118702080,
+      "step": 14490
+    },
+    {
+      "epoch": 1.8736270836025326,
+      "grad_norm": 0.7573959827423096,
+      "learning_rate": 1.5466130946505664e-05,
+      "loss": 0.7508,
+      "num_input_tokens_seen": 118784000,
+      "step": 14500
+    },
+    {
+      "epoch": 1.8749192402119137,
+      "grad_norm": 0.48801571130752563,
+      "learning_rate": 1.5434867570929724e-05,
+      "loss": 0.8983,
+      "num_input_tokens_seen": 118865920,
+      "step": 14510
+    },
+    {
+      "epoch": 1.8762113968212948,
+      "grad_norm": 0.34075334668159485,
+      "learning_rate": 1.5403621709076247e-05,
+      "loss": 0.6051,
+      "num_input_tokens_seen": 118947840,
+      "step": 14520
+    },
+    {
+      "epoch": 1.8775035534306759,
+      "grad_norm": 0.27687662839889526,
+      "learning_rate": 1.5372393418156323e-05,
+      "loss": 0.7321,
+      "num_input_tokens_seen": 119029760,
+      "step": 14530
+    },
+    {
+      "epoch": 1.8787957100400567,
+      "grad_norm": 0.7669917941093445,
+      "learning_rate": 1.5341182755348806e-05,
+      "loss": 0.887,
+      "num_input_tokens_seen": 119111680,
+      "step": 14540
+    },
+    {
+      "epoch": 1.880087866649438,
+      "grad_norm": 0.30796754360198975,
+      "learning_rate": 1.530998977780033e-05,
+      "loss": 0.7769,
+      "num_input_tokens_seen": 119193600,
+      "step": 14550
+    },
+    {
+      "epoch": 1.881380023258819,
+      "grad_norm": 0.6400482654571533,
+      "learning_rate": 1.5278814542625107e-05,
+      "loss": 0.7488,
+      "num_input_tokens_seen": 119275520,
+      "step": 14560
+    },
+    {
+      "epoch": 1.8826721798682,
+      "grad_norm": 0.6054263114929199,
+      "learning_rate": 1.5247657106904891e-05,
+      "loss": 0.8852,
+      "num_input_tokens_seen": 119357440,
+      "step": 14570
+    },
+    {
+      "epoch": 1.883964336477581,
+      "grad_norm": 0.19290970265865326,
+      "learning_rate": 1.5216517527688818e-05,
+      "loss": 0.8831,
+      "num_input_tokens_seen": 119439360,
+      "step": 14580
+    },
+    {
+      "epoch": 1.8852564930869622,
+      "grad_norm": 0.5725274085998535,
+      "learning_rate": 1.5185395861993353e-05,
+      "loss": 0.8191,
+      "num_input_tokens_seen": 119521280,
+      "step": 14590
+    },
+    {
+      "epoch": 1.8865486496963433,
+      "grad_norm": 0.7249799370765686,
+      "learning_rate": 1.515429216680216e-05,
+      "loss": 0.9853,
+      "num_input_tokens_seen": 119603200,
+      "step": 14600
+    },
+    {
+      "epoch": 1.8878408063057242,
+      "grad_norm": 0.7958240509033203,
+      "learning_rate": 1.5123206499065967e-05,
+      "loss": 0.9514,
+      "num_input_tokens_seen": 119685120,
+      "step": 14610
+    },
+    {
+      "epoch": 1.8891329629151055,
+      "grad_norm": 0.6748232245445251,
+      "learning_rate": 1.5092138915702545e-05,
+      "loss": 1.2975,
+      "num_input_tokens_seen": 119767040,
+      "step": 14620
+    },
+    {
+      "epoch": 1.8904251195244863,
+      "grad_norm": 0.9683583974838257,
+      "learning_rate": 1.5061089473596501e-05,
+      "loss": 0.8597,
+      "num_input_tokens_seen": 119848960,
+      "step": 14630
+    },
+    {
+      "epoch": 1.8917172761338674,
+      "grad_norm": 0.6267436742782593,
+      "learning_rate": 1.5030058229599275e-05,
+      "loss": 0.7876,
+      "num_input_tokens_seen": 119930880,
+      "step": 14640
+    },
+    {
+      "epoch": 1.8930094327432485,
+      "grad_norm": 0.18290306627750397,
+      "learning_rate": 1.4999045240528935e-05,
+      "loss": 0.4448,
+      "num_input_tokens_seen": 120012800,
+      "step": 14650
+    },
+    {
+      "epoch": 1.8943015893526294,
+      "grad_norm": 0.908818244934082,
+      "learning_rate": 1.4968050563170177e-05,
+      "loss": 0.9315,
+      "num_input_tokens_seen": 120094720,
+      "step": 14660
+    },
+    {
+      "epoch": 1.8955937459620107,
+      "grad_norm": 0.8734130859375,
+      "learning_rate": 1.4937074254274117e-05,
+      "loss": 1.0145,
+      "num_input_tokens_seen": 120176640,
+      "step": 14670
+    },
+    {
+      "epoch": 1.8968859025713916,
+      "grad_norm": 0.6656931042671204,
+      "learning_rate": 1.4906116370558276e-05,
+      "loss": 0.6827,
+      "num_input_tokens_seen": 120258560,
+      "step": 14680
+    },
+    {
+      "epoch": 1.8981780591807729,
+      "grad_norm": 0.612580418586731,
+      "learning_rate": 1.4875176968706434e-05,
+      "loss": 0.8853,
+      "num_input_tokens_seen": 120340480,
+      "step": 14690
+    },
+    {
+      "epoch": 1.8994702157901537,
+      "grad_norm": 0.30025026202201843,
+      "learning_rate": 1.4844256105368504e-05,
+      "loss": 0.7183,
+      "num_input_tokens_seen": 120422400,
+      "step": 14700
+    },
+    {
+      "epoch": 1.9007623723995348,
+      "grad_norm": 0.3702358603477478,
+      "learning_rate": 1.4813353837160488e-05,
+      "loss": 0.5776,
+      "num_input_tokens_seen": 120504320,
+      "step": 14710
+    },
+    {
+      "epoch": 1.902054529008916,
+      "grad_norm": 0.5429428219795227,
+      "learning_rate": 1.4782470220664313e-05,
+      "loss": 1.1882,
+      "num_input_tokens_seen": 120586240,
+      "step": 14720
+    },
+    {
+      "epoch": 1.9033466856182968,
+      "grad_norm": 0.8188362717628479,
+      "learning_rate": 1.4751605312427786e-05,
+      "loss": 0.9617,
+      "num_input_tokens_seen": 120668160,
+      "step": 14730
+    },
+    {
+      "epoch": 1.904638842227678,
+      "grad_norm": 0.8028152585029602,
+      "learning_rate": 1.472075916896442e-05,
+      "loss": 0.8622,
+      "num_input_tokens_seen": 120750080,
+      "step": 14740
+    },
+    {
+      "epoch": 1.905930998837059,
+      "grad_norm": 0.839156985282898,
+      "learning_rate": 1.4689931846753402e-05,
+      "loss": 0.9725,
+      "num_input_tokens_seen": 120832000,
+      "step": 14750
+    },
+    {
+      "epoch": 1.90722315544644,
+      "grad_norm": 0.7380186319351196,
+      "learning_rate": 1.4659123402239454e-05,
+      "loss": 1.2234,
+      "num_input_tokens_seen": 120913920,
+      "step": 14760
+    },
+    {
+      "epoch": 1.9085153120558211,
+      "grad_norm": 0.5705264806747437,
+      "learning_rate": 1.4628333891832713e-05,
+      "loss": 0.8926,
+      "num_input_tokens_seen": 120995840,
+      "step": 14770
+    },
+    {
+      "epoch": 1.9098074686652022,
+      "grad_norm": 0.7131834626197815,
+      "learning_rate": 1.4597563371908663e-05,
+      "loss": 1.0101,
+      "num_input_tokens_seen": 121077760,
+      "step": 14780
+    },
+    {
+      "epoch": 1.9110996252745833,
+      "grad_norm": 0.9028184413909912,
+      "learning_rate": 1.4566811898808013e-05,
+      "loss": 0.9358,
+      "num_input_tokens_seen": 121159680,
+      "step": 14790
+    },
+    {
+      "epoch": 1.9123917818839642,
+      "grad_norm": 0.5184687972068787,
+      "learning_rate": 1.4536079528836605e-05,
+      "loss": 0.7874,
+      "num_input_tokens_seen": 121241600,
+      "step": 14800
+    },
+    {
+      "epoch": 1.9136839384933455,
+      "grad_norm": 0.6445053815841675,
+      "learning_rate": 1.4505366318265278e-05,
+      "loss": 0.6709,
+      "num_input_tokens_seen": 121323520,
+      "step": 14810
+    },
+    {
+      "epoch": 1.9149760951027264,
+      "grad_norm": 0.3584889769554138,
+      "learning_rate": 1.4474672323329819e-05,
+      "loss": 0.6944,
+      "num_input_tokens_seen": 121405440,
+      "step": 14820
+    },
+    {
+      "epoch": 1.9162682517121075,
+      "grad_norm": 0.5627390742301941,
+      "learning_rate": 1.4443997600230832e-05,
+      "loss": 0.8012,
+      "num_input_tokens_seen": 121487360,
+      "step": 14830
+    },
+    {
+      "epoch": 1.9175604083214886,
+      "grad_norm": 0.6850423216819763,
+      "learning_rate": 1.4413342205133604e-05,
+      "loss": 0.9204,
+      "num_input_tokens_seen": 121569280,
+      "step": 14840
+    },
+    {
+      "epoch": 1.9188525649308696,
+      "grad_norm": 0.7621904611587524,
+      "learning_rate": 1.4382706194168066e-05,
+      "loss": 0.6648,
+      "num_input_tokens_seen": 121651200,
+      "step": 14850
+    },
+    {
+      "epoch": 1.9201447215402507,
+      "grad_norm": 0.636451244354248,
+      "learning_rate": 1.4352089623428627e-05,
+      "loss": 0.8368,
+      "num_input_tokens_seen": 121733120,
+      "step": 14860
+    },
+    {
+      "epoch": 1.9214368781496316,
+      "grad_norm": 0.5977798700332642,
+      "learning_rate": 1.4321492548974137e-05,
+      "loss": 1.248,
+      "num_input_tokens_seen": 121815040,
+      "step": 14870
+    },
+    {
+      "epoch": 1.922729034759013,
+      "grad_norm": 0.24377988278865814,
+      "learning_rate": 1.42909150268277e-05,
+      "loss": 1.0932,
+      "num_input_tokens_seen": 121896960,
+      "step": 14880
+    },
+    {
+      "epoch": 1.9240211913683938,
+      "grad_norm": 0.6114721894264221,
+      "learning_rate": 1.4260357112976664e-05,
+      "loss": 0.8562,
+      "num_input_tokens_seen": 121978880,
+      "step": 14890
+    },
+    {
+      "epoch": 1.9253133479777749,
+      "grad_norm": 0.5565524101257324,
+      "learning_rate": 1.4229818863372463e-05,
+      "loss": 0.9967,
+      "num_input_tokens_seen": 122060800,
+      "step": 14900
+    },
+    {
+      "epoch": 1.926605504587156,
+      "grad_norm": 0.9591652154922485,
+      "learning_rate": 1.4199300333930515e-05,
+      "loss": 1.2298,
+      "num_input_tokens_seen": 122142720,
+      "step": 14910
+    },
+    {
+      "epoch": 1.927897661196537,
+      "grad_norm": 0.7166655659675598,
+      "learning_rate": 1.4168801580530119e-05,
+      "loss": 0.6193,
+      "num_input_tokens_seen": 122224640,
+      "step": 14920
+    },
+    {
+      "epoch": 1.9291898178059181,
+      "grad_norm": 0.5449065566062927,
+      "learning_rate": 1.4138322659014408e-05,
+      "loss": 0.9493,
+      "num_input_tokens_seen": 122306560,
+      "step": 14930
+    },
+    {
+      "epoch": 1.930481974415299,
+      "grad_norm": 1.1408071517944336,
+      "learning_rate": 1.4107863625190163e-05,
+      "loss": 0.8788,
+      "num_input_tokens_seen": 122388480,
+      "step": 14940
+    },
+    {
+      "epoch": 1.9317741310246803,
+      "grad_norm": 0.3808801770210266,
+      "learning_rate": 1.4077424534827752e-05,
+      "loss": 0.7922,
+      "num_input_tokens_seen": 122470400,
+      "step": 14950
+    },
+    {
+      "epoch": 1.9330662876340612,
+      "grad_norm": 0.7727448344230652,
+      "learning_rate": 1.4047005443661048e-05,
+      "loss": 1.057,
+      "num_input_tokens_seen": 122552320,
+      "step": 14960
+    },
+    {
+      "epoch": 1.9343584442434423,
+      "grad_norm": 0.7547351121902466,
+      "learning_rate": 1.4016606407387312e-05,
+      "loss": 0.9366,
+      "num_input_tokens_seen": 122634240,
+      "step": 14970
+    },
+    {
+      "epoch": 1.9356506008528234,
+      "grad_norm": 0.9177370667457581,
+      "learning_rate": 1.398622748166704e-05,
+      "loss": 0.3856,
+      "num_input_tokens_seen": 122716160,
+      "step": 14980
+    },
+    {
+      "epoch": 1.9369427574622045,
+      "grad_norm": 0.519792914390564,
+      "learning_rate": 1.3955868722123955e-05,
+      "loss": 0.7705,
+      "num_input_tokens_seen": 122798080,
+      "step": 14990
+    },
+    {
+      "epoch": 1.9382349140715855,
+      "grad_norm": 0.8138200044631958,
+      "learning_rate": 1.3925530184344818e-05,
+      "loss": 0.7423,
+      "num_input_tokens_seen": 122880000,
+      "step": 15000
+    },
+    {
+      "epoch": 1.9395270706809664,
+      "grad_norm": 0.6865666508674622,
+      "learning_rate": 1.3895211923879397e-05,
+      "loss": 1.0697,
+      "num_input_tokens_seen": 122961920,
+      "step": 15010
+    },
+    {
+      "epoch": 1.9408192272903477,
+      "grad_norm": 0.6570258736610413,
+      "learning_rate": 1.3864913996240304e-05,
+      "loss": 0.967,
+      "num_input_tokens_seen": 123043840,
+      "step": 15020
+    },
+    {
+      "epoch": 1.9421113838997286,
+      "grad_norm": 1.129082441329956,
+      "learning_rate": 1.3834636456902944e-05,
+      "loss": 0.8996,
+      "num_input_tokens_seen": 123125760,
+      "step": 15030
+    },
+    {
+      "epoch": 1.9434035405091097,
+      "grad_norm": 0.1828085333108902,
+      "learning_rate": 1.3804379361305363e-05,
+      "loss": 0.5727,
+      "num_input_tokens_seen": 123207680,
+      "step": 15040
+    },
+    {
+      "epoch": 1.9446956971184908,
+      "grad_norm": 0.5514014363288879,
+      "learning_rate": 1.3774142764848207e-05,
+      "loss": 0.9822,
+      "num_input_tokens_seen": 123289600,
+      "step": 15050
+    },
+    {
+      "epoch": 1.9459878537278719,
+      "grad_norm": 0.51252681016922,
+      "learning_rate": 1.3743926722894579e-05,
+      "loss": 0.7735,
+      "num_input_tokens_seen": 123371520,
+      "step": 15060
+    },
+    {
+      "epoch": 1.947280010337253,
+      "grad_norm": 0.6373499631881714,
+      "learning_rate": 1.3713731290769921e-05,
+      "loss": 1.0148,
+      "num_input_tokens_seen": 123453440,
+      "step": 15070
+    },
+    {
+      "epoch": 1.9485721669466338,
+      "grad_norm": 0.7882586121559143,
+      "learning_rate": 1.3683556523761981e-05,
+      "loss": 0.8874,
+      "num_input_tokens_seen": 123535360,
+      "step": 15080
+    },
+    {
+      "epoch": 1.9498643235560151,
+      "grad_norm": 0.8441190719604492,
+      "learning_rate": 1.365340247712064e-05,
+      "loss": 0.9097,
+      "num_input_tokens_seen": 123617280,
+      "step": 15090
+    },
+    {
+      "epoch": 1.951156480165396,
+      "grad_norm": 0.6517019271850586,
+      "learning_rate": 1.362326920605783e-05,
+      "loss": 0.8473,
+      "num_input_tokens_seen": 123699200,
+      "step": 15100
+    },
+    {
+      "epoch": 1.952448636774777,
+      "grad_norm": 0.4873802363872528,
+      "learning_rate": 1.3593156765747483e-05,
+      "loss": 0.7964,
+      "num_input_tokens_seen": 123781120,
+      "step": 15110
+    },
+    {
+      "epoch": 1.9537407933841582,
+      "grad_norm": 0.6770155429840088,
+      "learning_rate": 1.3563065211325349e-05,
+      "loss": 0.795,
+      "num_input_tokens_seen": 123863040,
+      "step": 15120
+    },
+    {
+      "epoch": 1.9550329499935393,
+      "grad_norm": 0.4398241937160492,
+      "learning_rate": 1.3532994597888971e-05,
+      "loss": 0.6041,
+      "num_input_tokens_seen": 123944960,
+      "step": 15130
+    },
+    {
+      "epoch": 1.9563251066029204,
+      "grad_norm": 0.2949928939342499,
+      "learning_rate": 1.3502944980497514e-05,
+      "loss": 0.7019,
+      "num_input_tokens_seen": 124026880,
+      "step": 15140
+    },
+    {
+      "epoch": 1.9576172632123012,
+      "grad_norm": 0.7733688354492188,
+      "learning_rate": 1.3472916414171738e-05,
+      "loss": 1.2308,
+      "num_input_tokens_seen": 124108800,
+      "step": 15150
+    },
+    {
+      "epoch": 1.9589094198216825,
+      "grad_norm": 0.6464848518371582,
+      "learning_rate": 1.3442908953893816e-05,
+      "loss": 0.8531,
+      "num_input_tokens_seen": 124190720,
+      "step": 15160
+    },
+    {
+      "epoch": 1.9602015764310634,
+      "grad_norm": 0.7353049516677856,
+      "learning_rate": 1.3412922654607318e-05,
+      "loss": 0.8579,
+      "num_input_tokens_seen": 124272640,
+      "step": 15170
+    },
+    {
+      "epoch": 1.9614937330404445,
+      "grad_norm": 0.8654274344444275,
+      "learning_rate": 1.338295757121703e-05,
+      "loss": 0.7652,
+      "num_input_tokens_seen": 124354560,
+      "step": 15180
+    },
+    {
+      "epoch": 1.9627858896498256,
+      "grad_norm": 0.5775660872459412,
+      "learning_rate": 1.3353013758588923e-05,
+      "loss": 0.8972,
+      "num_input_tokens_seen": 124436480,
+      "step": 15190
+    },
+    {
+      "epoch": 1.9640780462592065,
+      "grad_norm": 0.5791903138160706,
+      "learning_rate": 1.3323091271550011e-05,
+      "loss": 0.8894,
+      "num_input_tokens_seen": 124518400,
+      "step": 15200
+    },
+    {
+      "epoch": 1.9653702028685878,
+      "grad_norm": 0.8218627572059631,
+      "learning_rate": 1.3293190164888242e-05,
+      "loss": 1.0825,
+      "num_input_tokens_seen": 124600320,
+      "step": 15210
+    },
+    {
+      "epoch": 1.9666623594779686,
+      "grad_norm": 0.3494950234889984,
+      "learning_rate": 1.3263310493352454e-05,
+      "loss": 0.8476,
+      "num_input_tokens_seen": 124682240,
+      "step": 15220
+    },
+    {
+      "epoch": 1.96795451608735,
+      "grad_norm": 0.5643335580825806,
+      "learning_rate": 1.3233452311652197e-05,
+      "loss": 0.6004,
+      "num_input_tokens_seen": 124764160,
+      "step": 15230
+    },
+    {
+      "epoch": 1.9692466726967308,
+      "grad_norm": 0.6971800923347473,
+      "learning_rate": 1.3203615674457709e-05,
+      "loss": 0.4468,
+      "num_input_tokens_seen": 124846080,
+      "step": 15240
+    },
+    {
+      "epoch": 1.970538829306112,
+      "grad_norm": 0.690518319606781,
+      "learning_rate": 1.3173800636399744e-05,
+      "loss": 0.704,
+      "num_input_tokens_seen": 124928000,
+      "step": 15250
+    },
+    {
+      "epoch": 1.971830985915493,
+      "grad_norm": 0.568469226360321,
+      "learning_rate": 1.3144007252069552e-05,
+      "loss": 0.5738,
+      "num_input_tokens_seen": 125009920,
+      "step": 15260
+    },
+    {
+      "epoch": 1.9731231425248739,
+      "grad_norm": 0.6457722783088684,
+      "learning_rate": 1.3114235576018686e-05,
+      "loss": 0.6915,
+      "num_input_tokens_seen": 125091840,
+      "step": 15270
+    },
+    {
+      "epoch": 1.9744152991342552,
+      "grad_norm": 0.6253653168678284,
+      "learning_rate": 1.3084485662758994e-05,
+      "loss": 0.6731,
+      "num_input_tokens_seen": 125173760,
+      "step": 15280
+    },
+    {
+      "epoch": 1.975707455743636,
+      "grad_norm": 0.6036949753761292,
+      "learning_rate": 1.3054757566762454e-05,
+      "loss": 0.8532,
+      "num_input_tokens_seen": 125255680,
+      "step": 15290
+    },
+    {
+      "epoch": 1.9769996123530171,
+      "grad_norm": 1.014550805091858,
+      "learning_rate": 1.3025051342461087e-05,
+      "loss": 1.3152,
+      "num_input_tokens_seen": 125337600,
+      "step": 15300
+    },
+    {
+      "epoch": 1.9782917689623982,
+      "grad_norm": 0.5557976961135864,
+      "learning_rate": 1.2995367044246903e-05,
+      "loss": 0.6177,
+      "num_input_tokens_seen": 125419520,
+      "step": 15310
+    },
+    {
+      "epoch": 1.9795839255717793,
+      "grad_norm": 1.1027765274047852,
+      "learning_rate": 1.2965704726471729e-05,
+      "loss": 0.7341,
+      "num_input_tokens_seen": 125501440,
+      "step": 15320
+    },
+    {
+      "epoch": 1.9808760821811604,
+      "grad_norm": 0.8762277364730835,
+      "learning_rate": 1.2936064443447157e-05,
+      "loss": 0.7912,
+      "num_input_tokens_seen": 125583360,
+      "step": 15330
+    },
+    {
+      "epoch": 1.9821682387905413,
+      "grad_norm": 0.5100016593933105,
+      "learning_rate": 1.2906446249444457e-05,
+      "loss": 1.0051,
+      "num_input_tokens_seen": 125665280,
+      "step": 15340
+    },
+    {
+      "epoch": 1.9834603953999226,
+      "grad_norm": 0.583299994468689,
+      "learning_rate": 1.2876850198694409e-05,
+      "loss": 0.4944,
+      "num_input_tokens_seen": 125747200,
+      "step": 15350
+    },
+    {
+      "epoch": 1.9847525520093035,
+      "grad_norm": 0.277055025100708,
+      "learning_rate": 1.2847276345387299e-05,
+      "loss": 0.7472,
+      "num_input_tokens_seen": 125829120,
+      "step": 15360
+    },
+    {
+      "epoch": 1.9860447086186845,
+      "grad_norm": 1.0343480110168457,
+      "learning_rate": 1.2817724743672715e-05,
+      "loss": 0.6023,
+      "num_input_tokens_seen": 125911040,
+      "step": 15370
+    },
+    {
+      "epoch": 1.9873368652280656,
+      "grad_norm": 0.6616178750991821,
+      "learning_rate": 1.2788195447659562e-05,
+      "loss": 1.1141,
+      "num_input_tokens_seen": 125992960,
+      "step": 15380
+    },
+    {
+      "epoch": 1.9886290218374467,
+      "grad_norm": 0.5394881367683411,
+      "learning_rate": 1.2758688511415848e-05,
+      "loss": 0.9213,
+      "num_input_tokens_seen": 126074880,
+      "step": 15390
+    },
+    {
+      "epoch": 1.9899211784468278,
+      "grad_norm": 0.7356629967689514,
+      "learning_rate": 1.2729203988968674e-05,
+      "loss": 0.7153,
+      "num_input_tokens_seen": 126156800,
+      "step": 15400
+    },
+    {
+      "epoch": 1.9912133350562087,
+      "grad_norm": 0.22516459226608276,
+      "learning_rate": 1.2699741934304104e-05,
+      "loss": 0.5926,
+      "num_input_tokens_seen": 126238720,
+      "step": 15410
+    },
+    {
+      "epoch": 1.99250549166559,
+      "grad_norm": 0.6779276728630066,
+      "learning_rate": 1.2670302401367035e-05,
+      "loss": 0.9406,
+      "num_input_tokens_seen": 126320640,
+      "step": 15420
+    },
+    {
+      "epoch": 1.9937976482749709,
+      "grad_norm": 0.5746923089027405,
+      "learning_rate": 1.2640885444061163e-05,
+      "loss": 0.9174,
+      "num_input_tokens_seen": 126402560,
+      "step": 15430
+    },
+    {
+      "epoch": 1.995089804884352,
+      "grad_norm": 0.6525429487228394,
+      "learning_rate": 1.2611491116248802e-05,
+      "loss": 1.254,
+      "num_input_tokens_seen": 126484480,
+      "step": 15440
+    },
+    {
+      "epoch": 1.996381961493733,
+      "grad_norm": 0.30466407537460327,
+      "learning_rate": 1.2582119471750888e-05,
+      "loss": 0.7516,
+      "num_input_tokens_seen": 126566400,
+      "step": 15450
+    },
+    {
+      "epoch": 1.9976741181031141,
+      "grad_norm": 0.6706373691558838,
+      "learning_rate": 1.2552770564346781e-05,
+      "loss": 0.9337,
+      "num_input_tokens_seen": 126648320,
+      "step": 15460
+    },
+    {
+      "epoch": 1.9989662747124952,
+      "grad_norm": 0.7642003297805786,
+      "learning_rate": 1.2523444447774213e-05,
+      "loss": 0.9429,
+      "num_input_tokens_seen": 126730240,
+      "step": 15470
+    },
+    {
+      "epoch": 2.000258431321876,
+      "grad_norm": 0.71830153465271,
+      "learning_rate": 1.2494141175729216e-05,
+      "loss": 0.9215,
+      "num_input_tokens_seen": 126812160,
+      "step": 15480
+    },
+    {
+      "epoch": 2.0015505879312574,
+      "grad_norm": 0.5191543698310852,
+      "learning_rate": 1.2464860801865954e-05,
+      "loss": 0.611,
+      "num_input_tokens_seen": 126894080,
+      "step": 15490
+    },
+    {
+      "epoch": 2.0028427445406383,
+      "grad_norm": 0.8038392066955566,
+      "learning_rate": 1.2435603379796704e-05,
+      "loss": 1.1059,
+      "num_input_tokens_seen": 126976000,
+      "step": 15500
+    },
+    {
+      "epoch": 2.0041349011500196,
+      "grad_norm": 0.5426739454269409,
+      "learning_rate": 1.240636896309168e-05,
+      "loss": 0.9048,
+      "num_input_tokens_seen": 127057920,
+      "step": 15510
+    },
+    {
+      "epoch": 2.0054270577594004,
+      "grad_norm": 0.6243565082550049,
+      "learning_rate": 1.237715760527901e-05,
+      "loss": 0.7904,
+      "num_input_tokens_seen": 127139840,
+      "step": 15520
+    },
+    {
+      "epoch": 2.0067192143687813,
+      "grad_norm": 0.35670045018196106,
+      "learning_rate": 1.2347969359844566e-05,
+      "loss": 0.7196,
+      "num_input_tokens_seen": 127221760,
+      "step": 15530
+    },
+    {
+      "epoch": 2.0080113709781626,
+      "grad_norm": 0.9198890924453735,
+      "learning_rate": 1.2318804280231939e-05,
+      "loss": 1.102,
+      "num_input_tokens_seen": 127303680,
+      "step": 15540
+    },
+    {
+      "epoch": 2.0093035275875435,
+      "grad_norm": 0.5498625636100769,
+      "learning_rate": 1.2289662419842258e-05,
+      "loss": 0.5505,
+      "num_input_tokens_seen": 127385600,
+      "step": 15550
+    },
+    {
+      "epoch": 2.010595684196925,
+      "grad_norm": 0.872312068939209,
+      "learning_rate": 1.2260543832034177e-05,
+      "loss": 0.6824,
+      "num_input_tokens_seen": 127467520,
+      "step": 15560
+    },
+    {
+      "epoch": 2.0118878408063057,
+      "grad_norm": 0.6140616536140442,
+      "learning_rate": 1.2231448570123732e-05,
+      "loss": 0.9454,
+      "num_input_tokens_seen": 127549440,
+      "step": 15570
+    },
+    {
+      "epoch": 2.013179997415687,
+      "grad_norm": 0.32139700651168823,
+      "learning_rate": 1.2202376687384223e-05,
+      "loss": 0.9467,
+      "num_input_tokens_seen": 127631360,
+      "step": 15580
+    },
+    {
+      "epoch": 2.014472154025068,
+      "grad_norm": 0.5901627540588379,
+      "learning_rate": 1.2173328237046178e-05,
+      "loss": 0.6482,
+      "num_input_tokens_seen": 127713280,
+      "step": 15590
+    },
+    {
+      "epoch": 2.0157643106344487,
+      "grad_norm": 0.9374563694000244,
+      "learning_rate": 1.2144303272297186e-05,
+      "loss": 0.547,
+      "num_input_tokens_seen": 127795200,
+      "step": 15600
+    },
+    {
+      "epoch": 2.01705646724383,
+      "grad_norm": 0.6503939032554626,
+      "learning_rate": 1.2115301846281871e-05,
+      "loss": 0.9459,
+      "num_input_tokens_seen": 127877120,
+      "step": 15610
+    },
+    {
+      "epoch": 2.018348623853211,
+      "grad_norm": 0.645089864730835,
+      "learning_rate": 1.2086324012101716e-05,
+      "loss": 0.8437,
+      "num_input_tokens_seen": 127959040,
+      "step": 15620
+    },
+    {
+      "epoch": 2.019640780462592,
+      "grad_norm": 0.6811534762382507,
+      "learning_rate": 1.2057369822815051e-05,
+      "loss": 0.9276,
+      "num_input_tokens_seen": 128040960,
+      "step": 15630
+    },
+    {
+      "epoch": 2.020932937071973,
+      "grad_norm": 0.5539160370826721,
+      "learning_rate": 1.2028439331436869e-05,
+      "loss": 0.8422,
+      "num_input_tokens_seen": 128122880,
+      "step": 15640
+    },
+    {
+      "epoch": 2.0222250936813544,
+      "grad_norm": 0.5917775630950928,
+      "learning_rate": 1.1999532590938817e-05,
+      "loss": 0.9474,
+      "num_input_tokens_seen": 128204800,
+      "step": 15650
+    },
+    {
+      "epoch": 2.0235172502907353,
+      "grad_norm": 0.35007065534591675,
+      "learning_rate": 1.1970649654249017e-05,
+      "loss": 0.5438,
+      "num_input_tokens_seen": 128286720,
+      "step": 15660
+    },
+    {
+      "epoch": 2.024809406900116,
+      "grad_norm": 0.8497019410133362,
+      "learning_rate": 1.1941790574252013e-05,
+      "loss": 0.5885,
+      "num_input_tokens_seen": 128368640,
+      "step": 15670
+    },
+    {
+      "epoch": 2.0261015635094974,
+      "grad_norm": 0.6658693552017212,
+      "learning_rate": 1.1912955403788695e-05,
+      "loss": 0.878,
+      "num_input_tokens_seen": 128450560,
+      "step": 15680
+    },
+    {
+      "epoch": 2.0273937201188783,
+      "grad_norm": 0.6294447779655457,
+      "learning_rate": 1.1884144195656133e-05,
+      "loss": 0.7301,
+      "num_input_tokens_seen": 128532480,
+      "step": 15690
+    },
+    {
+      "epoch": 2.0286858767282596,
+      "grad_norm": 1.484786868095398,
+      "learning_rate": 1.1855357002607556e-05,
+      "loss": 0.3667,
+      "num_input_tokens_seen": 128614400,
+      "step": 15700
+    },
+    {
+      "epoch": 2.0299780333376405,
+      "grad_norm": 1.5291662216186523,
+      "learning_rate": 1.1826593877352216e-05,
+      "loss": 0.6625,
+      "num_input_tokens_seen": 128696320,
+      "step": 15710
+    },
+    {
+      "epoch": 2.031270189947022,
+      "grad_norm": 0.41958874464035034,
+      "learning_rate": 1.1797854872555272e-05,
+      "loss": 0.8176,
+      "num_input_tokens_seen": 128778240,
+      "step": 15720
+    },
+    {
+      "epoch": 2.0325623465564027,
+      "grad_norm": 0.7266905307769775,
+      "learning_rate": 1.1769140040837755e-05,
+      "loss": 0.8307,
+      "num_input_tokens_seen": 128860160,
+      "step": 15730
+    },
+    {
+      "epoch": 2.0338545031657835,
+      "grad_norm": 0.606711208820343,
+      "learning_rate": 1.1740449434776402e-05,
+      "loss": 0.4084,
+      "num_input_tokens_seen": 128942080,
+      "step": 15740
+    },
+    {
+      "epoch": 2.035146659775165,
+      "grad_norm": 0.7246209979057312,
+      "learning_rate": 1.171178310690362e-05,
+      "loss": 0.664,
+      "num_input_tokens_seen": 129024000,
+      "step": 15750
+    },
+    {
+      "epoch": 2.0364388163845457,
+      "grad_norm": 0.588646650314331,
+      "learning_rate": 1.1683141109707339e-05,
+      "loss": 0.7079,
+      "num_input_tokens_seen": 129105920,
+      "step": 15760
+    },
+    {
+      "epoch": 2.037730972993927,
+      "grad_norm": 0.6580897569656372,
+      "learning_rate": 1.165452349563095e-05,
+      "loss": 0.9028,
+      "num_input_tokens_seen": 129187840,
+      "step": 15770
+    },
+    {
+      "epoch": 2.039023129603308,
+      "grad_norm": 0.7622004747390747,
+      "learning_rate": 1.1625930317073221e-05,
+      "loss": 0.6014,
+      "num_input_tokens_seen": 129269760,
+      "step": 15780
+    },
+    {
+      "epoch": 2.0403152862126888,
+      "grad_norm": 0.7260866761207581,
+      "learning_rate": 1.159736162638813e-05,
+      "loss": 0.7052,
+      "num_input_tokens_seen": 129351680,
+      "step": 15790
+    },
+    {
+      "epoch": 2.04160744282207,
+      "grad_norm": 0.6975520849227905,
+      "learning_rate": 1.1568817475884868e-05,
+      "loss": 0.5514,
+      "num_input_tokens_seen": 129433600,
+      "step": 15800
+    },
+    {
+      "epoch": 2.042899599431451,
+      "grad_norm": 1.0300558805465698,
+      "learning_rate": 1.154029791782765e-05,
+      "loss": 0.9024,
+      "num_input_tokens_seen": 129515520,
+      "step": 15810
+    },
+    {
+      "epoch": 2.0441917560408323,
+      "grad_norm": 0.5849384069442749,
+      "learning_rate": 1.1511803004435704e-05,
+      "loss": 0.8982,
+      "num_input_tokens_seen": 129597440,
+      "step": 15820
+    },
+    {
+      "epoch": 2.045483912650213,
+      "grad_norm": 0.6071699857711792,
+      "learning_rate": 1.1483332787883096e-05,
+      "loss": 0.8733,
+      "num_input_tokens_seen": 129679360,
+      "step": 15830
+    },
+    {
+      "epoch": 2.0467760692595944,
+      "grad_norm": 0.26067107915878296,
+      "learning_rate": 1.1454887320298686e-05,
+      "loss": 0.7505,
+      "num_input_tokens_seen": 129761280,
+      "step": 15840
+    },
+    {
+      "epoch": 2.0480682258689753,
+      "grad_norm": 0.8173131942749023,
+      "learning_rate": 1.1426466653766036e-05,
+      "loss": 0.6831,
+      "num_input_tokens_seen": 129843200,
+      "step": 15850
+    },
+    {
+      "epoch": 2.049360382478356,
+      "grad_norm": 0.7379295229911804,
+      "learning_rate": 1.1398070840323264e-05,
+      "loss": 0.8099,
+      "num_input_tokens_seen": 129925120,
+      "step": 15860
+    },
+    {
+      "epoch": 2.0506525390877375,
+      "grad_norm": 0.6924763917922974,
+      "learning_rate": 1.1369699931963018e-05,
+      "loss": 0.7674,
+      "num_input_tokens_seen": 130007040,
+      "step": 15870
+    },
+    {
+      "epoch": 2.0519446956971183,
+      "grad_norm": 0.8446022272109985,
+      "learning_rate": 1.1341353980632313e-05,
+      "loss": 0.9515,
+      "num_input_tokens_seen": 130088960,
+      "step": 15880
+    },
+    {
+      "epoch": 2.0532368523064997,
+      "grad_norm": 0.9424275755882263,
+      "learning_rate": 1.1313033038232498e-05,
+      "loss": 0.9029,
+      "num_input_tokens_seen": 130170880,
+      "step": 15890
+    },
+    {
+      "epoch": 2.0545290089158805,
+      "grad_norm": 0.9581369161605835,
+      "learning_rate": 1.1284737156619096e-05,
+      "loss": 0.7374,
+      "num_input_tokens_seen": 130252800,
+      "step": 15900
+    },
+    {
+      "epoch": 2.055821165525262,
+      "grad_norm": 0.6615799069404602,
+      "learning_rate": 1.1256466387601782e-05,
+      "loss": 0.7558,
+      "num_input_tokens_seen": 130334720,
+      "step": 15910
+    },
+    {
+      "epoch": 2.0571133221346427,
+      "grad_norm": 0.3666354715824127,
+      "learning_rate": 1.1228220782944212e-05,
+      "loss": 0.7905,
+      "num_input_tokens_seen": 130416640,
+      "step": 15920
+    },
+    {
+      "epoch": 2.0584054787440236,
+      "grad_norm": 0.6414427757263184,
+      "learning_rate": 1.1200000394363996e-05,
+      "loss": 0.9914,
+      "num_input_tokens_seen": 130498560,
+      "step": 15930
+    },
+    {
+      "epoch": 2.059697635353405,
+      "grad_norm": 0.23354114592075348,
+      "learning_rate": 1.1171805273532567e-05,
+      "loss": 0.7977,
+      "num_input_tokens_seen": 130580480,
+      "step": 15940
+    },
+    {
+      "epoch": 2.0609897919627858,
+      "grad_norm": 0.4102308750152588,
+      "learning_rate": 1.1143635472075074e-05,
+      "loss": 0.8725,
+      "num_input_tokens_seen": 130662400,
+      "step": 15950
+    },
+    {
+      "epoch": 2.062281948572167,
+      "grad_norm": 1.2794528007507324,
+      "learning_rate": 1.1115491041570337e-05,
+      "loss": 1.0455,
+      "num_input_tokens_seen": 130744320,
+      "step": 15960
+    },
+    {
+      "epoch": 2.063574105181548,
+      "grad_norm": 0.6293236613273621,
+      "learning_rate": 1.1087372033550685e-05,
+      "loss": 0.8646,
+      "num_input_tokens_seen": 130826240,
+      "step": 15970
+    },
+    {
+      "epoch": 2.0648662617909292,
+      "grad_norm": 0.7114168405532837,
+      "learning_rate": 1.105927849950194e-05,
+      "loss": 0.9453,
+      "num_input_tokens_seen": 130908160,
+      "step": 15980
+    },
+    {
+      "epoch": 2.06615841840031,
+      "grad_norm": 0.9560719728469849,
+      "learning_rate": 1.103121049086324e-05,
+      "loss": 0.822,
+      "num_input_tokens_seen": 130990080,
+      "step": 15990
+    },
+    {
+      "epoch": 2.067450575009691,
+      "grad_norm": 0.36301013827323914,
+      "learning_rate": 1.1003168059027025e-05,
+      "loss": 0.8453,
+      "num_input_tokens_seen": 131072000,
+      "step": 16000
+    },
+    {
+      "epoch": 2.0687427316190723,
+      "grad_norm": 0.5861186385154724,
+      "learning_rate": 1.0975151255338867e-05,
+      "loss": 0.9873,
+      "num_input_tokens_seen": 131153920,
+      "step": 16010
+    },
+    {
+      "epoch": 2.070034888228453,
+      "grad_norm": 0.6049230098724365,
+      "learning_rate": 1.094716013109745e-05,
+      "loss": 0.6615,
+      "num_input_tokens_seen": 131235840,
+      "step": 16020
+    },
+    {
+      "epoch": 2.0713270448378345,
+      "grad_norm": 0.3231169283390045,
+      "learning_rate": 1.0919194737554409e-05,
+      "loss": 0.5598,
+      "num_input_tokens_seen": 131317760,
+      "step": 16030
+    },
+    {
+      "epoch": 2.0726192014472153,
+      "grad_norm": 0.7107923030853271,
+      "learning_rate": 1.0891255125914269e-05,
+      "loss": 0.8139,
+      "num_input_tokens_seen": 131399680,
+      "step": 16040
+    },
+    {
+      "epoch": 2.0739113580565967,
+      "grad_norm": 0.196472629904747,
+      "learning_rate": 1.0863341347334376e-05,
+      "loss": 0.7574,
+      "num_input_tokens_seen": 131481600,
+      "step": 16050
+    },
+    {
+      "epoch": 2.0752035146659775,
+      "grad_norm": 0.9992514848709106,
+      "learning_rate": 1.0835453452924737e-05,
+      "loss": 0.578,
+      "num_input_tokens_seen": 131563520,
+      "step": 16060
+    },
+    {
+      "epoch": 2.0764956712753584,
+      "grad_norm": 0.8521533608436584,
+      "learning_rate": 1.0807591493747992e-05,
+      "loss": 0.6804,
+      "num_input_tokens_seen": 131645440,
+      "step": 16070
+    },
+    {
+      "epoch": 2.0777878278847397,
+      "grad_norm": 0.5789138078689575,
+      "learning_rate": 1.0779755520819302e-05,
+      "loss": 0.5871,
+      "num_input_tokens_seen": 131727360,
+      "step": 16080
+    },
+    {
+      "epoch": 2.0790799844941206,
+      "grad_norm": 0.27690398693084717,
+      "learning_rate": 1.0751945585106205e-05,
+      "loss": 0.8846,
+      "num_input_tokens_seen": 131809280,
+      "step": 16090
+    },
+    {
+      "epoch": 2.080372141103502,
+      "grad_norm": 0.5583693981170654,
+      "learning_rate": 1.0724161737528616e-05,
+      "loss": 0.9714,
+      "num_input_tokens_seen": 131891200,
+      "step": 16100
+    },
+    {
+      "epoch": 2.0816642977128827,
+      "grad_norm": 1.2210655212402344,
+      "learning_rate": 1.0696404028958634e-05,
+      "loss": 0.7469,
+      "num_input_tokens_seen": 131973120,
+      "step": 16110
+    },
+    {
+      "epoch": 2.082956454322264,
+      "grad_norm": 1.0592528581619263,
+      "learning_rate": 1.0668672510220548e-05,
+      "loss": 1.0468,
+      "num_input_tokens_seen": 132055040,
+      "step": 16120
+    },
+    {
+      "epoch": 2.084248610931645,
+      "grad_norm": 0.7899770140647888,
+      "learning_rate": 1.0640967232090643e-05,
+      "loss": 0.4886,
+      "num_input_tokens_seen": 132136960,
+      "step": 16130
+    },
+    {
+      "epoch": 2.085540767541026,
+      "grad_norm": 0.5646101832389832,
+      "learning_rate": 1.0613288245297193e-05,
+      "loss": 0.9022,
+      "num_input_tokens_seen": 132218880,
+      "step": 16140
+    },
+    {
+      "epoch": 2.086832924150407,
+      "grad_norm": 0.5926666855812073,
+      "learning_rate": 1.0585635600520327e-05,
+      "loss": 0.9351,
+      "num_input_tokens_seen": 132300800,
+      "step": 16150
+    },
+    {
+      "epoch": 2.088125080759788,
+      "grad_norm": 1.35215163230896,
+      "learning_rate": 1.0558009348391926e-05,
+      "loss": 0.7735,
+      "num_input_tokens_seen": 132382720,
+      "step": 16160
+    },
+    {
+      "epoch": 2.0894172373691693,
+      "grad_norm": 0.7683168649673462,
+      "learning_rate": 1.053040953949557e-05,
+      "loss": 0.6433,
+      "num_input_tokens_seen": 132464640,
+      "step": 16170
+    },
+    {
+      "epoch": 2.09070939397855,
+      "grad_norm": 0.49856263399124146,
+      "learning_rate": 1.0502836224366389e-05,
+      "loss": 0.9383,
+      "num_input_tokens_seen": 132546560,
+      "step": 16180
+    },
+    {
+      "epoch": 2.0920015505879315,
+      "grad_norm": 1.046005368232727,
+      "learning_rate": 1.0475289453491038e-05,
+      "loss": 0.8762,
+      "num_input_tokens_seen": 132628480,
+      "step": 16190
+    },
+    {
+      "epoch": 2.0932937071973123,
+      "grad_norm": 0.6279125213623047,
+      "learning_rate": 1.0447769277307554e-05,
+      "loss": 0.7354,
+      "num_input_tokens_seen": 132710400,
+      "step": 16200
+    },
+    {
+      "epoch": 2.094585863806693,
+      "grad_norm": 0.666612982749939,
+      "learning_rate": 1.042027574620526e-05,
+      "loss": 0.6087,
+      "num_input_tokens_seen": 132792320,
+      "step": 16210
+    },
+    {
+      "epoch": 2.0958780204160745,
+      "grad_norm": 0.6725567579269409,
+      "learning_rate": 1.0392808910524735e-05,
+      "loss": 0.8064,
+      "num_input_tokens_seen": 132874240,
+      "step": 16220
+    },
+    {
+      "epoch": 2.0971701770254554,
+      "grad_norm": 0.7392178773880005,
+      "learning_rate": 1.0365368820557633e-05,
+      "loss": 0.6069,
+      "num_input_tokens_seen": 132956160,
+      "step": 16230
+    },
+    {
+      "epoch": 2.0984623336348367,
+      "grad_norm": 0.7892594337463379,
+      "learning_rate": 1.0337955526546678e-05,
+      "loss": 0.6704,
+      "num_input_tokens_seen": 133038080,
+      "step": 16240
+    },
+    {
+      "epoch": 2.0997544902442176,
+      "grad_norm": 0.6627844572067261,
+      "learning_rate": 1.0310569078685494e-05,
+      "loss": 0.764,
+      "num_input_tokens_seen": 133120000,
+      "step": 16250
+    },
+    {
+      "epoch": 2.1010466468535984,
+      "grad_norm": 0.22952121496200562,
+      "learning_rate": 1.0283209527118584e-05,
+      "loss": 0.7597,
+      "num_input_tokens_seen": 133201920,
+      "step": 16260
+    },
+    {
+      "epoch": 2.1023388034629797,
+      "grad_norm": 1.4370572566986084,
+      "learning_rate": 1.0255876921941165e-05,
+      "loss": 0.8361,
+      "num_input_tokens_seen": 133283840,
+      "step": 16270
+    },
+    {
+      "epoch": 2.1036309600723606,
+      "grad_norm": 1.0919463634490967,
+      "learning_rate": 1.0228571313199161e-05,
+      "loss": 0.7833,
+      "num_input_tokens_seen": 133365760,
+      "step": 16280
+    },
+    {
+      "epoch": 2.104923116681742,
+      "grad_norm": 0.7926573753356934,
+      "learning_rate": 1.0201292750889022e-05,
+      "loss": 0.6994,
+      "num_input_tokens_seen": 133447680,
+      "step": 16290
+    },
+    {
+      "epoch": 2.106215273291123,
+      "grad_norm": 0.636835515499115,
+      "learning_rate": 1.0174041284957703e-05,
+      "loss": 0.549,
+      "num_input_tokens_seen": 133529600,
+      "step": 16300
+    },
+    {
+      "epoch": 2.107507429900504,
+      "grad_norm": 0.7572881579399109,
+      "learning_rate": 1.0146816965302546e-05,
+      "loss": 0.966,
+      "num_input_tokens_seen": 133611520,
+      "step": 16310
+    },
+    {
+      "epoch": 2.108799586509885,
+      "grad_norm": 0.6216884255409241,
+      "learning_rate": 1.011961984177117e-05,
+      "loss": 0.8817,
+      "num_input_tokens_seen": 133693440,
+      "step": 16320
+    },
+    {
+      "epoch": 2.1100917431192663,
+      "grad_norm": 1.0052217245101929,
+      "learning_rate": 1.0092449964161416e-05,
+      "loss": 0.7589,
+      "num_input_tokens_seen": 133775360,
+      "step": 16330
+    },
+    {
+      "epoch": 2.111383899728647,
+      "grad_norm": 0.8983517289161682,
+      "learning_rate": 1.006530738222122e-05,
+      "loss": 0.7143,
+      "num_input_tokens_seen": 133857280,
+      "step": 16340
+    },
+    {
+      "epoch": 2.112676056338028,
+      "grad_norm": 0.5024335384368896,
+      "learning_rate": 1.0038192145648567e-05,
+      "loss": 0.9325,
+      "num_input_tokens_seen": 133939200,
+      "step": 16350
+    },
+    {
+      "epoch": 2.1139682129474093,
+      "grad_norm": 1.1501423120498657,
+      "learning_rate": 1.001110430409134e-05,
+      "loss": 0.8195,
+      "num_input_tokens_seen": 134021120,
+      "step": 16360
+    },
+    {
+      "epoch": 2.11526036955679,
+      "grad_norm": 0.6280581951141357,
+      "learning_rate": 9.98404390714729e-06,
+      "loss": 0.6386,
+      "num_input_tokens_seen": 134103040,
+      "step": 16370
+    },
+    {
+      "epoch": 2.1165525261661715,
+      "grad_norm": 0.6069371104240417,
+      "learning_rate": 9.95701100436389e-06,
+      "loss": 0.8842,
+      "num_input_tokens_seen": 134184960,
+      "step": 16380
+    },
+    {
+      "epoch": 2.1178446827755524,
+      "grad_norm": 0.8397451639175415,
+      "learning_rate": 9.930005645238302e-06,
+      "loss": 1.0033,
+      "num_input_tokens_seen": 134266880,
+      "step": 16390
+    },
+    {
+      "epoch": 2.1191368393849332,
+      "grad_norm": 0.9031828045845032,
+      "learning_rate": 9.903027879217237e-06,
+      "loss": 0.7303,
+      "num_input_tokens_seen": 134348800,
+      "step": 16400
+    },
+    {
+      "epoch": 2.1204289959943146,
+      "grad_norm": 0.8852018117904663,
+      "learning_rate": 9.876077755696868e-06,
+      "loss": 0.4911,
+      "num_input_tokens_seen": 134430720,
+      "step": 16410
+    },
+    {
+      "epoch": 2.1217211526036954,
+      "grad_norm": 0.6135576963424683,
+      "learning_rate": 9.849155324022799e-06,
+      "loss": 0.8788,
+      "num_input_tokens_seen": 134512640,
+      "step": 16420
+    },
+    {
+      "epoch": 2.1230133092130767,
+      "grad_norm": 0.9423490762710571,
+      "learning_rate": 9.82226063348988e-06,
+      "loss": 0.7269,
+      "num_input_tokens_seen": 134594560,
+      "step": 16430
+    },
+    {
+      "epoch": 2.1243054658224576,
+      "grad_norm": 0.6482061147689819,
+      "learning_rate": 9.795393733342203e-06,
+      "loss": 0.8314,
+      "num_input_tokens_seen": 134676480,
+      "step": 16440
+    },
+    {
+      "epoch": 2.125597622431839,
+      "grad_norm": 1.0169423818588257,
+      "learning_rate": 9.76855467277297e-06,
+      "loss": 0.7974,
+      "num_input_tokens_seen": 134758400,
+      "step": 16450
+    },
+    {
+      "epoch": 2.12688977904122,
+      "grad_norm": 0.5455058813095093,
+      "learning_rate": 9.741743500924388e-06,
+      "loss": 0.8748,
+      "num_input_tokens_seen": 134840320,
+      "step": 16460
+    },
+    {
+      "epoch": 2.1281819356506007,
+      "grad_norm": 0.3267827033996582,
+      "learning_rate": 9.71496026688763e-06,
+      "loss": 0.6756,
+      "num_input_tokens_seen": 134922240,
+      "step": 16470
+    },
+    {
+      "epoch": 2.129474092259982,
+      "grad_norm": 0.6928475499153137,
+      "learning_rate": 9.688205019702684e-06,
+      "loss": 1.01,
+      "num_input_tokens_seen": 135004160,
+      "step": 16480
+    },
+    {
+      "epoch": 2.130766248869363,
+      "grad_norm": 0.7541202306747437,
+      "learning_rate": 9.661477808358323e-06,
+      "loss": 0.8252,
+      "num_input_tokens_seen": 135086080,
+      "step": 16490
+    },
+    {
+      "epoch": 2.132058405478744,
+      "grad_norm": 1.1462032794952393,
+      "learning_rate": 9.634778681791962e-06,
+      "loss": 0.8617,
+      "num_input_tokens_seen": 135168000,
+      "step": 16500
+    },
+    {
+      "epoch": 2.133350562088125,
+      "grad_norm": 0.8049100041389465,
+      "learning_rate": 9.608107688889609e-06,
+      "loss": 1.2436,
+      "num_input_tokens_seen": 135249920,
+      "step": 16510
+    },
+    {
+      "epoch": 2.1346427186975063,
+      "grad_norm": 0.6773203015327454,
+      "learning_rate": 9.581464878485764e-06,
+      "loss": 0.7527,
+      "num_input_tokens_seen": 135331840,
+      "step": 16520
+    },
+    {
+      "epoch": 2.135934875306887,
+      "grad_norm": 0.743137776851654,
+      "learning_rate": 9.554850299363294e-06,
+      "loss": 0.7333,
+      "num_input_tokens_seen": 135413760,
+      "step": 16530
+    },
+    {
+      "epoch": 2.137227031916268,
+      "grad_norm": 0.5760340094566345,
+      "learning_rate": 9.52826400025342e-06,
+      "loss": 1.0707,
+      "num_input_tokens_seen": 135495680,
+      "step": 16540
+    },
+    {
+      "epoch": 2.1385191885256494,
+      "grad_norm": 0.5763425230979919,
+      "learning_rate": 9.501706029835544e-06,
+      "loss": 0.7467,
+      "num_input_tokens_seen": 135577600,
+      "step": 16550
+    },
+    {
+      "epoch": 2.1398113451350302,
+      "grad_norm": 1.1024504899978638,
+      "learning_rate": 9.47517643673721e-06,
+      "loss": 0.6344,
+      "num_input_tokens_seen": 135659520,
+      "step": 16560
+    },
+    {
+      "epoch": 2.1411035017444116,
+      "grad_norm": 0.5545995235443115,
+      "learning_rate": 9.448675269534015e-06,
+      "loss": 0.844,
+      "num_input_tokens_seen": 135741440,
+      "step": 16570
+    },
+    {
+      "epoch": 2.1423956583537924,
+      "grad_norm": 0.710102915763855,
+      "learning_rate": 9.422202576749492e-06,
+      "loss": 0.9002,
+      "num_input_tokens_seen": 135823360,
+      "step": 16580
+    },
+    {
+      "epoch": 2.1436878149631737,
+      "grad_norm": 0.7064651846885681,
+      "learning_rate": 9.395758406855053e-06,
+      "loss": 0.9066,
+      "num_input_tokens_seen": 135905280,
+      "step": 16590
+    },
+    {
+      "epoch": 2.1449799715725546,
+      "grad_norm": 0.34538978338241577,
+      "learning_rate": 9.369342808269862e-06,
+      "loss": 0.8606,
+      "num_input_tokens_seen": 135987200,
+      "step": 16600
+    },
+    {
+      "epoch": 2.1462721281819355,
+      "grad_norm": 0.7004348635673523,
+      "learning_rate": 9.342955829360806e-06,
+      "loss": 0.8238,
+      "num_input_tokens_seen": 136069120,
+      "step": 16610
+    },
+    {
+      "epoch": 2.147564284791317,
+      "grad_norm": 0.664341390132904,
+      "learning_rate": 9.31659751844232e-06,
+      "loss": 0.571,
+      "num_input_tokens_seen": 136151040,
+      "step": 16620
+    },
+    {
+      "epoch": 2.1488564414006976,
+      "grad_norm": 0.9971612691879272,
+      "learning_rate": 9.290267923776397e-06,
+      "loss": 0.7521,
+      "num_input_tokens_seen": 136232960,
+      "step": 16630
+    },
+    {
+      "epoch": 2.150148598010079,
+      "grad_norm": 0.39424338936805725,
+      "learning_rate": 9.263967093572412e-06,
+      "loss": 0.3768,
+      "num_input_tokens_seen": 136314880,
+      "step": 16640
+    },
+    {
+      "epoch": 2.15144075461946,
+      "grad_norm": 0.7910518646240234,
+      "learning_rate": 9.237695075987106e-06,
+      "loss": 0.727,
+      "num_input_tokens_seen": 136396800,
+      "step": 16650
+    },
+    {
+      "epoch": 2.152732911228841,
+      "grad_norm": 0.6754958629608154,
+      "learning_rate": 9.211451919124429e-06,
+      "loss": 0.8798,
+      "num_input_tokens_seen": 136478720,
+      "step": 16660
+    },
+    {
+      "epoch": 2.154025067838222,
+      "grad_norm": 0.6639232039451599,
+      "learning_rate": 9.185237671035512e-06,
+      "loss": 1.1385,
+      "num_input_tokens_seen": 136560640,
+      "step": 16670
+    },
+    {
+      "epoch": 2.155317224447603,
+      "grad_norm": 0.8237486481666565,
+      "learning_rate": 9.15905237971856e-06,
+      "loss": 0.8854,
+      "num_input_tokens_seen": 136642560,
+      "step": 16680
+    },
+    {
+      "epoch": 2.156609381056984,
+      "grad_norm": 0.863280713558197,
+      "learning_rate": 9.132896093118726e-06,
+      "loss": 0.6467,
+      "num_input_tokens_seen": 136724480,
+      "step": 16690
+    },
+    {
+      "epoch": 2.157901537666365,
+      "grad_norm": 0.24760442972183228,
+      "learning_rate": 9.10676885912809e-06,
+      "loss": 0.7871,
+      "num_input_tokens_seen": 136806400,
+      "step": 16700
+    },
+    {
+      "epoch": 2.1591936942757464,
+      "grad_norm": 0.5246381759643555,
+      "learning_rate": 9.080670725585511e-06,
+      "loss": 0.3123,
+      "num_input_tokens_seen": 136888320,
+      "step": 16710
+    },
+    {
+      "epoch": 2.1604858508851272,
+      "grad_norm": 0.6992424726486206,
+      "learning_rate": 9.054601740276586e-06,
+      "loss": 0.7407,
+      "num_input_tokens_seen": 136970240,
+      "step": 16720
+    },
+    {
+      "epoch": 2.161778007494508,
+      "grad_norm": 0.7333977222442627,
+      "learning_rate": 9.028561950933517e-06,
+      "loss": 0.6363,
+      "num_input_tokens_seen": 137052160,
+      "step": 16730
+    },
+    {
+      "epoch": 2.1630701641038894,
+      "grad_norm": 0.5862302184104919,
+      "learning_rate": 9.002551405235082e-06,
+      "loss": 0.7121,
+      "num_input_tokens_seen": 137134080,
+      "step": 16740
+    },
+    {
+      "epoch": 2.1643623207132703,
+      "grad_norm": 0.36307257413864136,
+      "learning_rate": 8.976570150806486e-06,
+      "loss": 0.9597,
+      "num_input_tokens_seen": 137216000,
+      "step": 16750
+    },
+    {
+      "epoch": 2.1656544773226516,
+      "grad_norm": 0.8547699451446533,
+      "learning_rate": 8.950618235219302e-06,
+      "loss": 0.7746,
+      "num_input_tokens_seen": 137297920,
+      "step": 16760
+    },
+    {
+      "epoch": 2.1669466339320325,
+      "grad_norm": 0.5659860372543335,
+      "learning_rate": 8.924695705991407e-06,
+      "loss": 0.4788,
+      "num_input_tokens_seen": 137379840,
+      "step": 16770
+    },
+    {
+      "epoch": 2.1682387905414138,
+      "grad_norm": 0.7544788122177124,
+      "learning_rate": 8.898802610586843e-06,
+      "loss": 0.6243,
+      "num_input_tokens_seen": 137461760,
+      "step": 16780
+    },
+    {
+      "epoch": 2.1695309471507946,
+      "grad_norm": 0.728805661201477,
+      "learning_rate": 8.872938996415791e-06,
+      "loss": 0.9596,
+      "num_input_tokens_seen": 137543680,
+      "step": 16790
+    },
+    {
+      "epoch": 2.170823103760176,
+      "grad_norm": 0.7234178185462952,
+      "learning_rate": 8.847104910834414e-06,
+      "loss": 0.83,
+      "num_input_tokens_seen": 137625600,
+      "step": 16800
+    },
+    {
+      "epoch": 2.172115260369557,
+      "grad_norm": 0.29097720980644226,
+      "learning_rate": 8.821300401144836e-06,
+      "loss": 0.8546,
+      "num_input_tokens_seen": 137707520,
+      "step": 16810
+    },
+    {
+      "epoch": 2.1734074169789377,
+      "grad_norm": 0.6291516423225403,
+      "learning_rate": 8.795525514595032e-06,
+      "loss": 0.7418,
+      "num_input_tokens_seen": 137789440,
+      "step": 16820
+    },
+    {
+      "epoch": 2.174699573588319,
+      "grad_norm": 0.7936385273933411,
+      "learning_rate": 8.769780298378705e-06,
+      "loss": 0.9961,
+      "num_input_tokens_seen": 137871360,
+      "step": 16830
+    },
+    {
+      "epoch": 2.1759917301977,
+      "grad_norm": 0.8866328001022339,
+      "learning_rate": 8.74406479963527e-06,
+      "loss": 0.5871,
+      "num_input_tokens_seen": 137953280,
+      "step": 16840
+    },
+    {
+      "epoch": 2.177283886807081,
+      "grad_norm": 1.1491591930389404,
+      "learning_rate": 8.718379065449694e-06,
+      "loss": 1.1771,
+      "num_input_tokens_seen": 138035200,
+      "step": 16850
+    },
+    {
+      "epoch": 2.178576043416462,
+      "grad_norm": 0.9200372099876404,
+      "learning_rate": 8.69272314285248e-06,
+      "loss": 0.5562,
+      "num_input_tokens_seen": 138117120,
+      "step": 16860
+    },
+    {
+      "epoch": 2.179868200025843,
+      "grad_norm": 0.7428059577941895,
+      "learning_rate": 8.667097078819511e-06,
+      "loss": 1.0242,
+      "num_input_tokens_seen": 138199040,
+      "step": 16870
+    },
+    {
+      "epoch": 2.1811603566352242,
+      "grad_norm": 0.7725750803947449,
+      "learning_rate": 8.641500920272022e-06,
+      "loss": 0.8353,
+      "num_input_tokens_seen": 138280960,
+      "step": 16880
+    },
+    {
+      "epoch": 2.182452513244605,
+      "grad_norm": 0.8504687547683716,
+      "learning_rate": 8.6159347140765e-06,
+      "loss": 0.7933,
+      "num_input_tokens_seen": 138362880,
+      "step": 16890
+    },
+    {
+      "epoch": 2.1837446698539864,
+      "grad_norm": 0.8972610831260681,
+      "learning_rate": 8.59039850704455e-06,
+      "loss": 0.9106,
+      "num_input_tokens_seen": 138444800,
+      "step": 16900
+    },
+    {
+      "epoch": 2.1850368264633673,
+      "grad_norm": 0.9138893485069275,
+      "learning_rate": 8.564892345932899e-06,
+      "loss": 0.9258,
+      "num_input_tokens_seen": 138526720,
+      "step": 16910
+    },
+    {
+      "epoch": 2.1863289830727486,
+      "grad_norm": 0.8643404841423035,
+      "learning_rate": 8.539416277443218e-06,
+      "loss": 0.6456,
+      "num_input_tokens_seen": 138608640,
+      "step": 16920
+    },
+    {
+      "epoch": 2.1876211396821295,
+      "grad_norm": 0.5902085304260254,
+      "learning_rate": 8.513970348222095e-06,
+      "loss": 0.7899,
+      "num_input_tokens_seen": 138690560,
+      "step": 16930
+    },
+    {
+      "epoch": 2.1889132962915103,
+      "grad_norm": 0.7431287169456482,
+      "learning_rate": 8.488554604860947e-06,
+      "loss": 1.0066,
+      "num_input_tokens_seen": 138772480,
+      "step": 16940
+    },
+    {
+      "epoch": 2.1902054529008916,
+      "grad_norm": 0.330143541097641,
+      "learning_rate": 8.463169093895887e-06,
+      "loss": 0.5788,
+      "num_input_tokens_seen": 138854400,
+      "step": 16950
+    },
+    {
+      "epoch": 2.1914976095102725,
+      "grad_norm": 0.7542316913604736,
+      "learning_rate": 8.437813861807712e-06,
+      "loss": 0.6761,
+      "num_input_tokens_seen": 138936320,
+      "step": 16960
+    },
+    {
+      "epoch": 2.192789766119654,
+      "grad_norm": 0.6555983424186707,
+      "learning_rate": 8.412488955021744e-06,
+      "loss": 1.3443,
+      "num_input_tokens_seen": 139018240,
+      "step": 16970
+    },
+    {
+      "epoch": 2.1940819227290347,
+      "grad_norm": 0.9010647535324097,
+      "learning_rate": 8.38719441990781e-06,
+      "loss": 0.8272,
+      "num_input_tokens_seen": 139100160,
+      "step": 16980
+    },
+    {
+      "epoch": 2.195374079338416,
+      "grad_norm": 0.4530339539051056,
+      "learning_rate": 8.361930302780091e-06,
+      "loss": 0.6814,
+      "num_input_tokens_seen": 139182080,
+      "step": 16990
+    },
+    {
+      "epoch": 2.196666235947797,
+      "grad_norm": 0.5172061324119568,
+      "learning_rate": 8.336696649897116e-06,
+      "loss": 0.5844,
+      "num_input_tokens_seen": 139264000,
+      "step": 17000
+    },
+    {
+      "epoch": 2.1979583925571777,
+      "grad_norm": 1.0065109729766846,
+      "learning_rate": 8.311493507461593e-06,
+      "loss": 0.6501,
+      "num_input_tokens_seen": 139345920,
+      "step": 17010
+    },
+    {
+      "epoch": 2.199250549166559,
+      "grad_norm": 0.3245972990989685,
+      "learning_rate": 8.286320921620394e-06,
+      "loss": 0.6703,
+      "num_input_tokens_seen": 139427840,
+      "step": 17020
+    },
+    {
+      "epoch": 2.20054270577594,
+      "grad_norm": 0.8792958855628967,
+      "learning_rate": 8.261178938464422e-06,
+      "loss": 0.83,
+      "num_input_tokens_seen": 139509760,
+      "step": 17030
+    },
+    {
+      "epoch": 2.2018348623853212,
+      "grad_norm": 0.7680603265762329,
+      "learning_rate": 8.236067604028563e-06,
+      "loss": 0.6316,
+      "num_input_tokens_seen": 139591680,
+      "step": 17040
+    },
+    {
+      "epoch": 2.203127018994702,
+      "grad_norm": 0.8535128831863403,
+      "learning_rate": 8.210986964291587e-06,
+      "loss": 0.7002,
+      "num_input_tokens_seen": 139673600,
+      "step": 17050
+    },
+    {
+      "epoch": 2.2044191756040834,
+      "grad_norm": 0.7545835375785828,
+      "learning_rate": 8.185937065176033e-06,
+      "loss": 0.9392,
+      "num_input_tokens_seen": 139755520,
+      "step": 17060
+    },
+    {
+      "epoch": 2.2057113322134643,
+      "grad_norm": 0.5624322891235352,
+      "learning_rate": 8.160917952548197e-06,
+      "loss": 0.828,
+      "num_input_tokens_seen": 139837440,
+      "step": 17070
+    },
+    {
+      "epoch": 2.207003488822845,
+      "grad_norm": 1.1098763942718506,
+      "learning_rate": 8.13592967221796e-06,
+      "loss": 0.9871,
+      "num_input_tokens_seen": 139919360,
+      "step": 17080
+    },
+    {
+      "epoch": 2.2082956454322265,
+      "grad_norm": 0.5870392322540283,
+      "learning_rate": 8.110972269938793e-06,
+      "loss": 1.0047,
+      "num_input_tokens_seen": 140001280,
+      "step": 17090
+    },
+    {
+      "epoch": 2.2095878020416073,
+      "grad_norm": 0.8012586236000061,
+      "learning_rate": 8.08604579140759e-06,
+      "loss": 1.1041,
+      "num_input_tokens_seen": 140083200,
+      "step": 17100
+    },
+    {
+      "epoch": 2.2108799586509886,
+      "grad_norm": 0.7019762992858887,
+      "learning_rate": 8.06115028226466e-06,
+      "loss": 1.116,
+      "num_input_tokens_seen": 140165120,
+      "step": 17110
+    },
+    {
+      "epoch": 2.2121721152603695,
+      "grad_norm": 0.646838366985321,
+      "learning_rate": 8.036285788093578e-06,
+      "loss": 0.82,
+      "num_input_tokens_seen": 140247040,
+      "step": 17120
+    },
+    {
+      "epoch": 2.213464271869751,
+      "grad_norm": 0.8307018876075745,
+      "learning_rate": 8.011452354421136e-06,
+      "loss": 0.6641,
+      "num_input_tokens_seen": 140328960,
+      "step": 17130
+    },
+    {
+      "epoch": 2.2147564284791317,
+      "grad_norm": 0.39130905270576477,
+      "learning_rate": 7.986650026717277e-06,
+      "loss": 0.5812,
+      "num_input_tokens_seen": 140410880,
+      "step": 17140
+    },
+    {
+      "epoch": 2.2160485850885125,
+      "grad_norm": 0.7408241033554077,
+      "learning_rate": 7.961878850394952e-06,
+      "loss": 0.8512,
+      "num_input_tokens_seen": 140492800,
+      "step": 17150
+    },
+    {
+      "epoch": 2.217340741697894,
+      "grad_norm": 0.6536303758621216,
+      "learning_rate": 7.937138870810115e-06,
+      "loss": 0.7622,
+      "num_input_tokens_seen": 140574720,
+      "step": 17160
+    },
+    {
+      "epoch": 2.2186328983072747,
+      "grad_norm": 0.9935400485992432,
+      "learning_rate": 7.912430133261562e-06,
+      "loss": 0.7604,
+      "num_input_tokens_seen": 140656640,
+      "step": 17170
+    },
+    {
+      "epoch": 2.219925054916656,
+      "grad_norm": 0.6916456818580627,
+      "learning_rate": 7.887752682990903e-06,
+      "loss": 0.7567,
+      "num_input_tokens_seen": 140738560,
+      "step": 17180
+    },
+    {
+      "epoch": 2.221217211526037,
+      "grad_norm": 0.7032532095909119,
+      "learning_rate": 7.863106565182474e-06,
+      "loss": 1.0241,
+      "num_input_tokens_seen": 140820480,
+      "step": 17190
+    },
+    {
+      "epoch": 2.222509368135418,
+      "grad_norm": 0.6725844740867615,
+      "learning_rate": 7.838491824963207e-06,
+      "loss": 0.8592,
+      "num_input_tokens_seen": 140902400,
+      "step": 17200
+    },
+    {
+      "epoch": 2.223801524744799,
+      "grad_norm": 0.851714551448822,
+      "learning_rate": 7.81390850740262e-06,
+      "loss": 0.9155,
+      "num_input_tokens_seen": 140984320,
+      "step": 17210
+    },
+    {
+      "epoch": 2.22509368135418,
+      "grad_norm": 0.7544275522232056,
+      "learning_rate": 7.78935665751266e-06,
+      "loss": 1.1792,
+      "num_input_tokens_seen": 141066240,
+      "step": 17220
+    },
+    {
+      "epoch": 2.2263858379635613,
+      "grad_norm": 1.2108078002929688,
+      "learning_rate": 7.764836320247686e-06,
+      "loss": 0.7382,
+      "num_input_tokens_seen": 141148160,
+      "step": 17230
+    },
+    {
+      "epoch": 2.227677994572942,
+      "grad_norm": 0.5292445421218872,
+      "learning_rate": 7.740347540504336e-06,
+      "loss": 0.7866,
+      "num_input_tokens_seen": 141230080,
+      "step": 17240
+    },
+    {
+      "epoch": 2.2289701511823234,
+      "grad_norm": 0.5671383738517761,
+      "learning_rate": 7.715890363121484e-06,
+      "loss": 0.5538,
+      "num_input_tokens_seen": 141312000,
+      "step": 17250
+    },
+    {
+      "epoch": 2.2302623077917043,
+      "grad_norm": 1.1066720485687256,
+      "learning_rate": 7.691464832880135e-06,
+      "loss": 0.7333,
+      "num_input_tokens_seen": 141393920,
+      "step": 17260
+    },
+    {
+      "epoch": 2.2315544644010856,
+      "grad_norm": 0.3376465439796448,
+      "learning_rate": 7.667070994503334e-06,
+      "loss": 0.7558,
+      "num_input_tokens_seen": 141475840,
+      "step": 17270
+    },
+    {
+      "epoch": 2.2328466210104665,
+      "grad_norm": 0.6462758779525757,
+      "learning_rate": 7.642708892656125e-06,
+      "loss": 0.4585,
+      "num_input_tokens_seen": 141557760,
+      "step": 17280
+    },
+    {
+      "epoch": 2.2341387776198474,
+      "grad_norm": 0.7883977890014648,
+      "learning_rate": 7.618378571945417e-06,
+      "loss": 0.602,
+      "num_input_tokens_seen": 141639680,
+      "step": 17290
+    },
+    {
+      "epoch": 2.2354309342292287,
+      "grad_norm": 0.7147501707077026,
+      "learning_rate": 7.5940800769199345e-06,
+      "loss": 0.7789,
+      "num_input_tokens_seen": 141721600,
+      "step": 17300
+    },
+    {
+      "epoch": 2.2367230908386095,
+      "grad_norm": 0.7497975826263428,
+      "learning_rate": 7.569813452070146e-06,
+      "loss": 0.5047,
+      "num_input_tokens_seen": 141803520,
+      "step": 17310
+    },
+    {
+      "epoch": 2.238015247447991,
+      "grad_norm": 0.6466274261474609,
+      "learning_rate": 7.545578741828136e-06,
+      "loss": 0.6986,
+      "num_input_tokens_seen": 141885440,
+      "step": 17320
+    },
+    {
+      "epoch": 2.2393074040573717,
+      "grad_norm": 0.7558254599571228,
+      "learning_rate": 7.521375990567589e-06,
+      "loss": 0.8236,
+      "num_input_tokens_seen": 141967360,
+      "step": 17330
+    },
+    {
+      "epoch": 2.2405995606667526,
+      "grad_norm": 0.7305789589881897,
+      "learning_rate": 7.497205242603636e-06,
+      "loss": 0.469,
+      "num_input_tokens_seen": 142049280,
+      "step": 17340
+    },
+    {
+      "epoch": 2.241891717276134,
+      "grad_norm": 0.29660215973854065,
+      "learning_rate": 7.4730665421928445e-06,
+      "loss": 0.7266,
+      "num_input_tokens_seen": 142131200,
+      "step": 17350
+    },
+    {
+      "epoch": 2.2431838738855148,
+      "grad_norm": 0.26684778928756714,
+      "learning_rate": 7.4489599335330704e-06,
+      "loss": 1.0225,
+      "num_input_tokens_seen": 142213120,
+      "step": 17360
+    },
+    {
+      "epoch": 2.244476030494896,
+      "grad_norm": 0.6525532007217407,
+      "learning_rate": 7.424885460763442e-06,
+      "loss": 0.5151,
+      "num_input_tokens_seen": 142295040,
+      "step": 17370
+    },
+    {
+      "epoch": 2.245768187104277,
+      "grad_norm": 0.7799077033996582,
+      "learning_rate": 7.4008431679642165e-06,
+      "loss": 0.5883,
+      "num_input_tokens_seen": 142376960,
+      "step": 17380
+    },
+    {
+      "epoch": 2.2470603437136583,
+      "grad_norm": 0.9852105975151062,
+      "learning_rate": 7.3768330991567495e-06,
+      "loss": 0.6615,
+      "num_input_tokens_seen": 142458880,
+      "step": 17390
+    },
+    {
+      "epoch": 2.248352500323039,
+      "grad_norm": 0.7310599684715271,
+      "learning_rate": 7.3528552983033985e-06,
+      "loss": 0.7732,
+      "num_input_tokens_seen": 142540800,
+      "step": 17400
+    },
+    {
+      "epoch": 2.2496446569324204,
+      "grad_norm": 0.7955625653266907,
+      "learning_rate": 7.328909809307413e-06,
+      "loss": 0.954,
+      "num_input_tokens_seen": 142622720,
+      "step": 17410
+    },
+    {
+      "epoch": 2.2509368135418013,
+      "grad_norm": 0.8421788215637207,
+      "learning_rate": 7.304996676012913e-06,
+      "loss": 0.5088,
+      "num_input_tokens_seen": 142704640,
+      "step": 17420
+    },
+    {
+      "epoch": 2.252228970151182,
+      "grad_norm": 0.772693932056427,
+      "learning_rate": 7.281115942204739e-06,
+      "loss": 0.812,
+      "num_input_tokens_seen": 142786560,
+      "step": 17430
+    },
+    {
+      "epoch": 2.2535211267605635,
+      "grad_norm": 0.9953277707099915,
+      "learning_rate": 7.257267651608446e-06,
+      "loss": 0.9194,
+      "num_input_tokens_seen": 142868480,
+      "step": 17440
+    },
+    {
+      "epoch": 2.2548132833699444,
+      "grad_norm": 0.6572209000587463,
+      "learning_rate": 7.233451847890149e-06,
+      "loss": 0.6278,
+      "num_input_tokens_seen": 142950400,
+      "step": 17450
+    },
+    {
+      "epoch": 2.2561054399793257,
+      "grad_norm": 0.3987307846546173,
+      "learning_rate": 7.209668574656514e-06,
+      "loss": 0.7189,
+      "num_input_tokens_seen": 143032320,
+      "step": 17460
+    },
+    {
+      "epoch": 2.2573975965887065,
+      "grad_norm": 0.6912767291069031,
+      "learning_rate": 7.185917875454615e-06,
+      "loss": 0.8608,
+      "num_input_tokens_seen": 143114240,
+      "step": 17470
+    },
+    {
+      "epoch": 2.2586897531980874,
+      "grad_norm": 0.6624774932861328,
+      "learning_rate": 7.162199793771904e-06,
+      "loss": 0.8747,
+      "num_input_tokens_seen": 143196160,
+      "step": 17480
+    },
+    {
+      "epoch": 2.2599819098074687,
+      "grad_norm": 0.3295046091079712,
+      "learning_rate": 7.138514373036098e-06,
+      "loss": 0.8996,
+      "num_input_tokens_seen": 143278080,
+      "step": 17490
+    },
+    {
+      "epoch": 2.2612740664168496,
+      "grad_norm": 0.7381219267845154,
+      "learning_rate": 7.11486165661511e-06,
+      "loss": 0.6536,
+      "num_input_tokens_seen": 143360000,
+      "step": 17500
+    },
+    {
+      "epoch": 2.262566223026231,
+      "grad_norm": 0.9237964749336243,
+      "learning_rate": 7.091241687816988e-06,
+      "loss": 0.9105,
+      "num_input_tokens_seen": 143441920,
+      "step": 17510
+    },
+    {
+      "epoch": 2.2638583796356118,
+      "grad_norm": 0.6693175435066223,
+      "learning_rate": 7.0676545098897956e-06,
+      "loss": 0.7303,
+      "num_input_tokens_seen": 143523840,
+      "step": 17520
+    },
+    {
+      "epoch": 2.265150536244993,
+      "grad_norm": 0.803263247013092,
+      "learning_rate": 7.044100166021583e-06,
+      "loss": 0.6937,
+      "num_input_tokens_seen": 143605760,
+      "step": 17530
+    },
+    {
+      "epoch": 2.266442692854374,
+      "grad_norm": 1.2088638544082642,
+      "learning_rate": 7.020578699340255e-06,
+      "loss": 0.7094,
+      "num_input_tokens_seen": 143687680,
+      "step": 17540
+    },
+    {
+      "epoch": 2.2677348494637553,
+      "grad_norm": 0.6141706109046936,
+      "learning_rate": 6.997090152913535e-06,
+      "loss": 1.0294,
+      "num_input_tokens_seen": 143769600,
+      "step": 17550
+    },
+    {
+      "epoch": 2.269027006073136,
+      "grad_norm": 0.8863146305084229,
+      "learning_rate": 6.97363456974888e-06,
+      "loss": 0.5169,
+      "num_input_tokens_seen": 143851520,
+      "step": 17560
+    },
+    {
+      "epoch": 2.270319162682517,
+      "grad_norm": 0.5809889435768127,
+      "learning_rate": 6.950211992793354e-06,
+      "loss": 0.8054,
+      "num_input_tokens_seen": 143933440,
+      "step": 17570
+    },
+    {
+      "epoch": 2.2716113192918983,
+      "grad_norm": 0.6235867142677307,
+      "learning_rate": 6.92682246493363e-06,
+      "loss": 1.089,
+      "num_input_tokens_seen": 144015360,
+      "step": 17580
+    },
+    {
+      "epoch": 2.272903475901279,
+      "grad_norm": 0.7494005560874939,
+      "learning_rate": 6.903466028995828e-06,
+      "loss": 0.5911,
+      "num_input_tokens_seen": 144097280,
+      "step": 17590
+    },
+    {
+      "epoch": 2.2741956325106605,
+      "grad_norm": 0.3730751574039459,
+      "learning_rate": 6.880142727745517e-06,
+      "loss": 0.4274,
+      "num_input_tokens_seen": 144179200,
+      "step": 17600
+    },
+    {
+      "epoch": 2.2754877891200413,
+      "grad_norm": 0.6000313758850098,
+      "learning_rate": 6.856852603887556e-06,
+      "loss": 0.5875,
+      "num_input_tokens_seen": 144261120,
+      "step": 17610
+    },
+    {
+      "epoch": 2.276779945729422,
+      "grad_norm": 0.8631287217140198,
+      "learning_rate": 6.8335957000660925e-06,
+      "loss": 1.0119,
+      "num_input_tokens_seen": 144343040,
+      "step": 17620
+    },
+    {
+      "epoch": 2.2780721023388035,
+      "grad_norm": 0.606353223323822,
+      "learning_rate": 6.810372058864429e-06,
+      "loss": 0.8559,
+      "num_input_tokens_seen": 144424960,
+      "step": 17630
+    },
+    {
+      "epoch": 2.2793642589481844,
+      "grad_norm": 0.7653380036354065,
+      "learning_rate": 6.787181722804959e-06,
+      "loss": 0.9657,
+      "num_input_tokens_seen": 144506880,
+      "step": 17640
+    },
+    {
+      "epoch": 2.2806564155575657,
+      "grad_norm": 0.706365168094635,
+      "learning_rate": 6.764024734349117e-06,
+      "loss": 0.9287,
+      "num_input_tokens_seen": 144588800,
+      "step": 17650
+    },
+    {
+      "epoch": 2.2819485721669466,
+      "grad_norm": 1.115876317024231,
+      "learning_rate": 6.740901135897257e-06,
+      "loss": 0.8438,
+      "num_input_tokens_seen": 144670720,
+      "step": 17660
+    },
+    {
+      "epoch": 2.2832407287763274,
+      "grad_norm": 0.9500836133956909,
+      "learning_rate": 6.717810969788596e-06,
+      "loss": 0.8497,
+      "num_input_tokens_seen": 144752640,
+      "step": 17670
+    },
+    {
+      "epoch": 2.2845328853857088,
+      "grad_norm": 0.8140472769737244,
+      "learning_rate": 6.694754278301154e-06,
+      "loss": 0.8295,
+      "num_input_tokens_seen": 144834560,
+      "step": 17680
+    },
+    {
+      "epoch": 2.2858250419950896,
+      "grad_norm": 0.36906084418296814,
+      "learning_rate": 6.671731103651641e-06,
+      "loss": 0.6225,
+      "num_input_tokens_seen": 144916480,
+      "step": 17690
+    },
+    {
+      "epoch": 2.287117198604471,
+      "grad_norm": 0.9518002271652222,
+      "learning_rate": 6.648741487995416e-06,
+      "loss": 0.7609,
+      "num_input_tokens_seen": 144998400,
+      "step": 17700
+    },
+    {
+      "epoch": 2.288409355213852,
+      "grad_norm": 0.6692580580711365,
+      "learning_rate": 6.625785473426369e-06,
+      "loss": 0.9989,
+      "num_input_tokens_seen": 145080320,
+      "step": 17710
+    },
+    {
+      "epoch": 2.289701511823233,
+      "grad_norm": 1.0394073724746704,
+      "learning_rate": 6.602863101976886e-06,
+      "loss": 0.9415,
+      "num_input_tokens_seen": 145162240,
+      "step": 17720
+    },
+    {
+      "epoch": 2.290993668432614,
+      "grad_norm": 0.7470195889472961,
+      "learning_rate": 6.57997441561774e-06,
+      "loss": 0.8093,
+      "num_input_tokens_seen": 145244160,
+      "step": 17730
+    },
+    {
+      "epoch": 2.2922858250419953,
+      "grad_norm": 0.5699132084846497,
+      "learning_rate": 6.557119456258043e-06,
+      "loss": 0.6653,
+      "num_input_tokens_seen": 145326080,
+      "step": 17740
+    },
+    {
+      "epoch": 2.293577981651376,
+      "grad_norm": 0.5270111560821533,
+      "learning_rate": 6.534298265745128e-06,
+      "loss": 0.4557,
+      "num_input_tokens_seen": 145408000,
+      "step": 17750
+    },
+    {
+      "epoch": 2.294870138260757,
+      "grad_norm": 1.350411057472229,
+      "learning_rate": 6.511510885864516e-06,
+      "loss": 0.4692,
+      "num_input_tokens_seen": 145489920,
+      "step": 17760
+    },
+    {
+      "epoch": 2.2961622948701383,
+      "grad_norm": 0.2834679186344147,
+      "learning_rate": 6.4887573583398255e-06,
+      "loss": 0.6354,
+      "num_input_tokens_seen": 145571840,
+      "step": 17770
+    },
+    {
+      "epoch": 2.297454451479519,
+      "grad_norm": 0.27980419993400574,
+      "learning_rate": 6.466037724832666e-06,
+      "loss": 0.4408,
+      "num_input_tokens_seen": 145653760,
+      "step": 17780
+    },
+    {
+      "epoch": 2.2987466080889005,
+      "grad_norm": 0.7254844307899475,
+      "learning_rate": 6.44335202694262e-06,
+      "loss": 0.9355,
+      "num_input_tokens_seen": 145735680,
+      "step": 17790
+    },
+    {
+      "epoch": 2.3000387646982814,
+      "grad_norm": 0.7839716076850891,
+      "learning_rate": 6.420700306207103e-06,
+      "loss": 1.1966,
+      "num_input_tokens_seen": 145817600,
+      "step": 17800
+    },
+    {
+      "epoch": 2.3013309213076623,
+      "grad_norm": 0.8338444232940674,
+      "learning_rate": 6.3980826041013464e-06,
+      "loss": 0.3919,
+      "num_input_tokens_seen": 145899520,
+      "step": 17810
+    },
+    {
+      "epoch": 2.3026230779170436,
+      "grad_norm": 0.608752429485321,
+      "learning_rate": 6.375498962038265e-06,
+      "loss": 0.9223,
+      "num_input_tokens_seen": 145981440,
+      "step": 17820
+    },
+    {
+      "epoch": 2.3039152345264244,
+      "grad_norm": 0.5790471434593201,
+      "learning_rate": 6.35294942136844e-06,
+      "loss": 0.5995,
+      "num_input_tokens_seen": 146063360,
+      "step": 17830
+    },
+    {
+      "epoch": 2.3052073911358058,
+      "grad_norm": 0.9189697504043579,
+      "learning_rate": 6.3304340233799805e-06,
+      "loss": 0.6963,
+      "num_input_tokens_seen": 146145280,
+      "step": 17840
+    },
+    {
+      "epoch": 2.3064995477451866,
+      "grad_norm": 0.6007638573646545,
+      "learning_rate": 6.307952809298517e-06,
+      "loss": 0.6846,
+      "num_input_tokens_seen": 146227200,
+      "step": 17850
+    },
+    {
+      "epoch": 2.307791704354568,
+      "grad_norm": 0.8957340717315674,
+      "learning_rate": 6.28550582028706e-06,
+      "loss": 0.542,
+      "num_input_tokens_seen": 146309120,
+      "step": 17860
+    },
+    {
+      "epoch": 2.309083860963949,
+      "grad_norm": 0.8787885904312134,
+      "learning_rate": 6.263093097445957e-06,
+      "loss": 0.9602,
+      "num_input_tokens_seen": 146391040,
+      "step": 17870
+    },
+    {
+      "epoch": 2.31037601757333,
+      "grad_norm": 0.655099630355835,
+      "learning_rate": 6.240714681812837e-06,
+      "loss": 0.8196,
+      "num_input_tokens_seen": 146472960,
+      "step": 17880
+    },
+    {
+      "epoch": 2.311668174182711,
+      "grad_norm": 1.1180391311645508,
+      "learning_rate": 6.218370614362484e-06,
+      "loss": 0.6883,
+      "num_input_tokens_seen": 146554880,
+      "step": 17890
+    },
+    {
+      "epoch": 2.312960330792092,
+      "grad_norm": 0.5654224753379822,
+      "learning_rate": 6.196060936006817e-06,
+      "loss": 0.5604,
+      "num_input_tokens_seen": 146636800,
+      "step": 17900
+    },
+    {
+      "epoch": 2.314252487401473,
+      "grad_norm": 0.617138147354126,
+      "learning_rate": 6.173785687594761e-06,
+      "loss": 1.045,
+      "num_input_tokens_seen": 146718720,
+      "step": 17910
+    },
+    {
+      "epoch": 2.315544644010854,
+      "grad_norm": 0.2932003140449524,
+      "learning_rate": 6.1515449099122185e-06,
+      "loss": 0.8279,
+      "num_input_tokens_seen": 146800640,
+      "step": 17920
+    },
+    {
+      "epoch": 2.3168368006202353,
+      "grad_norm": 0.7684862017631531,
+      "learning_rate": 6.129338643681984e-06,
+      "loss": 1.0301,
+      "num_input_tokens_seen": 146882560,
+      "step": 17930
+    },
+    {
+      "epoch": 2.318128957229616,
+      "grad_norm": 0.6417585015296936,
+      "learning_rate": 6.107166929563629e-06,
+      "loss": 0.6595,
+      "num_input_tokens_seen": 146964480,
+      "step": 17940
+    },
+    {
+      "epoch": 2.319421113838997,
+      "grad_norm": 0.4829157292842865,
+      "learning_rate": 6.085029808153503e-06,
+      "loss": 0.672,
+      "num_input_tokens_seen": 147046400,
+      "step": 17950
+    },
+    {
+      "epoch": 2.3207132704483784,
+      "grad_norm": 0.6301985383033752,
+      "learning_rate": 6.062927319984576e-06,
+      "loss": 0.7306,
+      "num_input_tokens_seen": 147128320,
+      "step": 17960
+    },
+    {
+      "epoch": 2.3220054270577593,
+      "grad_norm": 0.5591002106666565,
+      "learning_rate": 6.040859505526439e-06,
+      "loss": 1.0703,
+      "num_input_tokens_seen": 147210240,
+      "step": 17970
+    },
+    {
+      "epoch": 2.3232975836671406,
+      "grad_norm": 0.6914545893669128,
+      "learning_rate": 6.018826405185163e-06,
+      "loss": 0.9107,
+      "num_input_tokens_seen": 147292160,
+      "step": 17980
+    },
+    {
+      "epoch": 2.3245897402765214,
+      "grad_norm": 0.6147554516792297,
+      "learning_rate": 5.99682805930328e-06,
+      "loss": 0.9921,
+      "num_input_tokens_seen": 147374080,
+      "step": 17990
+    },
+    {
+      "epoch": 2.3258818968859027,
+      "grad_norm": 0.8620073795318604,
+      "learning_rate": 5.974864508159692e-06,
+      "loss": 1.121,
+      "num_input_tokens_seen": 147456000,
+      "step": 18000
+    },
+    {
+      "epoch": 2.3271740534952836,
+      "grad_norm": 0.7225034236907959,
+      "learning_rate": 5.952935791969574e-06,
+      "loss": 0.7416,
+      "num_input_tokens_seen": 147537920,
+      "step": 18010
+    },
+    {
+      "epoch": 2.328466210104665,
+      "grad_norm": 0.9885997772216797,
+      "learning_rate": 5.931041950884314e-06,
+      "loss": 0.8224,
+      "num_input_tokens_seen": 147619840,
+      "step": 18020
+    },
+    {
+      "epoch": 2.329758366714046,
+      "grad_norm": 0.8367599844932556,
+      "learning_rate": 5.9091830249914685e-06,
+      "loss": 0.4575,
+      "num_input_tokens_seen": 147701760,
+      "step": 18030
+    },
+    {
+      "epoch": 2.3310505233234267,
+      "grad_norm": 0.685955286026001,
+      "learning_rate": 5.887359054314648e-06,
+      "loss": 0.8888,
+      "num_input_tokens_seen": 147783680,
+      "step": 18040
+    },
+    {
+      "epoch": 2.332342679932808,
+      "grad_norm": 0.8157021403312683,
+      "learning_rate": 5.8655700788134535e-06,
+      "loss": 1.0576,
+      "num_input_tokens_seen": 147865600,
+      "step": 18050
+    },
+    {
+      "epoch": 2.333634836542189,
+      "grad_norm": 0.7296263575553894,
+      "learning_rate": 5.843816138383429e-06,
+      "loss": 0.967,
+      "num_input_tokens_seen": 147947520,
+      "step": 18060
+    },
+    {
+      "epoch": 2.33492699315157,
+      "grad_norm": 0.6871107816696167,
+      "learning_rate": 5.822097272855964e-06,
+      "loss": 0.5175,
+      "num_input_tokens_seen": 148029440,
+      "step": 18070
+    },
+    {
+      "epoch": 2.336219149760951,
+      "grad_norm": 0.9326339364051819,
+      "learning_rate": 5.800413521998208e-06,
+      "loss": 0.8177,
+      "num_input_tokens_seen": 148111360,
+      "step": 18080
+    },
+    {
+      "epoch": 2.337511306370332,
+      "grad_norm": 0.6087706089019775,
+      "learning_rate": 5.778764925513045e-06,
+      "loss": 0.9179,
+      "num_input_tokens_seen": 148193280,
+      "step": 18090
+    },
+    {
+      "epoch": 2.338803462979713,
+      "grad_norm": 1.1223865747451782,
+      "learning_rate": 5.7571515230389586e-06,
+      "loss": 0.9355,
+      "num_input_tokens_seen": 148275200,
+      "step": 18100
+    },
+    {
+      "epoch": 2.340095619589094,
+      "grad_norm": 0.6350956559181213,
+      "learning_rate": 5.7355733541500285e-06,
+      "loss": 0.9177,
+      "num_input_tokens_seen": 148357120,
+      "step": 18110
+    },
+    {
+      "epoch": 2.3413877761984754,
+      "grad_norm": 0.543329656124115,
+      "learning_rate": 5.714030458355784e-06,
+      "loss": 0.8673,
+      "num_input_tokens_seen": 148439040,
+      "step": 18120
+    },
+    {
+      "epoch": 2.3426799328078562,
+      "grad_norm": 0.8380277752876282,
+      "learning_rate": 5.692522875101203e-06,
+      "loss": 0.8191,
+      "num_input_tokens_seen": 148520960,
+      "step": 18130
+    },
+    {
+      "epoch": 2.3439720894172376,
+      "grad_norm": 0.6148279905319214,
+      "learning_rate": 5.67105064376659e-06,
+      "loss": 0.4773,
+      "num_input_tokens_seen": 148602880,
+      "step": 18140
+    },
+    {
+      "epoch": 2.3452642460266184,
+      "grad_norm": 0.766385555267334,
+      "learning_rate": 5.649613803667511e-06,
+      "loss": 0.8423,
+      "num_input_tokens_seen": 148684800,
+      "step": 18150
+    },
+    {
+      "epoch": 2.3465564026359993,
+      "grad_norm": 0.24485957622528076,
+      "learning_rate": 5.628212394054758e-06,
+      "loss": 0.7039,
+      "num_input_tokens_seen": 148766720,
+      "step": 18160
+    },
+    {
+      "epoch": 2.3478485592453806,
+      "grad_norm": 0.30911335349082947,
+      "learning_rate": 5.606846454114218e-06,
+      "loss": 0.4886,
+      "num_input_tokens_seen": 148848640,
+      "step": 18170
+    },
+    {
+      "epoch": 2.3491407158547615,
+      "grad_norm": 1.1630913019180298,
+      "learning_rate": 5.5855160229668636e-06,
+      "loss": 0.8072,
+      "num_input_tokens_seen": 148930560,
+      "step": 18180
+    },
+    {
+      "epoch": 2.350432872464143,
+      "grad_norm": 0.8715269565582275,
+      "learning_rate": 5.564221139668621e-06,
+      "loss": 0.7343,
+      "num_input_tokens_seen": 149012480,
+      "step": 18190
+    },
+    {
+      "epoch": 2.3517250290735237,
+      "grad_norm": 0.7479017376899719,
+      "learning_rate": 5.542961843210359e-06,
+      "loss": 1.1034,
+      "num_input_tokens_seen": 149094400,
+      "step": 18200
+    },
+    {
+      "epoch": 2.353017185682905,
+      "grad_norm": 0.23909921944141388,
+      "learning_rate": 5.5217381725177624e-06,
+      "loss": 0.7438,
+      "num_input_tokens_seen": 149176320,
+      "step": 18210
+    },
+    {
+      "epoch": 2.354309342292286,
+      "grad_norm": 0.6988628506660461,
+      "learning_rate": 5.50055016645129e-06,
+      "loss": 0.6506,
+      "num_input_tokens_seen": 149258240,
+      "step": 18220
+    },
+    {
+      "epoch": 2.3556014989016667,
+      "grad_norm": 0.9495237469673157,
+      "learning_rate": 5.479397863806115e-06,
+      "loss": 0.7191,
+      "num_input_tokens_seen": 149340160,
+      "step": 18230
+    },
+    {
+      "epoch": 2.356893655511048,
+      "grad_norm": 0.9350152015686035,
+      "learning_rate": 5.458281303312016e-06,
+      "loss": 0.8379,
+      "num_input_tokens_seen": 149422080,
+      "step": 18240
+    },
+    {
+      "epoch": 2.358185812120429,
+      "grad_norm": 0.8469577431678772,
+      "learning_rate": 5.437200523633348e-06,
+      "loss": 0.388,
+      "num_input_tokens_seen": 149504000,
+      "step": 18250
+    },
+    {
+      "epoch": 2.35947796872981,
+      "grad_norm": 0.6918286085128784,
+      "learning_rate": 5.41615556336893e-06,
+      "loss": 1.0104,
+      "num_input_tokens_seen": 149585920,
+      "step": 18260
+    },
+    {
+      "epoch": 2.360770125339191,
+      "grad_norm": 0.8202119469642639,
+      "learning_rate": 5.39514646105202e-06,
+      "loss": 0.7642,
+      "num_input_tokens_seen": 149667840,
+      "step": 18270
+    },
+    {
+      "epoch": 2.362062281948572,
+      "grad_norm": 0.20039023458957672,
+      "learning_rate": 5.374173255150194e-06,
+      "loss": 0.6885,
+      "num_input_tokens_seen": 149749760,
+      "step": 18280
+    },
+    {
+      "epoch": 2.3633544385579532,
+      "grad_norm": 44.923805236816406,
+      "learning_rate": 5.353235984065321e-06,
+      "loss": 0.6808,
+      "num_input_tokens_seen": 149831680,
+      "step": 18290
+    },
+    {
+      "epoch": 2.364646595167334,
+      "grad_norm": 0.7638149261474609,
+      "learning_rate": 5.332334686133475e-06,
+      "loss": 0.7185,
+      "num_input_tokens_seen": 149913600,
+      "step": 18300
+    },
+    {
+      "epoch": 2.3659387517767154,
+      "grad_norm": 0.9998169541358948,
+      "learning_rate": 5.311469399624844e-06,
+      "loss": 0.8391,
+      "num_input_tokens_seen": 149995520,
+      "step": 18310
+    },
+    {
+      "epoch": 2.3672309083860963,
+      "grad_norm": 0.6908437609672546,
+      "learning_rate": 5.290640162743704e-06,
+      "loss": 0.6135,
+      "num_input_tokens_seen": 150077440,
+      "step": 18320
+    },
+    {
+      "epoch": 2.3685230649954776,
+      "grad_norm": 1.3081281185150146,
+      "learning_rate": 5.269847013628299e-06,
+      "loss": 0.9426,
+      "num_input_tokens_seen": 150159360,
+      "step": 18330
+    },
+    {
+      "epoch": 2.3698152216048585,
+      "grad_norm": 0.7654116749763489,
+      "learning_rate": 5.24908999035082e-06,
+      "loss": 0.9153,
+      "num_input_tokens_seen": 150241280,
+      "step": 18340
+    },
+    {
+      "epoch": 2.37110737821424,
+      "grad_norm": 1.00165855884552,
+      "learning_rate": 5.228369130917288e-06,
+      "loss": 0.4039,
+      "num_input_tokens_seen": 150323200,
+      "step": 18350
+    },
+    {
+      "epoch": 2.3723995348236206,
+      "grad_norm": 0.593526303768158,
+      "learning_rate": 5.207684473267527e-06,
+      "loss": 0.573,
+      "num_input_tokens_seen": 150405120,
+      "step": 18360
+    },
+    {
+      "epoch": 2.3736916914330015,
+      "grad_norm": 0.5847054719924927,
+      "learning_rate": 5.187036055275077e-06,
+      "loss": 1.009,
+      "num_input_tokens_seen": 150487040,
+      "step": 18370
+    },
+    {
+      "epoch": 2.374983848042383,
+      "grad_norm": 0.7985702157020569,
+      "learning_rate": 5.16642391474711e-06,
+      "loss": 0.6435,
+      "num_input_tokens_seen": 150568960,
+      "step": 18380
+    },
+    {
+      "epoch": 2.3762760046517637,
+      "grad_norm": 0.8118183612823486,
+      "learning_rate": 5.145848089424374e-06,
+      "loss": 0.884,
+      "num_input_tokens_seen": 150650880,
+      "step": 18390
+    },
+    {
+      "epoch": 2.377568161261145,
+      "grad_norm": 0.7361833453178406,
+      "learning_rate": 5.125308616981139e-06,
+      "loss": 0.7895,
+      "num_input_tokens_seen": 150732800,
+      "step": 18400
+    },
+    {
+      "epoch": 2.378860317870526,
+      "grad_norm": 0.6324991583824158,
+      "learning_rate": 5.1048055350251e-06,
+      "loss": 0.971,
+      "num_input_tokens_seen": 150814720,
+      "step": 18410
+    },
+    {
+      "epoch": 2.3801524744799067,
+      "grad_norm": 0.899131178855896,
+      "learning_rate": 5.0843388810973195e-06,
+      "loss": 0.5018,
+      "num_input_tokens_seen": 150896640,
+      "step": 18420
+    },
+    {
+      "epoch": 2.381444631089288,
+      "grad_norm": 0.6457487344741821,
+      "learning_rate": 5.06390869267217e-06,
+      "loss": 0.953,
+      "num_input_tokens_seen": 150978560,
+      "step": 18430
+    },
+    {
+      "epoch": 2.382736787698669,
+      "grad_norm": 0.8277502655982971,
+      "learning_rate": 5.043515007157263e-06,
+      "loss": 1.2336,
+      "num_input_tokens_seen": 151060480,
+      "step": 18440
+    },
+    {
+      "epoch": 2.3840289443080502,
+      "grad_norm": 0.7911684513092041,
+      "learning_rate": 5.02315786189334e-06,
+      "loss": 0.6531,
+      "num_input_tokens_seen": 151142400,
+      "step": 18450
+    },
+    {
+      "epoch": 2.385321100917431,
+      "grad_norm": 0.945289671421051,
+      "learning_rate": 5.002837294154283e-06,
+      "loss": 0.7825,
+      "num_input_tokens_seen": 151224320,
+      "step": 18460
+    },
+    {
+      "epoch": 2.3866132575268124,
+      "grad_norm": 0.6046878695487976,
+      "learning_rate": 4.982553341146956e-06,
+      "loss": 0.7779,
+      "num_input_tokens_seen": 151306240,
+      "step": 18470
+    },
+    {
+      "epoch": 2.3879054141361933,
+      "grad_norm": 0.8228399753570557,
+      "learning_rate": 4.962306040011222e-06,
+      "loss": 0.9538,
+      "num_input_tokens_seen": 151388160,
+      "step": 18480
+    },
+    {
+      "epoch": 2.3891975707455746,
+      "grad_norm": 0.7043998837471008,
+      "learning_rate": 4.942095427819796e-06,
+      "loss": 1.0866,
+      "num_input_tokens_seen": 151470080,
+      "step": 18490
+    },
+    {
+      "epoch": 2.3904897273549555,
+      "grad_norm": 0.5962804555892944,
+      "learning_rate": 4.921921541578248e-06,
+      "loss": 0.8085,
+      "num_input_tokens_seen": 151552000,
+      "step": 18500
+    },
+    {
+      "epoch": 2.3917818839643363,
+      "grad_norm": 0.6313290596008301,
+      "learning_rate": 4.901784418224892e-06,
+      "loss": 0.8541,
+      "num_input_tokens_seen": 151633920,
+      "step": 18510
+    },
+    {
+      "epoch": 2.3930740405737176,
+      "grad_norm": 0.6041284203529358,
+      "learning_rate": 4.881684094630712e-06,
+      "loss": 0.6467,
+      "num_input_tokens_seen": 151715840,
+      "step": 18520
+    },
+    {
+      "epoch": 2.3943661971830985,
+      "grad_norm": 0.7616141438484192,
+      "learning_rate": 4.861620607599346e-06,
+      "loss": 0.7586,
+      "num_input_tokens_seen": 151797760,
+      "step": 18530
+    },
+    {
+      "epoch": 2.39565835379248,
+      "grad_norm": 0.3103398084640503,
+      "learning_rate": 4.841593993866949e-06,
+      "loss": 0.853,
+      "num_input_tokens_seen": 151879680,
+      "step": 18540
+    },
+    {
+      "epoch": 2.3969505104018607,
+      "grad_norm": 0.7408877015113831,
+      "learning_rate": 4.821604290102191e-06,
+      "loss": 1.02,
+      "num_input_tokens_seen": 151961600,
+      "step": 18550
+    },
+    {
+      "epoch": 2.3982426670112416,
+      "grad_norm": 0.24368135631084442,
+      "learning_rate": 4.801651532906135e-06,
+      "loss": 0.8047,
+      "num_input_tokens_seen": 152043520,
+      "step": 18560
+    },
+    {
+      "epoch": 2.399534823620623,
+      "grad_norm": 0.7442583441734314,
+      "learning_rate": 4.781735758812217e-06,
+      "loss": 0.9561,
+      "num_input_tokens_seen": 152125440,
+      "step": 18570
+    },
+    {
+      "epoch": 2.4008269802300037,
+      "grad_norm": 0.6959922313690186,
+      "learning_rate": 4.761857004286141e-06,
+      "loss": 0.6804,
+      "num_input_tokens_seen": 152207360,
+      "step": 18580
+    },
+    {
+      "epoch": 2.402119136839385,
+      "grad_norm": 1.0393434762954712,
+      "learning_rate": 4.742015305725828e-06,
+      "loss": 0.9148,
+      "num_input_tokens_seen": 152289280,
+      "step": 18590
+    },
+    {
+      "epoch": 2.403411293448766,
+      "grad_norm": 0.9378628134727478,
+      "learning_rate": 4.7222106994613655e-06,
+      "loss": 0.6088,
+      "num_input_tokens_seen": 152371200,
+      "step": 18600
+    },
+    {
+      "epoch": 2.4047034500581472,
+      "grad_norm": 0.8462372422218323,
+      "learning_rate": 4.702443221754904e-06,
+      "loss": 0.7719,
+      "num_input_tokens_seen": 152453120,
+      "step": 18610
+    },
+    {
+      "epoch": 2.405995606667528,
+      "grad_norm": 0.6147511005401611,
+      "learning_rate": 4.6827129088006375e-06,
+      "loss": 0.8772,
+      "num_input_tokens_seen": 152535040,
+      "step": 18620
+    },
+    {
+      "epoch": 2.4072877632769094,
+      "grad_norm": 0.8971914052963257,
+      "learning_rate": 4.663019796724685e-06,
+      "loss": 0.7122,
+      "num_input_tokens_seen": 152616960,
+      "step": 18630
+    },
+    {
+      "epoch": 2.4085799198862903,
+      "grad_norm": 0.7405256628990173,
+      "learning_rate": 4.6433639215850696e-06,
+      "loss": 1.0226,
+      "num_input_tokens_seen": 152698880,
+      "step": 18640
+    },
+    {
+      "epoch": 2.409872076495671,
+      "grad_norm": 0.3762902021408081,
+      "learning_rate": 4.623745319371617e-06,
+      "loss": 0.6679,
+      "num_input_tokens_seen": 152780800,
+      "step": 18650
+    },
+    {
+      "epoch": 2.4111642331050525,
+      "grad_norm": 0.9347968697547913,
+      "learning_rate": 4.604164026005925e-06,
+      "loss": 1.0588,
+      "num_input_tokens_seen": 152862720,
+      "step": 18660
+    },
+    {
+      "epoch": 2.4124563897144333,
+      "grad_norm": 0.7165658473968506,
+      "learning_rate": 4.584620077341273e-06,
+      "loss": 0.8622,
+      "num_input_tokens_seen": 152944640,
+      "step": 18670
+    },
+    {
+      "epoch": 2.4137485463238146,
+      "grad_norm": 0.715330958366394,
+      "learning_rate": 4.565113509162547e-06,
+      "loss": 0.9585,
+      "num_input_tokens_seen": 153026560,
+      "step": 18680
+    },
+    {
+      "epoch": 2.4150407029331955,
+      "grad_norm": 0.5120941996574402,
+      "learning_rate": 4.5456443571862185e-06,
+      "loss": 0.5105,
+      "num_input_tokens_seen": 153108480,
+      "step": 18690
+    },
+    {
+      "epoch": 2.4163328595425764,
+      "grad_norm": 0.6818029284477234,
+      "learning_rate": 4.5262126570602135e-06,
+      "loss": 0.7058,
+      "num_input_tokens_seen": 153190400,
+      "step": 18700
+    },
+    {
+      "epoch": 2.4176250161519577,
+      "grad_norm": 0.3870543837547302,
+      "learning_rate": 4.506818444363925e-06,
+      "loss": 0.7815,
+      "num_input_tokens_seen": 153272320,
+      "step": 18710
+    },
+    {
+      "epoch": 2.4189171727613386,
+      "grad_norm": 0.23925046622753143,
+      "learning_rate": 4.487461754608066e-06,
+      "loss": 0.6579,
+      "num_input_tokens_seen": 153354240,
+      "step": 18720
+    },
+    {
+      "epoch": 2.42020932937072,
+      "grad_norm": 0.25111180543899536,
+      "learning_rate": 4.468142623234678e-06,
+      "loss": 0.5656,
+      "num_input_tokens_seen": 153436160,
+      "step": 18730
+    },
+    {
+      "epoch": 2.4215014859801007,
+      "grad_norm": 0.6497703790664673,
+      "learning_rate": 4.448861085617018e-06,
+      "loss": 0.9916,
+      "num_input_tokens_seen": 153518080,
+      "step": 18740
+    },
+    {
+      "epoch": 2.4227936425894816,
+      "grad_norm": 0.30985990166664124,
+      "learning_rate": 4.429617177059508e-06,
+      "loss": 0.3525,
+      "num_input_tokens_seen": 153600000,
+      "step": 18750
+    },
+    {
+      "epoch": 2.424085799198863,
+      "grad_norm": 0.848656177520752,
+      "learning_rate": 4.410410932797671e-06,
+      "loss": 0.912,
+      "num_input_tokens_seen": 153681920,
+      "step": 18760
+    },
+    {
+      "epoch": 2.425377955808244,
+      "grad_norm": 0.6000288724899292,
+      "learning_rate": 4.391242387998079e-06,
+      "loss": 0.8084,
+      "num_input_tokens_seen": 153763840,
+      "step": 18770
+    },
+    {
+      "epoch": 2.426670112417625,
+      "grad_norm": 0.6745223999023438,
+      "learning_rate": 4.372111577758261e-06,
+      "loss": 0.8788,
+      "num_input_tokens_seen": 153845760,
+      "step": 18780
+    },
+    {
+      "epoch": 2.427962269027006,
+      "grad_norm": 0.9669370651245117,
+      "learning_rate": 4.353018537106657e-06,
+      "loss": 0.8239,
+      "num_input_tokens_seen": 153927680,
+      "step": 18790
+    },
+    {
+      "epoch": 2.4292544256363873,
+      "grad_norm": 0.9055847525596619,
+      "learning_rate": 4.333963301002558e-06,
+      "loss": 0.6755,
+      "num_input_tokens_seen": 154009600,
+      "step": 18800
+    },
+    {
+      "epoch": 2.430546582245768,
+      "grad_norm": 0.5974034070968628,
+      "learning_rate": 4.314945904336037e-06,
+      "loss": 0.8883,
+      "num_input_tokens_seen": 154091520,
+      "step": 18810
+    },
+    {
+      "epoch": 2.4318387388551495,
+      "grad_norm": 0.6422154307365417,
+      "learning_rate": 4.295966381927871e-06,
+      "loss": 0.9761,
+      "num_input_tokens_seen": 154173440,
+      "step": 18820
+    },
+    {
+      "epoch": 2.4331308954645303,
+      "grad_norm": 1.6391959190368652,
+      "learning_rate": 4.2770247685295e-06,
+      "loss": 0.7079,
+      "num_input_tokens_seen": 154255360,
+      "step": 18830
+    },
+    {
+      "epoch": 2.434423052073911,
+      "grad_norm": 1.442500114440918,
+      "learning_rate": 4.258121098822945e-06,
+      "loss": 0.8145,
+      "num_input_tokens_seen": 154337280,
+      "step": 18840
+    },
+    {
+      "epoch": 2.4357152086832925,
+      "grad_norm": 0.6912276744842529,
+      "learning_rate": 4.239255407420764e-06,
+      "loss": 0.668,
+      "num_input_tokens_seen": 154419200,
+      "step": 18850
+    },
+    {
+      "epoch": 2.4370073652926734,
+      "grad_norm": 0.5514190196990967,
+      "learning_rate": 4.220427728865956e-06,
+      "loss": 0.635,
+      "num_input_tokens_seen": 154501120,
+      "step": 18860
+    },
+    {
+      "epoch": 2.4382995219020547,
+      "grad_norm": 0.8538394570350647,
+      "learning_rate": 4.201638097631938e-06,
+      "loss": 0.8883,
+      "num_input_tokens_seen": 154583040,
+      "step": 18870
+    },
+    {
+      "epoch": 2.4395916785114355,
+      "grad_norm": 1.320961356163025,
+      "learning_rate": 4.182886548122464e-06,
+      "loss": 0.6052,
+      "num_input_tokens_seen": 154664960,
+      "step": 18880
+    },
+    {
+      "epoch": 2.4408838351208164,
+      "grad_norm": 0.70893394947052,
+      "learning_rate": 4.164173114671538e-06,
+      "loss": 0.5808,
+      "num_input_tokens_seen": 154746880,
+      "step": 18890
+    },
+    {
+      "epoch": 2.4421759917301977,
+      "grad_norm": 0.386322021484375,
+      "learning_rate": 4.145497831543402e-06,
+      "loss": 1.0335,
+      "num_input_tokens_seen": 154828800,
+      "step": 18900
+    },
+    {
+      "epoch": 2.4434681483395786,
+      "grad_norm": 0.5800808668136597,
+      "learning_rate": 4.1268607329324195e-06,
+      "loss": 0.8685,
+      "num_input_tokens_seen": 154910720,
+      "step": 18910
+    },
+    {
+      "epoch": 2.44476030494896,
+      "grad_norm": 1.0976125001907349,
+      "learning_rate": 4.108261852963061e-06,
+      "loss": 0.9031,
+      "num_input_tokens_seen": 154992640,
+      "step": 18920
+    },
+    {
+      "epoch": 2.4460524615583408,
+      "grad_norm": 0.9657114148139954,
+      "learning_rate": 4.089701225689793e-06,
+      "loss": 0.6177,
+      "num_input_tokens_seen": 155074560,
+      "step": 18930
+    },
+    {
+      "epoch": 2.447344618167722,
+      "grad_norm": 0.718108594417572,
+      "learning_rate": 4.071178885097074e-06,
+      "loss": 0.6072,
+      "num_input_tokens_seen": 155156480,
+      "step": 18940
+    },
+    {
+      "epoch": 2.448636774777103,
+      "grad_norm": 0.8545438051223755,
+      "learning_rate": 4.052694865099232e-06,
+      "loss": 0.8248,
+      "num_input_tokens_seen": 155238400,
+      "step": 18950
+    },
+    {
+      "epoch": 2.4499289313864843,
+      "grad_norm": 0.739714503288269,
+      "learning_rate": 4.034249199540432e-06,
+      "loss": 0.8951,
+      "num_input_tokens_seen": 155320320,
+      "step": 18960
+    },
+    {
+      "epoch": 2.451221087995865,
+      "grad_norm": 0.3881072998046875,
+      "learning_rate": 4.015841922194638e-06,
+      "loss": 0.6641,
+      "num_input_tokens_seen": 155402240,
+      "step": 18970
+    },
+    {
+      "epoch": 2.452513244605246,
+      "grad_norm": 0.2270699441432953,
+      "learning_rate": 3.997473066765489e-06,
+      "loss": 0.3194,
+      "num_input_tokens_seen": 155484160,
+      "step": 18980
+    },
+    {
+      "epoch": 2.4538054012146273,
+      "grad_norm": 0.6020457744598389,
+      "learning_rate": 3.97914266688631e-06,
+      "loss": 0.8105,
+      "num_input_tokens_seen": 155566080,
+      "step": 18990
+    },
+    {
+      "epoch": 2.455097557824008,
+      "grad_norm": 1.502171277999878,
+      "learning_rate": 3.96085075611998e-06,
+      "loss": 0.7238,
+      "num_input_tokens_seen": 155648000,
+      "step": 19000
+    },
+    {
+      "epoch": 2.4563897144333895,
+      "grad_norm": 0.7338537573814392,
+      "learning_rate": 3.942597367958928e-06,
+      "loss": 0.9272,
+      "num_input_tokens_seen": 155729920,
+      "step": 19010
+    },
+    {
+      "epoch": 2.4576818710427704,
+      "grad_norm": 0.7727378010749817,
+      "learning_rate": 3.924382535825047e-06,
+      "loss": 0.5264,
+      "num_input_tokens_seen": 155811840,
+      "step": 19020
+    },
+    {
+      "epoch": 2.4589740276521512,
+      "grad_norm": 0.7787820100784302,
+      "learning_rate": 3.906206293069617e-06,
+      "loss": 0.8021,
+      "num_input_tokens_seen": 155893760,
+      "step": 19030
+    },
+    {
+      "epoch": 2.4602661842615325,
+      "grad_norm": 0.7348833084106445,
+      "learning_rate": 3.88806867297328e-06,
+      "loss": 0.9558,
+      "num_input_tokens_seen": 155975680,
+      "step": 19040
+    },
+    {
+      "epoch": 2.4615583408709134,
+      "grad_norm": 0.7787769436836243,
+      "learning_rate": 3.869969708745946e-06,
+      "loss": 1.0861,
+      "num_input_tokens_seen": 156057600,
+      "step": 19050
+    },
+    {
+      "epoch": 2.4628504974802947,
+      "grad_norm": 0.7689657807350159,
+      "learning_rate": 3.85190943352676e-06,
+      "loss": 0.9696,
+      "num_input_tokens_seen": 156139520,
+      "step": 19060
+    },
+    {
+      "epoch": 2.4641426540896756,
+      "grad_norm": 1.0891441106796265,
+      "learning_rate": 3.833887880384007e-06,
+      "loss": 0.7405,
+      "num_input_tokens_seen": 156221440,
+      "step": 19070
+    },
+    {
+      "epoch": 2.465434810699057,
+      "grad_norm": 0.4140090048313141,
+      "learning_rate": 3.815905082315102e-06,
+      "loss": 0.6941,
+      "num_input_tokens_seen": 156303360,
+      "step": 19080
+    },
+    {
+      "epoch": 2.4667269673084378,
+      "grad_norm": 0.9087291359901428,
+      "learning_rate": 3.7979610722464643e-06,
+      "loss": 0.472,
+      "num_input_tokens_seen": 156385280,
+      "step": 19090
+    },
+    {
+      "epoch": 2.468019123917819,
+      "grad_norm": 0.9746045470237732,
+      "learning_rate": 3.780055883033523e-06,
+      "loss": 1.1217,
+      "num_input_tokens_seen": 156467200,
+      "step": 19100
+    },
+    {
+      "epoch": 2.4693112805272,
+      "grad_norm": 0.43733227252960205,
+      "learning_rate": 3.762189547460615e-06,
+      "loss": 0.7157,
+      "num_input_tokens_seen": 156549120,
+      "step": 19110
+    },
+    {
+      "epoch": 2.470603437136581,
+      "grad_norm": 0.6600936055183411,
+      "learning_rate": 3.7443620982409305e-06,
+      "loss": 0.9332,
+      "num_input_tokens_seen": 156631040,
+      "step": 19120
+    },
+    {
+      "epoch": 2.471895593745962,
+      "grad_norm": 0.22970102727413177,
+      "learning_rate": 3.7265735680164615e-06,
+      "loss": 0.5993,
+      "num_input_tokens_seen": 156712960,
+      "step": 19130
+    },
+    {
+      "epoch": 2.473187750355343,
+      "grad_norm": 0.37218940258026123,
+      "learning_rate": 3.7088239893579456e-06,
+      "loss": 0.8641,
+      "num_input_tokens_seen": 156794880,
+      "step": 19140
+    },
+    {
+      "epoch": 2.4744799069647243,
+      "grad_norm": 0.8411920666694641,
+      "learning_rate": 3.6911133947648002e-06,
+      "loss": 0.5555,
+      "num_input_tokens_seen": 156876800,
+      "step": 19150
+    },
+    {
+      "epoch": 2.475772063574105,
+      "grad_norm": 0.7639785408973694,
+      "learning_rate": 3.6734418166650436e-06,
+      "loss": 0.7846,
+      "num_input_tokens_seen": 156958720,
+      "step": 19160
+    },
+    {
+      "epoch": 2.477064220183486,
+      "grad_norm": 0.5358415842056274,
+      "learning_rate": 3.655809287415285e-06,
+      "loss": 0.7447,
+      "num_input_tokens_seen": 157040640,
+      "step": 19170
+    },
+    {
+      "epoch": 2.4783563767928674,
+      "grad_norm": 0.6833941340446472,
+      "learning_rate": 3.638215839300624e-06,
+      "loss": 0.8456,
+      "num_input_tokens_seen": 157122560,
+      "step": 19180
+    },
+    {
+      "epoch": 2.4796485334022482,
+      "grad_norm": 0.5907963514328003,
+      "learning_rate": 3.6206615045345837e-06,
+      "loss": 0.9318,
+      "num_input_tokens_seen": 157204480,
+      "step": 19190
+    },
+    {
+      "epoch": 2.4809406900116295,
+      "grad_norm": 0.9480105638504028,
+      "learning_rate": 3.603146315259104e-06,
+      "loss": 1.074,
+      "num_input_tokens_seen": 157286400,
+      "step": 19200
+    },
+    {
+      "epoch": 2.4822328466210104,
+      "grad_norm": 0.9866535067558289,
+      "learning_rate": 3.5856703035444196e-06,
+      "loss": 0.6913,
+      "num_input_tokens_seen": 157368320,
+      "step": 19210
+    },
+    {
+      "epoch": 2.4835250032303917,
+      "grad_norm": 0.965872585773468,
+      "learning_rate": 3.568233501389054e-06,
+      "loss": 0.7036,
+      "num_input_tokens_seen": 157450240,
+      "step": 19220
+    },
+    {
+      "epoch": 2.4848171598397726,
+      "grad_norm": 0.752018928527832,
+      "learning_rate": 3.5508359407197157e-06,
+      "loss": 0.7991,
+      "num_input_tokens_seen": 157532160,
+      "step": 19230
+    },
+    {
+      "epoch": 2.4861093164491535,
+      "grad_norm": 0.6639631986618042,
+      "learning_rate": 3.5334776533912846e-06,
+      "loss": 0.6216,
+      "num_input_tokens_seen": 157614080,
+      "step": 19240
+    },
+    {
+      "epoch": 2.4874014730585348,
+      "grad_norm": 0.7294473052024841,
+      "learning_rate": 3.516158671186723e-06,
+      "loss": 0.9294,
+      "num_input_tokens_seen": 157696000,
+      "step": 19250
+    },
+    {
+      "epoch": 2.4886936296679156,
+      "grad_norm": 0.920391321182251,
+      "learning_rate": 3.4988790258170146e-06,
+      "loss": 0.8402,
+      "num_input_tokens_seen": 157777920,
+      "step": 19260
+    },
+    {
+      "epoch": 2.489985786277297,
+      "grad_norm": 1.0001167058944702,
+      "learning_rate": 3.481638748921137e-06,
+      "loss": 0.8772,
+      "num_input_tokens_seen": 157859840,
+      "step": 19270
+    },
+    {
+      "epoch": 2.491277942886678,
+      "grad_norm": 1.0365166664123535,
+      "learning_rate": 3.4644378720659648e-06,
+      "loss": 0.5995,
+      "num_input_tokens_seen": 157941760,
+      "step": 19280
+    },
+    {
+      "epoch": 2.492570099496059,
+      "grad_norm": 0.8598686456680298,
+      "learning_rate": 3.4472764267462486e-06,
+      "loss": 0.8775,
+      "num_input_tokens_seen": 158023680,
+      "step": 19290
+    },
+    {
+      "epoch": 2.49386225610544,
+      "grad_norm": 0.7252724170684814,
+      "learning_rate": 3.430154444384523e-06,
+      "loss": 0.827,
+      "num_input_tokens_seen": 158105600,
+      "step": 19300
+    },
+    {
+      "epoch": 2.495154412714821,
+      "grad_norm": 0.9044560194015503,
+      "learning_rate": 3.4130719563310877e-06,
+      "loss": 0.6568,
+      "num_input_tokens_seen": 158187520,
+      "step": 19310
+    },
+    {
+      "epoch": 2.496446569324202,
+      "grad_norm": 0.3147662281990051,
+      "learning_rate": 3.396028993863906e-06,
+      "loss": 0.8331,
+      "num_input_tokens_seen": 158269440,
+      "step": 19320
+    },
+    {
+      "epoch": 2.497738725933583,
+      "grad_norm": 0.6290098428726196,
+      "learning_rate": 3.379025588188578e-06,
+      "loss": 0.6592,
+      "num_input_tokens_seen": 158351360,
+      "step": 19330
+    },
+    {
+      "epoch": 2.4990308825429643,
+      "grad_norm": 0.49024534225463867,
+      "learning_rate": 3.362061770438285e-06,
+      "loss": 0.9447,
+      "num_input_tokens_seen": 158433280,
+      "step": 19340
+    },
+    {
+      "epoch": 2.500323039152345,
+      "grad_norm": 1.2774690389633179,
+      "learning_rate": 3.3451375716737067e-06,
+      "loss": 0.4547,
+      "num_input_tokens_seen": 158515200,
+      "step": 19350
+    },
+    {
+      "epoch": 2.501615195761726,
+      "grad_norm": 0.6173495650291443,
+      "learning_rate": 3.328253022883002e-06,
+      "loss": 0.8312,
+      "num_input_tokens_seen": 158597120,
+      "step": 19360
+    },
+    {
+      "epoch": 2.5029073523711074,
+      "grad_norm": 0.7313811779022217,
+      "learning_rate": 3.3114081549817018e-06,
+      "loss": 0.744,
+      "num_input_tokens_seen": 158679040,
+      "step": 19370
+    },
+    {
+      "epoch": 2.5041995089804887,
+      "grad_norm": 0.28148242831230164,
+      "learning_rate": 3.2946029988127068e-06,
+      "loss": 0.6551,
+      "num_input_tokens_seen": 158760960,
+      "step": 19380
+    },
+    {
+      "epoch": 2.5054916655898696,
+      "grad_norm": 0.8187404870986938,
+      "learning_rate": 3.2778375851462013e-06,
+      "loss": 1.1236,
+      "num_input_tokens_seen": 158842880,
+      "step": 19390
+    },
+    {
+      "epoch": 2.5067838221992504,
+      "grad_norm": 0.7850127220153809,
+      "learning_rate": 3.2611119446795844e-06,
+      "loss": 0.7408,
+      "num_input_tokens_seen": 158924800,
+      "step": 19400
+    },
+    {
+      "epoch": 2.5080759788086318,
+      "grad_norm": 0.9370409250259399,
+      "learning_rate": 3.2444261080374546e-06,
+      "loss": 0.8042,
+      "num_input_tokens_seen": 159006720,
+      "step": 19410
+    },
+    {
+      "epoch": 2.5093681354180126,
+      "grad_norm": 0.6807863116264343,
+      "learning_rate": 3.227780105771505e-06,
+      "loss": 0.7672,
+      "num_input_tokens_seen": 159088640,
+      "step": 19420
+    },
+    {
+      "epoch": 2.510660292027394,
+      "grad_norm": 0.713300347328186,
+      "learning_rate": 3.2111739683605204e-06,
+      "loss": 0.9146,
+      "num_input_tokens_seen": 159170560,
+      "step": 19430
+    },
+    {
+      "epoch": 2.511952448636775,
+      "grad_norm": 0.7816091179847717,
+      "learning_rate": 3.194607726210261e-06,
+      "loss": 0.8129,
+      "num_input_tokens_seen": 159252480,
+      "step": 19440
+    },
+    {
+      "epoch": 2.5132446052461557,
+      "grad_norm": 0.1920579969882965,
+      "learning_rate": 3.178081409653469e-06,
+      "loss": 0.6187,
+      "num_input_tokens_seen": 159334400,
+      "step": 19450
+    },
+    {
+      "epoch": 2.514536761855537,
+      "grad_norm": 0.6369209885597229,
+      "learning_rate": 3.1615950489497587e-06,
+      "loss": 0.7696,
+      "num_input_tokens_seen": 159416320,
+      "step": 19460
+    },
+    {
+      "epoch": 2.515828918464918,
+      "grad_norm": 0.4802658259868622,
+      "learning_rate": 3.1451486742856055e-06,
+      "loss": 0.8106,
+      "num_input_tokens_seen": 159498240,
+      "step": 19470
+    },
+    {
+      "epoch": 2.517121075074299,
+      "grad_norm": 0.8319321274757385,
+      "learning_rate": 3.128742315774255e-06,
+      "loss": 0.8289,
+      "num_input_tokens_seen": 159580160,
+      "step": 19480
+    },
+    {
+      "epoch": 2.51841323168368,
+      "grad_norm": 0.8333976864814758,
+      "learning_rate": 3.1123760034556943e-06,
+      "loss": 0.4341,
+      "num_input_tokens_seen": 159662080,
+      "step": 19490
+    },
+    {
+      "epoch": 2.519705388293061,
+      "grad_norm": 1.237228274345398,
+      "learning_rate": 3.0960497672965825e-06,
+      "loss": 0.7386,
+      "num_input_tokens_seen": 159744000,
+      "step": 19500
+    },
+    {
+      "epoch": 2.520997544902442,
+      "grad_norm": 0.8341740369796753,
+      "learning_rate": 3.0797636371901863e-06,
+      "loss": 0.6727,
+      "num_input_tokens_seen": 159825920,
+      "step": 19510
+    },
+    {
+      "epoch": 2.522289701511823,
+      "grad_norm": 0.6844837069511414,
+      "learning_rate": 3.063517642956365e-06,
+      "loss": 0.8113,
+      "num_input_tokens_seen": 159907840,
+      "step": 19520
+    },
+    {
+      "epoch": 2.5235818581212044,
+      "grad_norm": 0.8005863428115845,
+      "learning_rate": 3.0473118143414634e-06,
+      "loss": 0.793,
+      "num_input_tokens_seen": 159989760,
+      "step": 19530
+    },
+    {
+      "epoch": 2.5248740147305853,
+      "grad_norm": 0.6440595984458923,
+      "learning_rate": 3.031146181018299e-06,
+      "loss": 0.8666,
+      "num_input_tokens_seen": 160071680,
+      "step": 19540
+    },
+    {
+      "epoch": 2.526166171339966,
+      "grad_norm": 0.6987878680229187,
+      "learning_rate": 3.0150207725860912e-06,
+      "loss": 0.6603,
+      "num_input_tokens_seen": 160153600,
+      "step": 19550
+    },
+    {
+      "epoch": 2.5274583279493474,
+      "grad_norm": 0.9995496273040771,
+      "learning_rate": 2.9989356185703975e-06,
+      "loss": 0.6602,
+      "num_input_tokens_seen": 160235520,
+      "step": 19560
+    },
+    {
+      "epoch": 2.5287504845587288,
+      "grad_norm": 0.8544566035270691,
+      "learning_rate": 2.982890748423084e-06,
+      "loss": 0.4753,
+      "num_input_tokens_seen": 160317440,
+      "step": 19570
+    },
+    {
+      "epoch": 2.5300426411681096,
+      "grad_norm": 0.9914432764053345,
+      "learning_rate": 2.9668861915222364e-06,
+      "loss": 0.9147,
+      "num_input_tokens_seen": 160399360,
+      "step": 19580
+    },
+    {
+      "epoch": 2.5313347977774905,
+      "grad_norm": 0.5936645865440369,
+      "learning_rate": 2.950921977172155e-06,
+      "loss": 0.9882,
+      "num_input_tokens_seen": 160481280,
+      "step": 19590
+    },
+    {
+      "epoch": 2.532626954386872,
+      "grad_norm": 0.45081570744514465,
+      "learning_rate": 2.934998134603245e-06,
+      "loss": 0.5187,
+      "num_input_tokens_seen": 160563200,
+      "step": 19600
+    },
+    {
+      "epoch": 2.5339191109962527,
+      "grad_norm": 0.7718636393547058,
+      "learning_rate": 2.919114692972008e-06,
+      "loss": 0.7276,
+      "num_input_tokens_seen": 160645120,
+      "step": 19610
+    },
+    {
+      "epoch": 2.535211267605634,
+      "grad_norm": 0.371920108795166,
+      "learning_rate": 2.9032716813609723e-06,
+      "loss": 0.6929,
+      "num_input_tokens_seen": 160727040,
+      "step": 19620
+    },
+    {
+      "epoch": 2.536503424215015,
+      "grad_norm": 0.7801430225372314,
+      "learning_rate": 2.8874691287786275e-06,
+      "loss": 0.5275,
+      "num_input_tokens_seen": 160808960,
+      "step": 19630
+    },
+    {
+      "epoch": 2.5377955808243957,
+      "grad_norm": 0.8307329416275024,
+      "learning_rate": 2.8717070641593987e-06,
+      "loss": 0.6655,
+      "num_input_tokens_seen": 160890880,
+      "step": 19640
+    },
+    {
+      "epoch": 2.539087737433777,
+      "grad_norm": 0.44690385460853577,
+      "learning_rate": 2.8559855163635544e-06,
+      "loss": 0.6951,
+      "num_input_tokens_seen": 160972800,
+      "step": 19650
+    },
+    {
+      "epoch": 2.540379894043158,
+      "grad_norm": 1.1183629035949707,
+      "learning_rate": 2.8403045141772054e-06,
+      "loss": 0.6766,
+      "num_input_tokens_seen": 161054720,
+      "step": 19660
+    },
+    {
+      "epoch": 2.541672050652539,
+      "grad_norm": 0.7904671430587769,
+      "learning_rate": 2.824664086312204e-06,
+      "loss": 0.7144,
+      "num_input_tokens_seen": 161136640,
+      "step": 19670
+    },
+    {
+      "epoch": 2.54296420726192,
+      "grad_norm": 0.6050812602043152,
+      "learning_rate": 2.809064261406111e-06,
+      "loss": 0.8016,
+      "num_input_tokens_seen": 161218560,
+      "step": 19680
+    },
+    {
+      "epoch": 2.544256363871301,
+      "grad_norm": 0.7455185651779175,
+      "learning_rate": 2.7935050680221565e-06,
+      "loss": 0.808,
+      "num_input_tokens_seen": 161300480,
+      "step": 19690
+    },
+    {
+      "epoch": 2.5455485204806823,
+      "grad_norm": 0.6713374853134155,
+      "learning_rate": 2.7779865346491576e-06,
+      "loss": 0.6117,
+      "num_input_tokens_seen": 161382400,
+      "step": 19700
+    },
+    {
+      "epoch": 2.5468406770900636,
+      "grad_norm": 0.7479690909385681,
+      "learning_rate": 2.762508689701504e-06,
+      "loss": 0.8894,
+      "num_input_tokens_seen": 161464320,
+      "step": 19710
+    },
+    {
+      "epoch": 2.5481328336994444,
+      "grad_norm": 0.8708047270774841,
+      "learning_rate": 2.74707156151906e-06,
+      "loss": 0.9919,
+      "num_input_tokens_seen": 161546240,
+      "step": 19720
+    },
+    {
+      "epoch": 2.5494249903088253,
+      "grad_norm": 0.33944186568260193,
+      "learning_rate": 2.7316751783671655e-06,
+      "loss": 1.0815,
+      "num_input_tokens_seen": 161628160,
+      "step": 19730
+    },
+    {
+      "epoch": 2.5507171469182066,
+      "grad_norm": 0.6622774004936218,
+      "learning_rate": 2.716319568436529e-06,
+      "loss": 0.9753,
+      "num_input_tokens_seen": 161710080,
+      "step": 19740
+    },
+    {
+      "epoch": 2.5520093035275875,
+      "grad_norm": 1.0370920896530151,
+      "learning_rate": 2.7010047598432205e-06,
+      "loss": 1.0534,
+      "num_input_tokens_seen": 161792000,
+      "step": 19750
+    },
+    {
+      "epoch": 2.553301460136969,
+      "grad_norm": 0.35142797231674194,
+      "learning_rate": 2.6857307806286037e-06,
+      "loss": 0.5268,
+      "num_input_tokens_seen": 161873920,
+      "step": 19760
+    },
+    {
+      "epoch": 2.5545936167463497,
+      "grad_norm": 1.3889226913452148,
+      "learning_rate": 2.6704976587592688e-06,
+      "loss": 0.7309,
+      "num_input_tokens_seen": 161955840,
+      "step": 19770
+    },
+    {
+      "epoch": 2.5558857733557305,
+      "grad_norm": 0.6227067708969116,
+      "learning_rate": 2.655305422127016e-06,
+      "loss": 0.9389,
+      "num_input_tokens_seen": 162037760,
+      "step": 19780
+    },
+    {
+      "epoch": 2.557177929965112,
+      "grad_norm": 0.35299497842788696,
+      "learning_rate": 2.6401540985487667e-06,
+      "loss": 0.6891,
+      "num_input_tokens_seen": 162119680,
+      "step": 19790
+    },
+    {
+      "epoch": 2.5584700865744927,
+      "grad_norm": 1.1497657299041748,
+      "learning_rate": 2.6250437157665455e-06,
+      "loss": 0.6537,
+      "num_input_tokens_seen": 162201600,
+      "step": 19800
+    },
+    {
+      "epoch": 2.559762243183874,
+      "grad_norm": 0.7763955593109131,
+      "learning_rate": 2.6099743014474014e-06,
+      "loss": 0.7283,
+      "num_input_tokens_seen": 162283520,
+      "step": 19810
+    },
+    {
+      "epoch": 2.561054399793255,
+      "grad_norm": 0.6793556809425354,
+      "learning_rate": 2.594945883183386e-06,
+      "loss": 0.6683,
+      "num_input_tokens_seen": 162365440,
+      "step": 19820
+    },
+    {
+      "epoch": 2.5623465564026358,
+      "grad_norm": 0.6524314880371094,
+      "learning_rate": 2.5799584884914685e-06,
+      "loss": 0.855,
+      "num_input_tokens_seen": 162447360,
+      "step": 19830
+    },
+    {
+      "epoch": 2.563638713012017,
+      "grad_norm": 1.6065878868103027,
+      "learning_rate": 2.5650121448135222e-06,
+      "loss": 0.9368,
+      "num_input_tokens_seen": 162529280,
+      "step": 19840
+    },
+    {
+      "epoch": 2.5649308696213984,
+      "grad_norm": 0.7918826937675476,
+      "learning_rate": 2.550106879516237e-06,
+      "loss": 0.5496,
+      "num_input_tokens_seen": 162611200,
+      "step": 19850
+    },
+    {
+      "epoch": 2.5662230262307792,
+      "grad_norm": 0.7481831312179565,
+      "learning_rate": 2.535242719891112e-06,
+      "loss": 1.0861,
+      "num_input_tokens_seen": 162693120,
+      "step": 19860
+    },
+    {
+      "epoch": 2.56751518284016,
+      "grad_norm": 0.519053041934967,
+      "learning_rate": 2.5204196931543635e-06,
+      "loss": 0.4623,
+      "num_input_tokens_seen": 162775040,
+      "step": 19870
+    },
+    {
+      "epoch": 2.5688073394495414,
+      "grad_norm": 0.7055248618125916,
+      "learning_rate": 2.505637826446891e-06,
+      "loss": 1.0828,
+      "num_input_tokens_seen": 162856960,
+      "step": 19880
+    },
+    {
+      "epoch": 2.5700994960589223,
+      "grad_norm": 0.2592199444770813,
+      "learning_rate": 2.4908971468342535e-06,
+      "loss": 0.8837,
+      "num_input_tokens_seen": 162938880,
+      "step": 19890
+    },
+    {
+      "epoch": 2.5713916526683036,
+      "grad_norm": 0.6449967622756958,
+      "learning_rate": 2.4761976813065663e-06,
+      "loss": 0.986,
+      "num_input_tokens_seen": 163020800,
+      "step": 19900
+    },
+    {
+      "epoch": 2.5726838092776845,
+      "grad_norm": 0.601736307144165,
+      "learning_rate": 2.4615394567785055e-06,
+      "loss": 0.9548,
+      "num_input_tokens_seen": 163102720,
+      "step": 19910
+    },
+    {
+      "epoch": 2.5739759658870653,
+      "grad_norm": 0.7824547290802002,
+      "learning_rate": 2.44692250008923e-06,
+      "loss": 1.0377,
+      "num_input_tokens_seen": 163184640,
+      "step": 19920
+    },
+    {
+      "epoch": 2.5752681224964467,
+      "grad_norm": 0.4208737015724182,
+      "learning_rate": 2.432346838002325e-06,
+      "loss": 0.8889,
+      "num_input_tokens_seen": 163266560,
+      "step": 19930
+    },
+    {
+      "epoch": 2.5765602791058275,
+      "grad_norm": 0.7913038730621338,
+      "learning_rate": 2.417812497205782e-06,
+      "loss": 0.7366,
+      "num_input_tokens_seen": 163348480,
+      "step": 19940
+    },
+    {
+      "epoch": 2.577852435715209,
+      "grad_norm": 1.0949358940124512,
+      "learning_rate": 2.403319504311921e-06,
+      "loss": 0.9359,
+      "num_input_tokens_seen": 163430400,
+      "step": 19950
+    },
+    {
+      "epoch": 2.5791445923245897,
+      "grad_norm": 0.877521276473999,
+      "learning_rate": 2.3888678858573625e-06,
+      "loss": 1.1218,
+      "num_input_tokens_seen": 163512320,
+      "step": 19960
+    },
+    {
+      "epoch": 2.5804367489339706,
+      "grad_norm": 0.8615255951881409,
+      "learning_rate": 2.374457668302962e-06,
+      "loss": 0.8245,
+      "num_input_tokens_seen": 163594240,
+      "step": 19970
+    },
+    {
+      "epoch": 2.581728905543352,
+      "grad_norm": 0.5929073691368103,
+      "learning_rate": 2.360088878033778e-06,
+      "loss": 1.0923,
+      "num_input_tokens_seen": 163676160,
+      "step": 19980
+    },
+    {
+      "epoch": 2.5830210621527327,
+      "grad_norm": 0.7284789681434631,
+      "learning_rate": 2.3457615413590177e-06,
+      "loss": 1.161,
+      "num_input_tokens_seen": 163758080,
+      "step": 19990
+    },
+    {
+      "epoch": 2.584313218762114,
+      "grad_norm": 1.2667911052703857,
+      "learning_rate": 2.3314756845119746e-06,
+      "loss": 0.9189,
+      "num_input_tokens_seen": 163840000,
+      "step": 20000
+    },
+    {
+      "epoch": 2.585605375371495,
+      "grad_norm": 0.672854483127594,
+      "learning_rate": 2.317231333650005e-06,
+      "loss": 0.8559,
+      "num_input_tokens_seen": 163921920,
+      "step": 20010
+    },
+    {
+      "epoch": 2.5868975319808762,
+      "grad_norm": 0.6795851588249207,
+      "learning_rate": 2.3030285148544577e-06,
+      "loss": 0.9338,
+      "num_input_tokens_seen": 164003840,
+      "step": 20020
+    },
+    {
+      "epoch": 2.588189688590257,
+      "grad_norm": 0.6645368933677673,
+      "learning_rate": 2.2888672541306525e-06,
+      "loss": 0.7432,
+      "num_input_tokens_seen": 164085760,
+      "step": 20030
+    },
+    {
+      "epoch": 2.5894818451996384,
+      "grad_norm": 0.7248007655143738,
+      "learning_rate": 2.2747475774077986e-06,
+      "loss": 0.7769,
+      "num_input_tokens_seen": 164167680,
+      "step": 20040
+    },
+    {
+      "epoch": 2.5907740018090193,
+      "grad_norm": 0.7931993007659912,
+      "learning_rate": 2.2606695105389653e-06,
+      "loss": 0.8439,
+      "num_input_tokens_seen": 164249600,
+      "step": 20050
+    },
+    {
+      "epoch": 2.5920661584184,
+      "grad_norm": 0.8612631559371948,
+      "learning_rate": 2.2466330793010555e-06,
+      "loss": 0.772,
+      "num_input_tokens_seen": 164331520,
+      "step": 20060
+    },
+    {
+      "epoch": 2.5933583150277815,
+      "grad_norm": 0.7526144981384277,
+      "learning_rate": 2.2326383093947135e-06,
+      "loss": 0.7069,
+      "num_input_tokens_seen": 164413440,
+      "step": 20070
+    },
+    {
+      "epoch": 2.5946504716371623,
+      "grad_norm": 1.1278603076934814,
+      "learning_rate": 2.2186852264443196e-06,
+      "loss": 0.9164,
+      "num_input_tokens_seen": 164495360,
+      "step": 20080
+    },
+    {
+      "epoch": 2.5959426282465436,
+      "grad_norm": 0.7620711326599121,
+      "learning_rate": 2.2047738559979104e-06,
+      "loss": 0.8105,
+      "num_input_tokens_seen": 164577280,
+      "step": 20090
+    },
+    {
+      "epoch": 2.5972347848559245,
+      "grad_norm": 0.6490961909294128,
+      "learning_rate": 2.1909042235271597e-06,
+      "loss": 1.0952,
+      "num_input_tokens_seen": 164659200,
+      "step": 20100
+    },
+    {
+      "epoch": 2.5985269414653054,
+      "grad_norm": 1.2163642644882202,
+      "learning_rate": 2.1770763544273098e-06,
+      "loss": 0.9695,
+      "num_input_tokens_seen": 164741120,
+      "step": 20110
+    },
+    {
+      "epoch": 2.5998190980746867,
+      "grad_norm": 0.4267684519290924,
+      "learning_rate": 2.1632902740171378e-06,
+      "loss": 0.8496,
+      "num_input_tokens_seen": 164823040,
+      "step": 20120
+    },
+    {
+      "epoch": 2.6011112546840676,
+      "grad_norm": 0.7075666785240173,
+      "learning_rate": 2.1495460075389133e-06,
+      "loss": 0.7883,
+      "num_input_tokens_seen": 164904960,
+      "step": 20130
+    },
+    {
+      "epoch": 2.602403411293449,
+      "grad_norm": 0.6190396547317505,
+      "learning_rate": 2.1358435801583283e-06,
+      "loss": 0.6911,
+      "num_input_tokens_seen": 164986880,
+      "step": 20140
+    },
+    {
+      "epoch": 2.6036955679028297,
+      "grad_norm": 0.3564266264438629,
+      "learning_rate": 2.122183016964488e-06,
+      "loss": 0.5446,
+      "num_input_tokens_seen": 165068800,
+      "step": 20150
+    },
+    {
+      "epoch": 2.6049877245122106,
+      "grad_norm": 0.7317832112312317,
+      "learning_rate": 2.1085643429698236e-06,
+      "loss": 0.7104,
+      "num_input_tokens_seen": 165150720,
+      "step": 20160
+    },
+    {
+      "epoch": 2.606279881121592,
+      "grad_norm": 0.6708076000213623,
+      "learning_rate": 2.094987583110086e-06,
+      "loss": 0.9044,
+      "num_input_tokens_seen": 165232640,
+      "step": 20170
+    },
+    {
+      "epoch": 2.6075720377309732,
+      "grad_norm": 0.6246494650840759,
+      "learning_rate": 2.0814527622442626e-06,
+      "loss": 0.687,
+      "num_input_tokens_seen": 165314560,
+      "step": 20180
+    },
+    {
+      "epoch": 2.608864194340354,
+      "grad_norm": 0.7877633571624756,
+      "learning_rate": 2.067959905154568e-06,
+      "loss": 0.791,
+      "num_input_tokens_seen": 165396480,
+      "step": 20190
+    },
+    {
+      "epoch": 2.610156350949735,
+      "grad_norm": 0.33362698554992676,
+      "learning_rate": 2.0545090365463788e-06,
+      "loss": 0.508,
+      "num_input_tokens_seen": 165478400,
+      "step": 20200
+    },
+    {
+      "epoch": 2.6114485075591163,
+      "grad_norm": 0.5828604102134705,
+      "learning_rate": 2.041100181048178e-06,
+      "loss": 0.9302,
+      "num_input_tokens_seen": 165560320,
+      "step": 20210
+    },
+    {
+      "epoch": 2.612740664168497,
+      "grad_norm": 0.7072322964668274,
+      "learning_rate": 2.0277333632115288e-06,
+      "loss": 0.7966,
+      "num_input_tokens_seen": 165642240,
+      "step": 20220
+    },
+    {
+      "epoch": 2.6140328207778785,
+      "grad_norm": 0.760685920715332,
+      "learning_rate": 2.0144086075110367e-06,
+      "loss": 1.1115,
+      "num_input_tokens_seen": 165724160,
+      "step": 20230
+    },
+    {
+      "epoch": 2.6153249773872593,
+      "grad_norm": 1.190176248550415,
+      "learning_rate": 2.001125938344273e-06,
+      "loss": 0.6785,
+      "num_input_tokens_seen": 165806080,
+      "step": 20240
+    },
+    {
+      "epoch": 2.61661713399664,
+      "grad_norm": 0.2874329686164856,
+      "learning_rate": 1.9878853800317535e-06,
+      "loss": 0.8553,
+      "num_input_tokens_seen": 165888000,
+      "step": 20250
+    },
+    {
+      "epoch": 2.6179092906060215,
+      "grad_norm": 1.3199117183685303,
+      "learning_rate": 1.9746869568168985e-06,
+      "loss": 0.7277,
+      "num_input_tokens_seen": 165969920,
+      "step": 20260
+    },
+    {
+      "epoch": 2.6192014472154024,
+      "grad_norm": 0.7847501039505005,
+      "learning_rate": 1.9615306928659677e-06,
+      "loss": 0.823,
+      "num_input_tokens_seen": 166051840,
+      "step": 20270
+    },
+    {
+      "epoch": 2.6204936038247837,
+      "grad_norm": 0.6932337284088135,
+      "learning_rate": 1.948416612268034e-06,
+      "loss": 0.9189,
+      "num_input_tokens_seen": 166133760,
+      "step": 20280
+    },
+    {
+      "epoch": 2.6217857604341646,
+      "grad_norm": 0.6914857029914856,
+      "learning_rate": 1.935344739034936e-06,
+      "loss": 0.8662,
+      "num_input_tokens_seen": 166215680,
+      "step": 20290
+    },
+    {
+      "epoch": 2.6230779170435454,
+      "grad_norm": 0.6377202868461609,
+      "learning_rate": 1.922315097101218e-06,
+      "loss": 0.9959,
+      "num_input_tokens_seen": 166297600,
+      "step": 20300
+    },
+    {
+      "epoch": 2.6243700736529267,
+      "grad_norm": 0.9114018082618713,
+      "learning_rate": 1.909327710324116e-06,
+      "loss": 0.4612,
+      "num_input_tokens_seen": 166379520,
+      "step": 20310
+    },
+    {
+      "epoch": 2.625662230262308,
+      "grad_norm": 0.7677037715911865,
+      "learning_rate": 1.8963826024834734e-06,
+      "loss": 0.9159,
+      "num_input_tokens_seen": 166461440,
+      "step": 20320
+    },
+    {
+      "epoch": 2.626954386871689,
+      "grad_norm": 0.4625994563102722,
+      "learning_rate": 1.8834797972817508e-06,
+      "loss": 0.6083,
+      "num_input_tokens_seen": 166543360,
+      "step": 20330
+    },
+    {
+      "epoch": 2.62824654348107,
+      "grad_norm": 0.6611701250076294,
+      "learning_rate": 1.8706193183439247e-06,
+      "loss": 0.8399,
+      "num_input_tokens_seen": 166625280,
+      "step": 20340
+    },
+    {
+      "epoch": 2.629538700090451,
+      "grad_norm": 0.528550386428833,
+      "learning_rate": 1.8578011892174924e-06,
+      "loss": 0.4694,
+      "num_input_tokens_seen": 166707200,
+      "step": 20350
+    },
+    {
+      "epoch": 2.630830856699832,
+      "grad_norm": 0.7156257033348083,
+      "learning_rate": 1.845025433372402e-06,
+      "loss": 1.0965,
+      "num_input_tokens_seen": 166789120,
+      "step": 20360
+    },
+    {
+      "epoch": 2.6321230133092133,
+      "grad_norm": 0.6360397934913635,
+      "learning_rate": 1.8322920742010086e-06,
+      "loss": 0.9704,
+      "num_input_tokens_seen": 166871040,
+      "step": 20370
+    },
+    {
+      "epoch": 2.633415169918594,
+      "grad_norm": 1.0048847198486328,
+      "learning_rate": 1.8196011350180563e-06,
+      "loss": 1.1178,
+      "num_input_tokens_seen": 166952960,
+      "step": 20380
+    },
+    {
+      "epoch": 2.634707326527975,
+      "grad_norm": 0.7727967500686646,
+      "learning_rate": 1.8069526390605968e-06,
+      "loss": 0.6866,
+      "num_input_tokens_seen": 167034880,
+      "step": 20390
+    },
+    {
+      "epoch": 2.6359994831373563,
+      "grad_norm": 0.8432693481445312,
+      "learning_rate": 1.7943466094879902e-06,
+      "loss": 0.8549,
+      "num_input_tokens_seen": 167116800,
+      "step": 20400
+    },
+    {
+      "epoch": 2.637291639746737,
+      "grad_norm": 0.6722807884216309,
+      "learning_rate": 1.7817830693818288e-06,
+      "loss": 0.7212,
+      "num_input_tokens_seen": 167198720,
+      "step": 20410
+    },
+    {
+      "epoch": 2.6385837963561185,
+      "grad_norm": 0.6678622961044312,
+      "learning_rate": 1.7692620417459004e-06,
+      "loss": 0.3765,
+      "num_input_tokens_seen": 167280640,
+      "step": 20420
+    },
+    {
+      "epoch": 2.6398759529654994,
+      "grad_norm": 0.7946398854255676,
+      "learning_rate": 1.7567835495061718e-06,
+      "loss": 0.9786,
+      "num_input_tokens_seen": 167362560,
+      "step": 20430
+    },
+    {
+      "epoch": 2.6411681095748802,
+      "grad_norm": 0.26235970854759216,
+      "learning_rate": 1.7443476155107052e-06,
+      "loss": 0.8443,
+      "num_input_tokens_seen": 167444480,
+      "step": 20440
+    },
+    {
+      "epoch": 2.6424602661842616,
+      "grad_norm": 0.9268401861190796,
+      "learning_rate": 1.7319542625296613e-06,
+      "loss": 0.9905,
+      "num_input_tokens_seen": 167526400,
+      "step": 20450
+    },
+    {
+      "epoch": 2.643752422793643,
+      "grad_norm": 0.7280152440071106,
+      "learning_rate": 1.7196035132552135e-06,
+      "loss": 0.7714,
+      "num_input_tokens_seen": 167608320,
+      "step": 20460
+    },
+    {
+      "epoch": 2.6450445794030237,
+      "grad_norm": 0.6951790452003479,
+      "learning_rate": 1.7072953903015498e-06,
+      "loss": 0.9703,
+      "num_input_tokens_seen": 167690240,
+      "step": 20470
+    },
+    {
+      "epoch": 2.6463367360124046,
+      "grad_norm": 0.5963300466537476,
+      "learning_rate": 1.6950299162047878e-06,
+      "loss": 0.7194,
+      "num_input_tokens_seen": 167772160,
+      "step": 20480
+    },
+    {
+      "epoch": 2.647628892621786,
+      "grad_norm": 1.0526518821716309,
+      "learning_rate": 1.682807113422971e-06,
+      "loss": 0.7812,
+      "num_input_tokens_seen": 167854080,
+      "step": 20490
+    },
+    {
+      "epoch": 2.648921049231167,
+      "grad_norm": 0.33847522735595703,
+      "learning_rate": 1.6706270043360117e-06,
+      "loss": 0.6726,
+      "num_input_tokens_seen": 167936000,
+      "step": 20500
+    },
+    {
+      "epoch": 2.650213205840548,
+      "grad_norm": 0.7132168412208557,
+      "learning_rate": 1.6584896112456338e-06,
+      "loss": 0.921,
+      "num_input_tokens_seen": 168017920,
+      "step": 20510
+    },
+    {
+      "epoch": 2.651505362449929,
+      "grad_norm": 0.6660868525505066,
+      "learning_rate": 1.646394956375369e-06,
+      "loss": 0.8371,
+      "num_input_tokens_seen": 168099840,
+      "step": 20520
+    },
+    {
+      "epoch": 2.65279751905931,
+      "grad_norm": 0.7083340287208557,
+      "learning_rate": 1.6343430618704775e-06,
+      "loss": 0.6175,
+      "num_input_tokens_seen": 168181760,
+      "step": 20530
+    },
+    {
+      "epoch": 2.654089675668691,
+      "grad_norm": 0.35995742678642273,
+      "learning_rate": 1.622333949797944e-06,
+      "loss": 0.6474,
+      "num_input_tokens_seen": 168263680,
+      "step": 20540
+    },
+    {
+      "epoch": 2.655381832278072,
+      "grad_norm": 0.6416080594062805,
+      "learning_rate": 1.6103676421463986e-06,
+      "loss": 0.6756,
+      "num_input_tokens_seen": 168345600,
+      "step": 20550
+    },
+    {
+      "epoch": 2.6566739888874533,
+      "grad_norm": 0.6493408679962158,
+      "learning_rate": 1.5984441608261152e-06,
+      "loss": 0.8549,
+      "num_input_tokens_seen": 168427520,
+      "step": 20560
+    },
+    {
+      "epoch": 2.657966145496834,
+      "grad_norm": 1.0145015716552734,
+      "learning_rate": 1.5865635276689412e-06,
+      "loss": 0.6964,
+      "num_input_tokens_seen": 168509440,
+      "step": 20570
+    },
+    {
+      "epoch": 2.659258302106215,
+      "grad_norm": 0.8252952098846436,
+      "learning_rate": 1.5747257644282726e-06,
+      "loss": 0.8505,
+      "num_input_tokens_seen": 168591360,
+      "step": 20580
+    },
+    {
+      "epoch": 2.6605504587155964,
+      "grad_norm": 1.0364081859588623,
+      "learning_rate": 1.5629308927790077e-06,
+      "loss": 0.4486,
+      "num_input_tokens_seen": 168673280,
+      "step": 20590
+    },
+    {
+      "epoch": 2.6618426153249772,
+      "grad_norm": 0.7188685536384583,
+      "learning_rate": 1.551178934317521e-06,
+      "loss": 0.7925,
+      "num_input_tokens_seen": 168755200,
+      "step": 20600
+    },
+    {
+      "epoch": 2.6631347719343585,
+      "grad_norm": 0.6135556101799011,
+      "learning_rate": 1.5394699105616002e-06,
+      "loss": 0.9941,
+      "num_input_tokens_seen": 168837120,
+      "step": 20610
+    },
+    {
+      "epoch": 2.6644269285437394,
+      "grad_norm": 0.39184391498565674,
+      "learning_rate": 1.5278038429504177e-06,
+      "loss": 0.8127,
+      "num_input_tokens_seen": 168919040,
+      "step": 20620
+    },
+    {
+      "epoch": 2.6657190851531203,
+      "grad_norm": 0.3787175416946411,
+      "learning_rate": 1.516180752844515e-06,
+      "loss": 0.3516,
+      "num_input_tokens_seen": 169000960,
+      "step": 20630
+    },
+    {
+      "epoch": 2.6670112417625016,
+      "grad_norm": 0.7270440459251404,
+      "learning_rate": 1.504600661525718e-06,
+      "loss": 0.7978,
+      "num_input_tokens_seen": 169082880,
+      "step": 20640
+    },
+    {
+      "epoch": 2.668303398371883,
+      "grad_norm": 0.8899626731872559,
+      "learning_rate": 1.493063590197133e-06,
+      "loss": 0.6851,
+      "num_input_tokens_seen": 169164800,
+      "step": 20650
+    },
+    {
+      "epoch": 2.6695955549812638,
+      "grad_norm": 0.7514258027076721,
+      "learning_rate": 1.4815695599830981e-06,
+      "loss": 1.0456,
+      "num_input_tokens_seen": 169246720,
+      "step": 20660
+    },
+    {
+      "epoch": 2.6708877115906446,
+      "grad_norm": 1.2644627094268799,
+      "learning_rate": 1.4701185919291372e-06,
+      "loss": 0.7482,
+      "num_input_tokens_seen": 169328640,
+      "step": 20670
+    },
+    {
+      "epoch": 2.672179868200026,
+      "grad_norm": 0.5645007491111755,
+      "learning_rate": 1.4587107070019368e-06,
+      "loss": 0.8197,
+      "num_input_tokens_seen": 169410560,
+      "step": 20680
+    },
+    {
+      "epoch": 2.673472024809407,
+      "grad_norm": 0.5567523241043091,
+      "learning_rate": 1.447345926089283e-06,
+      "loss": 0.8722,
+      "num_input_tokens_seen": 169492480,
+      "step": 20690
+    },
+    {
+      "epoch": 2.674764181418788,
+      "grad_norm": 1.2603198289871216,
+      "learning_rate": 1.436024270000058e-06,
+      "loss": 0.9786,
+      "num_input_tokens_seen": 169574400,
+      "step": 20700
+    },
+    {
+      "epoch": 2.676056338028169,
+      "grad_norm": 1.105191946029663,
+      "learning_rate": 1.4247457594641662e-06,
+      "loss": 0.8513,
+      "num_input_tokens_seen": 169656320,
+      "step": 20710
+    },
+    {
+      "epoch": 2.67734849463755,
+      "grad_norm": 0.5803696513175964,
+      "learning_rate": 1.4135104151325184e-06,
+      "loss": 0.8101,
+      "num_input_tokens_seen": 169738240,
+      "step": 20720
+    },
+    {
+      "epoch": 2.678640651246931,
+      "grad_norm": 0.6467001438140869,
+      "learning_rate": 1.4023182575769956e-06,
+      "loss": 0.5225,
+      "num_input_tokens_seen": 169820160,
+      "step": 20730
+    },
+    {
+      "epoch": 2.679932807856312,
+      "grad_norm": 1.2854204177856445,
+      "learning_rate": 1.391169307290391e-06,
+      "loss": 0.7088,
+      "num_input_tokens_seen": 169902080,
+      "step": 20740
+    },
+    {
+      "epoch": 2.6812249644656934,
+      "grad_norm": 0.6026389002799988,
+      "learning_rate": 1.3800635846863973e-06,
+      "loss": 0.9884,
+      "num_input_tokens_seen": 169984000,
+      "step": 20750
+    },
+    {
+      "epoch": 2.6825171210750742,
+      "grad_norm": 0.8329545259475708,
+      "learning_rate": 1.3690011100995437e-06,
+      "loss": 1.097,
+      "num_input_tokens_seen": 170065920,
+      "step": 20760
+    },
+    {
+      "epoch": 2.683809277684455,
+      "grad_norm": 1.0132107734680176,
+      "learning_rate": 1.357981903785191e-06,
+      "loss": 0.7501,
+      "num_input_tokens_seen": 170147840,
+      "step": 20770
+    },
+    {
+      "epoch": 2.6851014342938364,
+      "grad_norm": 0.5987882614135742,
+      "learning_rate": 1.3470059859194583e-06,
+      "loss": 0.9277,
+      "num_input_tokens_seen": 170229760,
+      "step": 20780
+    },
+    {
+      "epoch": 2.6863935909032177,
+      "grad_norm": 1.0641794204711914,
+      "learning_rate": 1.3360733765992116e-06,
+      "loss": 0.6113,
+      "num_input_tokens_seen": 170311680,
+      "step": 20790
+    },
+    {
+      "epoch": 2.6876857475125986,
+      "grad_norm": 0.6288622617721558,
+      "learning_rate": 1.325184095842022e-06,
+      "loss": 0.7514,
+      "num_input_tokens_seen": 170393600,
+      "step": 20800
+    },
+    {
+      "epoch": 2.6889779041219795,
+      "grad_norm": 1.762484073638916,
+      "learning_rate": 1.3143381635861207e-06,
+      "loss": 0.7529,
+      "num_input_tokens_seen": 170475520,
+      "step": 20810
+    },
+    {
+      "epoch": 2.6902700607313608,
+      "grad_norm": 0.8595255017280579,
+      "learning_rate": 1.3035355996903697e-06,
+      "loss": 0.8549,
+      "num_input_tokens_seen": 170557440,
+      "step": 20820
+    },
+    {
+      "epoch": 2.6915622173407416,
+      "grad_norm": 0.663703441619873,
+      "learning_rate": 1.2927764239342221e-06,
+      "loss": 0.7501,
+      "num_input_tokens_seen": 170639360,
+      "step": 20830
+    },
+    {
+      "epoch": 2.692854373950123,
+      "grad_norm": 0.9071587920188904,
+      "learning_rate": 1.2820606560176945e-06,
+      "loss": 0.9612,
+      "num_input_tokens_seen": 170721280,
+      "step": 20840
+    },
+    {
+      "epoch": 2.694146530559504,
+      "grad_norm": 1.5256946086883545,
+      "learning_rate": 1.2713883155613144e-06,
+      "loss": 0.8317,
+      "num_input_tokens_seen": 170803200,
+      "step": 20850
+    },
+    {
+      "epoch": 2.6954386871688847,
+      "grad_norm": 1.048254370689392,
+      "learning_rate": 1.2607594221060975e-06,
+      "loss": 0.6523,
+      "num_input_tokens_seen": 170885120,
+      "step": 20860
+    },
+    {
+      "epoch": 2.696730843778266,
+      "grad_norm": 0.8906717896461487,
+      "learning_rate": 1.2501739951135155e-06,
+      "loss": 0.8824,
+      "num_input_tokens_seen": 170967040,
+      "step": 20870
+    },
+    {
+      "epoch": 2.698023000387647,
+      "grad_norm": 0.6903151869773865,
+      "learning_rate": 1.2396320539654366e-06,
+      "loss": 0.8567,
+      "num_input_tokens_seen": 171048960,
+      "step": 20880
+    },
+    {
+      "epoch": 2.699315156997028,
+      "grad_norm": 1.00175940990448,
+      "learning_rate": 1.229133617964126e-06,
+      "loss": 1.1474,
+      "num_input_tokens_seen": 171130880,
+      "step": 20890
+    },
+    {
+      "epoch": 2.700607313606409,
+      "grad_norm": 0.7980899810791016,
+      "learning_rate": 1.2186787063321743e-06,
+      "loss": 0.5778,
+      "num_input_tokens_seen": 171212800,
+      "step": 20900
+    },
+    {
+      "epoch": 2.70189947021579,
+      "grad_norm": 0.848283052444458,
+      "learning_rate": 1.208267338212493e-06,
+      "loss": 0.9957,
+      "num_input_tokens_seen": 171294720,
+      "step": 20910
+    },
+    {
+      "epoch": 2.7031916268251712,
+      "grad_norm": 0.7135230898857117,
+      "learning_rate": 1.1978995326682535e-06,
+      "loss": 0.8154,
+      "num_input_tokens_seen": 171376640,
+      "step": 20920
+    },
+    {
+      "epoch": 2.7044837834345525,
+      "grad_norm": 0.7431100010871887,
+      "learning_rate": 1.1875753086828727e-06,
+      "loss": 0.3983,
+      "num_input_tokens_seen": 171458560,
+      "step": 20930
+    },
+    {
+      "epoch": 2.7057759400439334,
+      "grad_norm": 0.3723915219306946,
+      "learning_rate": 1.177294685159963e-06,
+      "loss": 0.8132,
+      "num_input_tokens_seen": 171540480,
+      "step": 20940
+    },
+    {
+      "epoch": 2.7070680966533143,
+      "grad_norm": 0.5851907134056091,
+      "learning_rate": 1.167057680923317e-06,
+      "loss": 0.6962,
+      "num_input_tokens_seen": 171622400,
+      "step": 20950
+    },
+    {
+      "epoch": 2.7083602532626956,
+      "grad_norm": 0.5370262861251831,
+      "learning_rate": 1.1568643147168434e-06,
+      "loss": 0.5466,
+      "num_input_tokens_seen": 171704320,
+      "step": 20960
+    },
+    {
+      "epoch": 2.7096524098720765,
+      "grad_norm": 0.6362653374671936,
+      "learning_rate": 1.1467146052045603e-06,
+      "loss": 0.9384,
+      "num_input_tokens_seen": 171786240,
+      "step": 20970
+    },
+    {
+      "epoch": 2.7109445664814578,
+      "grad_norm": 1.7740426063537598,
+      "learning_rate": 1.1366085709705515e-06,
+      "loss": 0.3847,
+      "num_input_tokens_seen": 171868160,
+      "step": 20980
+    },
+    {
+      "epoch": 2.7122367230908386,
+      "grad_norm": 0.7024763226509094,
+      "learning_rate": 1.1265462305189268e-06,
+      "loss": 0.7502,
+      "num_input_tokens_seen": 171950080,
+      "step": 20990
+    },
+    {
+      "epoch": 2.7135288797002195,
+      "grad_norm": 0.7445507645606995,
+      "learning_rate": 1.1165276022737926e-06,
+      "loss": 0.8216,
+      "num_input_tokens_seen": 172032000,
+      "step": 21000
+    },
+    {
+      "epoch": 2.714821036309601,
+      "grad_norm": 0.6791986227035522,
+      "learning_rate": 1.1065527045792251e-06,
+      "loss": 0.6687,
+      "num_input_tokens_seen": 172113920,
+      "step": 21010
+    },
+    {
+      "epoch": 2.7161131929189817,
+      "grad_norm": 0.999146580696106,
+      "learning_rate": 1.0966215556992231e-06,
+      "loss": 1.0147,
+      "num_input_tokens_seen": 172195840,
+      "step": 21020
+    },
+    {
+      "epoch": 2.717405349528363,
+      "grad_norm": 1.1012611389160156,
+      "learning_rate": 1.0867341738176857e-06,
+      "loss": 0.9388,
+      "num_input_tokens_seen": 172277760,
+      "step": 21030
+    },
+    {
+      "epoch": 2.718697506137744,
+      "grad_norm": 0.8878210783004761,
+      "learning_rate": 1.076890577038367e-06,
+      "loss": 0.9003,
+      "num_input_tokens_seen": 172359680,
+      "step": 21040
+    },
+    {
+      "epoch": 2.7199896627471247,
+      "grad_norm": 0.6756826043128967,
+      "learning_rate": 1.0670907833848664e-06,
+      "loss": 0.6774,
+      "num_input_tokens_seen": 172441600,
+      "step": 21050
+    },
+    {
+      "epoch": 2.721281819356506,
+      "grad_norm": 0.6432533860206604,
+      "learning_rate": 1.0573348108005614e-06,
+      "loss": 0.9375,
+      "num_input_tokens_seen": 172523520,
+      "step": 21060
+    },
+    {
+      "epoch": 2.722573975965887,
+      "grad_norm": 0.5785542726516724,
+      "learning_rate": 1.0476226771486074e-06,
+      "loss": 0.5949,
+      "num_input_tokens_seen": 172605440,
+      "step": 21070
+    },
+    {
+      "epoch": 2.723866132575268,
+      "grad_norm": 0.6814249753952026,
+      "learning_rate": 1.0379544002118824e-06,
+      "loss": 0.5954,
+      "num_input_tokens_seen": 172687360,
+      "step": 21080
+    },
+    {
+      "epoch": 2.725158289184649,
+      "grad_norm": 0.6625814437866211,
+      "learning_rate": 1.0283299976929672e-06,
+      "loss": 0.4849,
+      "num_input_tokens_seen": 172769280,
+      "step": 21090
+    },
+    {
+      "epoch": 2.7264504457940304,
+      "grad_norm": 0.4581983983516693,
+      "learning_rate": 1.0187494872141102e-06,
+      "loss": 0.6476,
+      "num_input_tokens_seen": 172851200,
+      "step": 21100
+    },
+    {
+      "epoch": 2.7277426024034113,
+      "grad_norm": 0.617347240447998,
+      "learning_rate": 1.0092128863171846e-06,
+      "loss": 1.1527,
+      "num_input_tokens_seen": 172933120,
+      "step": 21110
+    },
+    {
+      "epoch": 2.7290347590127926,
+      "grad_norm": 0.6780828833580017,
+      "learning_rate": 9.997202124636785e-07,
+      "loss": 0.7512,
+      "num_input_tokens_seen": 173015040,
+      "step": 21120
+    },
+    {
+      "epoch": 2.7303269156221734,
+      "grad_norm": 1.0733816623687744,
+      "learning_rate": 9.902714830346437e-07,
+      "loss": 0.6027,
+      "num_input_tokens_seen": 173096960,
+      "step": 21130
+    },
+    {
+      "epoch": 2.7316190722315543,
+      "grad_norm": 0.6169953942298889,
+      "learning_rate": 9.808667153306612e-07,
+      "loss": 0.5431,
+      "num_input_tokens_seen": 173178880,
+      "step": 21140
+    },
+    {
+      "epoch": 2.7329112288409356,
+      "grad_norm": 0.3251383602619171,
+      "learning_rate": 9.715059265718335e-07,
+      "loss": 0.7888,
+      "num_input_tokens_seen": 173260800,
+      "step": 21150
+    },
+    {
+      "epoch": 2.7342033854503165,
+      "grad_norm": 0.7879831194877625,
+      "learning_rate": 9.62189133897723e-07,
+      "loss": 0.9271,
+      "num_input_tokens_seen": 173342720,
+      "step": 21160
+    },
+    {
+      "epoch": 2.735495542059698,
+      "grad_norm": 0.988198459148407,
+      "learning_rate": 9.52916354367353e-07,
+      "loss": 0.8073,
+      "num_input_tokens_seen": 173424640,
+      "step": 21170
+    },
+    {
+      "epoch": 2.7367876986690787,
+      "grad_norm": 0.5871654152870178,
+      "learning_rate": 9.436876049591398e-07,
+      "loss": 1.2963,
+      "num_input_tokens_seen": 173506560,
+      "step": 21180
+    },
+    {
+      "epoch": 2.7380798552784595,
+      "grad_norm": 0.9331020712852478,
+      "learning_rate": 9.345029025708995e-07,
+      "loss": 0.6766,
+      "num_input_tokens_seen": 173588480,
+      "step": 21190
+    },
+    {
+      "epoch": 2.739372011887841,
+      "grad_norm": 1.031977891921997,
+      "learning_rate": 9.253622640197773e-07,
+      "loss": 0.5872,
+      "num_input_tokens_seen": 173670400,
+      "step": 21200
+    },
+    {
+      "epoch": 2.7406641684972217,
+      "grad_norm": 0.28784510493278503,
+      "learning_rate": 9.162657060422574e-07,
+      "loss": 0.5351,
+      "num_input_tokens_seen": 173752320,
+      "step": 21210
+    },
+    {
+      "epoch": 2.741956325106603,
+      "grad_norm": 0.7928471565246582,
+      "learning_rate": 9.072132452941002e-07,
+      "loss": 0.6825,
+      "num_input_tokens_seen": 173834240,
+      "step": 21220
+    },
+    {
+      "epoch": 2.743248481715984,
+      "grad_norm": 0.6947720646858215,
+      "learning_rate": 8.982048983503271e-07,
+      "loss": 0.84,
+      "num_input_tokens_seen": 173916160,
+      "step": 21230
+    },
+    {
+      "epoch": 2.7445406383253648,
+      "grad_norm": 0.6338273882865906,
+      "learning_rate": 8.892406817051946e-07,
+      "loss": 0.4665,
+      "num_input_tokens_seen": 173998080,
+      "step": 21240
+    },
+    {
+      "epoch": 2.745832794934746,
+      "grad_norm": 0.6995043754577637,
+      "learning_rate": 8.803206117721424e-07,
+      "loss": 0.8926,
+      "num_input_tokens_seen": 174080000,
+      "step": 21250
+    },
+    {
+      "epoch": 2.7471249515441274,
+      "grad_norm": 0.9222922325134277,
+      "learning_rate": 8.714447048837948e-07,
+      "loss": 0.8874,
+      "num_input_tokens_seen": 174161920,
+      "step": 21260
+    },
+    {
+      "epoch": 2.7484171081535083,
+      "grad_norm": 1.1156824827194214,
+      "learning_rate": 8.626129772918962e-07,
+      "loss": 0.9445,
+      "num_input_tokens_seen": 174243840,
+      "step": 21270
+    },
+    {
+      "epoch": 2.749709264762889,
+      "grad_norm": 0.2606189250946045,
+      "learning_rate": 8.538254451673138e-07,
+      "loss": 0.7208,
+      "num_input_tokens_seen": 174325760,
+      "step": 21280
+    },
+    {
+      "epoch": 2.7510014213722704,
+      "grad_norm": 1.007570743560791,
+      "learning_rate": 8.450821245999829e-07,
+      "loss": 0.6216,
+      "num_input_tokens_seen": 174407680,
+      "step": 21290
+    },
+    {
+      "epoch": 2.7522935779816513,
+      "grad_norm": 1.0008625984191895,
+      "learning_rate": 8.363830315988947e-07,
+      "loss": 0.9531,
+      "num_input_tokens_seen": 174489600,
+      "step": 21300
+    },
+    {
+      "epoch": 2.7535857345910326,
+      "grad_norm": 0.3577563166618347,
+      "learning_rate": 8.277281820920523e-07,
+      "loss": 0.7485,
+      "num_input_tokens_seen": 174571520,
+      "step": 21310
+    },
+    {
+      "epoch": 2.7548778912004135,
+      "grad_norm": 0.6377021074295044,
+      "learning_rate": 8.191175919264604e-07,
+      "loss": 1.0028,
+      "num_input_tokens_seen": 174653440,
+      "step": 21320
+    },
+    {
+      "epoch": 2.7561700478097944,
+      "grad_norm": 1.360044002532959,
+      "learning_rate": 8.105512768680712e-07,
+      "loss": 0.7594,
+      "num_input_tokens_seen": 174735360,
+      "step": 21330
+    },
+    {
+      "epoch": 2.7574622044191757,
+      "grad_norm": 0.5991158485412598,
+      "learning_rate": 8.02029252601777e-07,
+      "loss": 0.7408,
+      "num_input_tokens_seen": 174817280,
+      "step": 21340
+    },
+    {
+      "epoch": 2.7587543610285565,
+      "grad_norm": 0.593151330947876,
+      "learning_rate": 7.935515347313793e-07,
+      "loss": 0.845,
+      "num_input_tokens_seen": 174899200,
+      "step": 21350
+    },
+    {
+      "epoch": 2.760046517637938,
+      "grad_norm": 0.6272279620170593,
+      "learning_rate": 7.851181387795392e-07,
+      "loss": 0.9223,
+      "num_input_tokens_seen": 174981120,
+      "step": 21360
+    },
+    {
+      "epoch": 2.7613386742473187,
+      "grad_norm": 0.7071083188056946,
+      "learning_rate": 7.767290801877796e-07,
+      "loss": 0.5397,
+      "num_input_tokens_seen": 175063040,
+      "step": 21370
+    },
+    {
+      "epoch": 2.7626308308566996,
+      "grad_norm": 0.23855863511562347,
+      "learning_rate": 7.683843743164359e-07,
+      "loss": 0.7537,
+      "num_input_tokens_seen": 175144960,
+      "step": 21380
+    },
+    {
+      "epoch": 2.763922987466081,
+      "grad_norm": 0.9484770894050598,
+      "learning_rate": 7.600840364446333e-07,
+      "loss": 0.7232,
+      "num_input_tokens_seen": 175226880,
+      "step": 21390
+    },
+    {
+      "epoch": 2.765215144075462,
+      "grad_norm": 0.7923175096511841,
+      "learning_rate": 7.518280817702616e-07,
+      "loss": 0.683,
+      "num_input_tokens_seen": 175308800,
+      "step": 21400
+    },
+    {
+      "epoch": 2.766507300684843,
+      "grad_norm": 1.019843578338623,
+      "learning_rate": 7.436165254099376e-07,
+      "loss": 0.9568,
+      "num_input_tokens_seen": 175390720,
+      "step": 21410
+    },
+    {
+      "epoch": 2.767799457294224,
+      "grad_norm": 0.6530530452728271,
+      "learning_rate": 7.354493823990006e-07,
+      "loss": 0.8146,
+      "num_input_tokens_seen": 175472640,
+      "step": 21420
+    },
+    {
+      "epoch": 2.7690916139036053,
+      "grad_norm": 0.8259232044219971,
+      "learning_rate": 7.273266676914498e-07,
+      "loss": 0.8714,
+      "num_input_tokens_seen": 175554560,
+      "step": 21430
+    },
+    {
+      "epoch": 2.770383770512986,
+      "grad_norm": 0.8644856214523315,
+      "learning_rate": 7.19248396159955e-07,
+      "loss": 0.621,
+      "num_input_tokens_seen": 175636480,
+      "step": 21440
+    },
+    {
+      "epoch": 2.7716759271223674,
+      "grad_norm": 0.7736506462097168,
+      "learning_rate": 7.112145825957927e-07,
+      "loss": 0.8045,
+      "num_input_tokens_seen": 175718400,
+      "step": 21450
+    },
+    {
+      "epoch": 2.7729680837317483,
+      "grad_norm": 0.886471152305603,
+      "learning_rate": 7.03225241708852e-07,
+      "loss": 1.0378,
+      "num_input_tokens_seen": 175800320,
+      "step": 21460
+    },
+    {
+      "epoch": 2.774260240341129,
+      "grad_norm": 0.7383142113685608,
+      "learning_rate": 6.952803881275894e-07,
+      "loss": 0.5995,
+      "num_input_tokens_seen": 175882240,
+      "step": 21470
+    },
+    {
+      "epoch": 2.7755523969505105,
+      "grad_norm": 1.1481050252914429,
+      "learning_rate": 6.873800363989935e-07,
+      "loss": 0.9361,
+      "num_input_tokens_seen": 175964160,
+      "step": 21480
+    },
+    {
+      "epoch": 2.7768445535598913,
+      "grad_norm": 1.066712737083435,
+      "learning_rate": 6.795242009885905e-07,
+      "loss": 0.7132,
+      "num_input_tokens_seen": 176046080,
+      "step": 21490
+    },
+    {
+      "epoch": 2.7781367101692727,
+      "grad_norm": 0.7255296111106873,
+      "learning_rate": 6.717128962803798e-07,
+      "loss": 0.9569,
+      "num_input_tokens_seen": 176128000,
+      "step": 21500
+    },
+    {
+      "epoch": 2.7794288667786535,
+      "grad_norm": 0.3221868574619293,
+      "learning_rate": 6.63946136576829e-07,
+      "loss": 0.5613,
+      "num_input_tokens_seen": 176209920,
+      "step": 21510
+    },
+    {
+      "epoch": 2.7807210233880344,
+      "grad_norm": 0.8177877068519592,
+      "learning_rate": 6.562239360988542e-07,
+      "loss": 0.8761,
+      "num_input_tokens_seen": 176291840,
+      "step": 21520
+    },
+    {
+      "epoch": 2.7820131799974157,
+      "grad_norm": 0.910959780216217,
+      "learning_rate": 6.485463089857674e-07,
+      "loss": 0.5046,
+      "num_input_tokens_seen": 176373760,
+      "step": 21530
+    },
+    {
+      "epoch": 2.783305336606797,
+      "grad_norm": 1.2782275676727295,
+      "learning_rate": 6.409132692952874e-07,
+      "loss": 0.8931,
+      "num_input_tokens_seen": 176455680,
+      "step": 21540
+    },
+    {
+      "epoch": 2.784597493216178,
+      "grad_norm": 0.6053476333618164,
+      "learning_rate": 6.333248310034706e-07,
+      "loss": 0.9367,
+      "num_input_tokens_seen": 176537600,
+      "step": 21550
+    },
+    {
+      "epoch": 2.7858896498255588,
+      "grad_norm": 0.9305321574211121,
+      "learning_rate": 6.257810080047249e-07,
+      "loss": 1.0246,
+      "num_input_tokens_seen": 176619520,
+      "step": 21560
+    },
+    {
+      "epoch": 2.78718180643494,
+      "grad_norm": 1.467142105102539,
+      "learning_rate": 6.182818141117625e-07,
+      "loss": 0.7575,
+      "num_input_tokens_seen": 176701440,
+      "step": 21570
+    },
+    {
+      "epoch": 2.788473963044321,
+      "grad_norm": 0.3563095033168793,
+      "learning_rate": 6.1082726305558e-07,
+      "loss": 0.8363,
+      "num_input_tokens_seen": 176783360,
+      "step": 21580
+    },
+    {
+      "epoch": 2.7897661196537022,
+      "grad_norm": 0.7321064472198486,
+      "learning_rate": 6.034173684854316e-07,
+      "loss": 0.8788,
+      "num_input_tokens_seen": 176865280,
+      "step": 21590
+    },
+    {
+      "epoch": 2.791058276263083,
+      "grad_norm": 0.5982012152671814,
+      "learning_rate": 5.960521439688088e-07,
+      "loss": 0.7845,
+      "num_input_tokens_seen": 176947200,
+      "step": 21600
+    },
+    {
+      "epoch": 2.792350432872464,
+      "grad_norm": 0.965398371219635,
+      "learning_rate": 5.88731602991413e-07,
+      "loss": 0.8485,
+      "num_input_tokens_seen": 177029120,
+      "step": 21610
+    },
+    {
+      "epoch": 2.7936425894818453,
+      "grad_norm": 0.33368125557899475,
+      "learning_rate": 5.814557589571223e-07,
+      "loss": 0.9114,
+      "num_input_tokens_seen": 177111040,
+      "step": 21620
+    },
+    {
+      "epoch": 2.794934746091226,
+      "grad_norm": 1.154457688331604,
+      "learning_rate": 5.742246251879829e-07,
+      "loss": 0.6277,
+      "num_input_tokens_seen": 177192960,
+      "step": 21630
+    },
+    {
+      "epoch": 2.7962269027006075,
+      "grad_norm": 0.728986382484436,
+      "learning_rate": 5.67038214924176e-07,
+      "loss": 0.8553,
+      "num_input_tokens_seen": 177274880,
+      "step": 21640
+    },
+    {
+      "epoch": 2.7975190593099883,
+      "grad_norm": 0.7881826758384705,
+      "learning_rate": 5.598965413239926e-07,
+      "loss": 0.7787,
+      "num_input_tokens_seen": 177356800,
+      "step": 21650
+    },
+    {
+      "epoch": 2.798811215919369,
+      "grad_norm": 0.42167970538139343,
+      "learning_rate": 5.527996174638061e-07,
+      "loss": 0.5017,
+      "num_input_tokens_seen": 177438720,
+      "step": 21660
+    },
+    {
+      "epoch": 2.8001033725287505,
+      "grad_norm": 0.8099974393844604,
+      "learning_rate": 5.457474563380638e-07,
+      "loss": 1.0671,
+      "num_input_tokens_seen": 177520640,
+      "step": 21670
+    },
+    {
+      "epoch": 2.8013955291381314,
+      "grad_norm": 0.619404673576355,
+      "learning_rate": 5.387400708592422e-07,
+      "loss": 1.0223,
+      "num_input_tokens_seen": 177602560,
+      "step": 21680
+    },
+    {
+      "epoch": 2.8026876857475127,
+      "grad_norm": 1.066259741783142,
+      "learning_rate": 5.317774738578446e-07,
+      "loss": 0.7952,
+      "num_input_tokens_seen": 177684480,
+      "step": 21690
+    },
+    {
+      "epoch": 2.8039798423568936,
+      "grad_norm": 0.8455848097801208,
+      "learning_rate": 5.248596780823567e-07,
+      "loss": 0.5789,
+      "num_input_tokens_seen": 177766400,
+      "step": 21700
+    },
+    {
+      "epoch": 2.8052719989662744,
+      "grad_norm": 0.40524500608444214,
+      "learning_rate": 5.179866961992353e-07,
+      "loss": 0.9732,
+      "num_input_tokens_seen": 177848320,
+      "step": 21710
+    },
+    {
+      "epoch": 2.8065641555756558,
+      "grad_norm": 0.7112755179405212,
+      "learning_rate": 5.111585407928887e-07,
+      "loss": 0.6634,
+      "num_input_tokens_seen": 177930240,
+      "step": 21720
+    },
+    {
+      "epoch": 2.807856312185037,
+      "grad_norm": 0.7933624386787415,
+      "learning_rate": 5.043752243656414e-07,
+      "loss": 0.6454,
+      "num_input_tokens_seen": 178012160,
+      "step": 21730
+    },
+    {
+      "epoch": 2.809148468794418,
+      "grad_norm": 0.6080639362335205,
+      "learning_rate": 4.976367593377218e-07,
+      "loss": 0.8702,
+      "num_input_tokens_seen": 178094080,
+      "step": 21740
+    },
+    {
+      "epoch": 2.810440625403799,
+      "grad_norm": 0.7494640946388245,
+      "learning_rate": 4.909431580472385e-07,
+      "loss": 1.0812,
+      "num_input_tokens_seen": 178176000,
+      "step": 21750
+    },
+    {
+      "epoch": 2.81173278201318,
+      "grad_norm": 0.49116799235343933,
+      "learning_rate": 4.842944327501458e-07,
+      "loss": 0.9232,
+      "num_input_tokens_seen": 178257920,
+      "step": 21760
+    },
+    {
+      "epoch": 2.813024938622561,
+      "grad_norm": 0.8744192719459534,
+      "learning_rate": 4.776905956202393e-07,
+      "loss": 0.7583,
+      "num_input_tokens_seen": 178339840,
+      "step": 21770
+    },
+    {
+      "epoch": 2.8143170952319423,
+      "grad_norm": 0.6495558619499207,
+      "learning_rate": 4.711316587491188e-07,
+      "loss": 0.626,
+      "num_input_tokens_seen": 178421760,
+      "step": 21780
+    },
+    {
+      "epoch": 2.815609251841323,
+      "grad_norm": 0.5544187426567078,
+      "learning_rate": 4.646176341461722e-07,
+      "loss": 0.7818,
+      "num_input_tokens_seen": 178503680,
+      "step": 21790
+    },
+    {
+      "epoch": 2.816901408450704,
+      "grad_norm": 0.26983362436294556,
+      "learning_rate": 4.581485337385588e-07,
+      "loss": 0.7148,
+      "num_input_tokens_seen": 178585600,
+      "step": 21800
+    },
+    {
+      "epoch": 2.8181935650600853,
+      "grad_norm": 0.6384357213973999,
+      "learning_rate": 4.5172436937117036e-07,
+      "loss": 0.8498,
+      "num_input_tokens_seen": 178667520,
+      "step": 21810
+    },
+    {
+      "epoch": 2.819485721669466,
+      "grad_norm": 1.1162294149398804,
+      "learning_rate": 4.4534515280663937e-07,
+      "loss": 1.0099,
+      "num_input_tokens_seen": 178749440,
+      "step": 21820
+    },
+    {
+      "epoch": 2.8207778782788475,
+      "grad_norm": 1.0189695358276367,
+      "learning_rate": 4.390108957252781e-07,
+      "loss": 0.878,
+      "num_input_tokens_seen": 178831360,
+      "step": 21830
+    },
+    {
+      "epoch": 2.8220700348882284,
+      "grad_norm": 0.7732197046279907,
+      "learning_rate": 4.3272160972509524e-07,
+      "loss": 0.7988,
+      "num_input_tokens_seen": 178913280,
+      "step": 21840
+    },
+    {
+      "epoch": 2.8233621914976093,
+      "grad_norm": 0.9585608839988708,
+      "learning_rate": 4.264773063217431e-07,
+      "loss": 1.0454,
+      "num_input_tokens_seen": 178995200,
+      "step": 21850
+    },
+    {
+      "epoch": 2.8246543481069906,
+      "grad_norm": 0.694817841053009,
+      "learning_rate": 4.20277996948526e-07,
+      "loss": 0.9275,
+      "num_input_tokens_seen": 179077120,
+      "step": 21860
+    },
+    {
+      "epoch": 2.825946504716372,
+      "grad_norm": 0.8328717350959778,
+      "learning_rate": 4.1412369295635023e-07,
+      "loss": 0.95,
+      "num_input_tokens_seen": 179159040,
+      "step": 21870
+    },
+    {
+      "epoch": 2.8272386613257527,
+      "grad_norm": 0.9796121716499329,
+      "learning_rate": 4.0801440561372694e-07,
+      "loss": 0.8007,
+      "num_input_tokens_seen": 179240960,
+      "step": 21880
+    },
+    {
+      "epoch": 2.8285308179351336,
+      "grad_norm": 1.0329798460006714,
+      "learning_rate": 4.0195014610674153e-07,
+      "loss": 0.8786,
+      "num_input_tokens_seen": 179322880,
+      "step": 21890
+    },
+    {
+      "epoch": 2.829822974544515,
+      "grad_norm": 0.4001299738883972,
+      "learning_rate": 3.9593092553902587e-07,
+      "loss": 0.8457,
+      "num_input_tokens_seen": 179404800,
+      "step": 21900
+    },
+    {
+      "epoch": 2.831115131153896,
+      "grad_norm": 0.5981205701828003,
+      "learning_rate": 3.899567549317529e-07,
+      "loss": 0.7406,
+      "num_input_tokens_seen": 179486720,
+      "step": 21910
+    },
+    {
+      "epoch": 2.832407287763277,
+      "grad_norm": 0.5832979679107666,
+      "learning_rate": 3.840276452236058e-07,
+      "loss": 0.7223,
+      "num_input_tokens_seen": 179568640,
+      "step": 21920
+    },
+    {
+      "epoch": 2.833699444372658,
+      "grad_norm": 0.6982936263084412,
+      "learning_rate": 3.7814360727076724e-07,
+      "loss": 0.8734,
+      "num_input_tokens_seen": 179650560,
+      "step": 21930
+    },
+    {
+      "epoch": 2.834991600982039,
+      "grad_norm": 0.81357342004776,
+      "learning_rate": 3.723046518468859e-07,
+      "loss": 0.5382,
+      "num_input_tokens_seen": 179732480,
+      "step": 21940
+    },
+    {
+      "epoch": 2.83628375759142,
+      "grad_norm": 0.6205422282218933,
+      "learning_rate": 3.6651078964306807e-07,
+      "loss": 1.0019,
+      "num_input_tokens_seen": 179814400,
+      "step": 21950
+    },
+    {
+      "epoch": 2.837575914200801,
+      "grad_norm": 0.670599102973938,
+      "learning_rate": 3.607620312678528e-07,
+      "loss": 0.7448,
+      "num_input_tokens_seen": 179896320,
+      "step": 21960
+    },
+    {
+      "epoch": 2.8388680708101823,
+      "grad_norm": 1.0795679092407227,
+      "learning_rate": 3.550583872471952e-07,
+      "loss": 0.553,
+      "num_input_tokens_seen": 179978240,
+      "step": 21970
+    },
+    {
+      "epoch": 2.840160227419563,
+      "grad_norm": 0.7728472352027893,
+      "learning_rate": 3.4939986802445256e-07,
+      "loss": 0.9766,
+      "num_input_tokens_seen": 180060160,
+      "step": 21980
+    },
+    {
+      "epoch": 2.841452384028944,
+      "grad_norm": 0.24065038561820984,
+      "learning_rate": 3.437864839603455e-07,
+      "loss": 0.3731,
+      "num_input_tokens_seen": 180142080,
+      "step": 21990
+    },
+    {
+      "epoch": 2.8427445406383254,
+      "grad_norm": 1.1644612550735474,
+      "learning_rate": 3.3821824533296633e-07,
+      "loss": 0.5934,
+      "num_input_tokens_seen": 180224000,
+      "step": 22000
+    },
+    {
+      "epoch": 2.8440366972477067,
+      "grad_norm": 0.8106747269630432,
+      "learning_rate": 3.3269516233773446e-07,
+      "loss": 0.7189,
+      "num_input_tokens_seen": 180305920,
+      "step": 22010
+    },
+    {
+      "epoch": 2.8453288538570876,
+      "grad_norm": 0.24490030109882355,
+      "learning_rate": 3.272172450873967e-07,
+      "loss": 0.6045,
+      "num_input_tokens_seen": 180387840,
+      "step": 22020
+    },
+    {
+      "epoch": 2.8466210104664684,
+      "grad_norm": 1.151792287826538,
+      "learning_rate": 3.217845036119993e-07,
+      "loss": 0.4563,
+      "num_input_tokens_seen": 180469760,
+      "step": 22030
+    },
+    {
+      "epoch": 2.8479131670758497,
+      "grad_norm": 0.4980817139148712,
+      "learning_rate": 3.163969478588713e-07,
+      "loss": 0.7337,
+      "num_input_tokens_seen": 180551680,
+      "step": 22040
+    },
+    {
+      "epoch": 2.8492053236852306,
+      "grad_norm": 0.7074184417724609,
+      "learning_rate": 3.11054587692608e-07,
+      "loss": 0.8102,
+      "num_input_tokens_seen": 180633600,
+      "step": 22050
+    },
+    {
+      "epoch": 2.850497480294612,
+      "grad_norm": 0.7132773995399475,
+      "learning_rate": 3.057574328950541e-07,
+      "loss": 0.841,
+      "num_input_tokens_seen": 180715520,
+      "step": 22060
+    },
+    {
+      "epoch": 2.851789636903993,
+      "grad_norm": 1.5704045295715332,
+      "learning_rate": 3.005054931652762e-07,
+      "loss": 0.5297,
+      "num_input_tokens_seen": 180797440,
+      "step": 22070
+    },
+    {
+      "epoch": 2.8530817935133737,
+      "grad_norm": 1.3629149198532104,
+      "learning_rate": 2.952987781195599e-07,
+      "loss": 0.7995,
+      "num_input_tokens_seen": 180879360,
+      "step": 22080
+    },
+    {
+      "epoch": 2.854373950122755,
+      "grad_norm": 0.6526079773902893,
+      "learning_rate": 2.901372972913791e-07,
+      "loss": 0.8583,
+      "num_input_tokens_seen": 180961280,
+      "step": 22090
+    },
+    {
+      "epoch": 2.855666106732136,
+      "grad_norm": 0.8587889075279236,
+      "learning_rate": 2.8502106013138516e-07,
+      "loss": 0.8178,
+      "num_input_tokens_seen": 181043200,
+      "step": 22100
+    },
+    {
+      "epoch": 2.856958263341517,
+      "grad_norm": 0.6336101293563843,
+      "learning_rate": 2.799500760073931e-07,
+      "loss": 0.7179,
+      "num_input_tokens_seen": 181125120,
+      "step": 22110
+    },
+    {
+      "epoch": 2.858250419950898,
+      "grad_norm": 0.9315862059593201,
+      "learning_rate": 2.749243542043561e-07,
+      "loss": 0.4239,
+      "num_input_tokens_seen": 181207040,
+      "step": 22120
+    },
+    {
+      "epoch": 2.859542576560279,
+      "grad_norm": 0.7689865827560425,
+      "learning_rate": 2.699439039243523e-07,
+      "loss": 0.88,
+      "num_input_tokens_seen": 181288960,
+      "step": 22130
+    },
+    {
+      "epoch": 2.86083473316966,
+      "grad_norm": 0.37839198112487793,
+      "learning_rate": 2.6500873428656483e-07,
+      "loss": 0.4962,
+      "num_input_tokens_seen": 181370880,
+      "step": 22140
+    },
+    {
+      "epoch": 2.862126889779041,
+      "grad_norm": 0.6896770596504211,
+      "learning_rate": 2.601188543272737e-07,
+      "loss": 0.7783,
+      "num_input_tokens_seen": 181452800,
+      "step": 22150
+    },
+    {
+      "epoch": 2.8634190463884224,
+      "grad_norm": 0.639472484588623,
+      "learning_rate": 2.552742729998309e-07,
+      "loss": 1.1341,
+      "num_input_tokens_seen": 181534720,
+      "step": 22160
+    },
+    {
+      "epoch": 2.8647112029978032,
+      "grad_norm": 0.6958195567131042,
+      "learning_rate": 2.5047499917464636e-07,
+      "loss": 0.6448,
+      "num_input_tokens_seen": 181616640,
+      "step": 22170
+    },
+    {
+      "epoch": 2.8660033596071846,
+      "grad_norm": 0.9285911321640015,
+      "learning_rate": 2.457210416391742e-07,
+      "loss": 0.7003,
+      "num_input_tokens_seen": 181698560,
+      "step": 22180
+    },
+    {
+      "epoch": 2.8672955162165654,
+      "grad_norm": 1.0297704935073853,
+      "learning_rate": 2.4101240909789325e-07,
+      "loss": 0.9887,
+      "num_input_tokens_seen": 181780480,
+      "step": 22190
+    },
+    {
+      "epoch": 2.8685876728259467,
+      "grad_norm": 0.7845831513404846,
+      "learning_rate": 2.3634911017229034e-07,
+      "loss": 0.6259,
+      "num_input_tokens_seen": 181862400,
+      "step": 22200
+    },
+    {
+      "epoch": 2.8698798294353276,
+      "grad_norm": 0.7428514361381531,
+      "learning_rate": 2.3173115340085204e-07,
+      "loss": 0.9638,
+      "num_input_tokens_seen": 181944320,
+      "step": 22210
+    },
+    {
+      "epoch": 2.8711719860447085,
+      "grad_norm": 0.7519323825836182,
+      "learning_rate": 2.2715854723903974e-07,
+      "loss": 0.9015,
+      "num_input_tokens_seen": 182026240,
+      "step": 22220
+    },
+    {
+      "epoch": 2.87246414265409,
+      "grad_norm": 1.0879383087158203,
+      "learning_rate": 2.2263130005927558e-07,
+      "loss": 0.7179,
+      "num_input_tokens_seen": 182108160,
+      "step": 22230
+    },
+    {
+      "epoch": 2.8737562992634706,
+      "grad_norm": 0.7938571572303772,
+      "learning_rate": 2.181494201509343e-07,
+      "loss": 0.936,
+      "num_input_tokens_seen": 182190080,
+      "step": 22240
+    },
+    {
+      "epoch": 2.875048455872852,
+      "grad_norm": 0.4322587549686432,
+      "learning_rate": 2.1371291572032382e-07,
+      "loss": 0.5952,
+      "num_input_tokens_seen": 182272000,
+      "step": 22250
+    },
+    {
+      "epoch": 2.876340612482233,
+      "grad_norm": 0.47111037373542786,
+      "learning_rate": 2.0932179489066006e-07,
+      "loss": 0.7432,
+      "num_input_tokens_seen": 182353920,
+      "step": 22260
+    },
+    {
+      "epoch": 2.8776327690916137,
+      "grad_norm": 0.7983747720718384,
+      "learning_rate": 2.0497606570207829e-07,
+      "loss": 0.8684,
+      "num_input_tokens_seen": 182435840,
+      "step": 22270
+    },
+    {
+      "epoch": 2.878924925700995,
+      "grad_norm": 0.3726508915424347,
+      "learning_rate": 2.0067573611158853e-07,
+      "loss": 0.6558,
+      "num_input_tokens_seen": 182517760,
+      "step": 22280
+    },
+    {
+      "epoch": 2.880217082310376,
+      "grad_norm": 1.232807993888855,
+      "learning_rate": 1.9642081399307844e-07,
+      "loss": 0.562,
+      "num_input_tokens_seen": 182599680,
+      "step": 22290
+    },
+    {
+      "epoch": 2.881509238919757,
+      "grad_norm": 1.0021343231201172,
+      "learning_rate": 1.9221130713729663e-07,
+      "loss": 0.7792,
+      "num_input_tokens_seen": 182681600,
+      "step": 22300
+    },
+    {
+      "epoch": 2.882801395529138,
+      "grad_norm": 0.8979313969612122,
+      "learning_rate": 1.8804722325183044e-07,
+      "loss": 0.7825,
+      "num_input_tokens_seen": 182763520,
+      "step": 22310
+    },
+    {
+      "epoch": 2.884093552138519,
+      "grad_norm": 0.4306122660636902,
+      "learning_rate": 1.8392856996110875e-07,
+      "loss": 0.6558,
+      "num_input_tokens_seen": 182845440,
+      "step": 22320
+    },
+    {
+      "epoch": 2.8853857087479002,
+      "grad_norm": 0.7078767418861389,
+      "learning_rate": 1.7985535480636584e-07,
+      "loss": 0.6038,
+      "num_input_tokens_seen": 182927360,
+      "step": 22330
+    },
+    {
+      "epoch": 2.8866778653572815,
+      "grad_norm": 0.7182630300521851,
+      "learning_rate": 1.7582758524564425e-07,
+      "loss": 0.787,
+      "num_input_tokens_seen": 183009280,
+      "step": 22340
+    },
+    {
+      "epoch": 2.8879700219666624,
+      "grad_norm": 0.6843809485435486,
+      "learning_rate": 1.7184526865377805e-07,
+      "loss": 0.56,
+      "num_input_tokens_seen": 183091200,
+      "step": 22350
+    },
+    {
+      "epoch": 2.8892621785760433,
+      "grad_norm": 0.5071055293083191,
+      "learning_rate": 1.6790841232237064e-07,
+      "loss": 0.494,
+      "num_input_tokens_seen": 183173120,
+      "step": 22360
+    },
+    {
+      "epoch": 2.8905543351854246,
+      "grad_norm": 0.43069300055503845,
+      "learning_rate": 1.6401702345979485e-07,
+      "loss": 1.0883,
+      "num_input_tokens_seen": 183255040,
+      "step": 22370
+    },
+    {
+      "epoch": 2.8918464917948055,
+      "grad_norm": 0.9334409832954407,
+      "learning_rate": 1.6017110919116786e-07,
+      "loss": 0.7701,
+      "num_input_tokens_seen": 183336960,
+      "step": 22380
+    },
+    {
+      "epoch": 2.8931386484041868,
+      "grad_norm": 0.9663013815879822,
+      "learning_rate": 1.5637067655834282e-07,
+      "loss": 0.7901,
+      "num_input_tokens_seen": 183418880,
+      "step": 22390
+    },
+    {
+      "epoch": 2.8944308050135676,
+      "grad_norm": 0.45232853293418884,
+      "learning_rate": 1.526157325199007e-07,
+      "loss": 0.8333,
+      "num_input_tokens_seen": 183500800,
+      "step": 22400
+    },
+    {
+      "epoch": 2.8957229616229485,
+      "grad_norm": 1.2626268863677979,
+      "learning_rate": 1.4890628395113072e-07,
+      "loss": 0.3667,
+      "num_input_tokens_seen": 183582720,
+      "step": 22410
+    },
+    {
+      "epoch": 2.89701511823233,
+      "grad_norm": 1.018869400024414,
+      "learning_rate": 1.452423376440193e-07,
+      "loss": 1.0644,
+      "num_input_tokens_seen": 183664640,
+      "step": 22420
+    },
+    {
+      "epoch": 2.8983072748417107,
+      "grad_norm": 0.7966929078102112,
+      "learning_rate": 1.4162390030723617e-07,
+      "loss": 0.6229,
+      "num_input_tokens_seen": 183746560,
+      "step": 22430
+    },
+    {
+      "epoch": 2.899599431451092,
+      "grad_norm": 1.1223193407058716,
+      "learning_rate": 1.380509785661288e-07,
+      "loss": 0.8879,
+      "num_input_tokens_seen": 183828480,
+      "step": 22440
+    },
+    {
+      "epoch": 2.900891588060473,
+      "grad_norm": 0.751388430595398,
+      "learning_rate": 1.3452357896270308e-07,
+      "loss": 0.7747,
+      "num_input_tokens_seen": 183910400,
+      "step": 22450
+    },
+    {
+      "epoch": 2.9021837446698537,
+      "grad_norm": 0.7575998902320862,
+      "learning_rate": 1.3104170795561477e-07,
+      "loss": 0.7773,
+      "num_input_tokens_seen": 183992320,
+      "step": 22460
+    },
+    {
+      "epoch": 2.903475901279235,
+      "grad_norm": 0.6274828314781189,
+      "learning_rate": 1.2760537192015866e-07,
+      "loss": 0.7762,
+      "num_input_tokens_seen": 184074240,
+      "step": 22470
+    },
+    {
+      "epoch": 2.9047680578886164,
+      "grad_norm": 0.9277952909469604,
+      "learning_rate": 1.242145771482489e-07,
+      "loss": 0.7858,
+      "num_input_tokens_seen": 184156160,
+      "step": 22480
+    },
+    {
+      "epoch": 2.9060602144979972,
+      "grad_norm": 1.2103768587112427,
+      "learning_rate": 1.2086932984842758e-07,
+      "loss": 0.9757,
+      "num_input_tokens_seen": 184238080,
+      "step": 22490
+    },
+    {
+      "epoch": 2.907352371107378,
+      "grad_norm": 0.7629841566085815,
+      "learning_rate": 1.1756963614582006e-07,
+      "loss": 0.8254,
+      "num_input_tokens_seen": 184320000,
+      "step": 22500
+    },
+    {
+      "epoch": 2.9086445277167594,
+      "grad_norm": 0.32851213216781616,
+      "learning_rate": 1.1431550208215736e-07,
+      "loss": 0.2809,
+      "num_input_tokens_seen": 184401920,
+      "step": 22510
+    },
+    {
+      "epoch": 2.9099366843261403,
+      "grad_norm": 0.37616801261901855,
+      "learning_rate": 1.1110693361574831e-07,
+      "loss": 0.779,
+      "num_input_tokens_seen": 184483840,
+      "step": 22520
+    },
+    {
+      "epoch": 2.9112288409355216,
+      "grad_norm": 1.0593737363815308,
+      "learning_rate": 1.0794393662147129e-07,
+      "loss": 0.9263,
+      "num_input_tokens_seen": 184565760,
+      "step": 22530
+    },
+    {
+      "epoch": 2.9125209975449025,
+      "grad_norm": 0.390419602394104,
+      "learning_rate": 1.0482651689075751e-07,
+      "loss": 0.6628,
+      "num_input_tokens_seen": 184647680,
+      "step": 22540
+    },
+    {
+      "epoch": 2.9138131541542833,
+      "grad_norm": 0.5607829689979553,
+      "learning_rate": 1.0175468013159384e-07,
+      "loss": 0.7321,
+      "num_input_tokens_seen": 184729600,
+      "step": 22550
+    },
+    {
+      "epoch": 2.9151053107636646,
+      "grad_norm": 0.9104856252670288,
+      "learning_rate": 9.872843196850057e-08,
+      "loss": 0.8234,
+      "num_input_tokens_seen": 184811520,
+      "step": 22560
+    },
+    {
+      "epoch": 2.9163974673730455,
+      "grad_norm": 1.4063791036605835,
+      "learning_rate": 9.574777794253143e-08,
+      "loss": 0.5345,
+      "num_input_tokens_seen": 184893440,
+      "step": 22570
+    },
+    {
+      "epoch": 2.917689623982427,
+      "grad_norm": 2.190293073654175,
+      "learning_rate": 9.281272351124859e-08,
+      "loss": 0.9455,
+      "num_input_tokens_seen": 184975360,
+      "step": 22580
+    },
+    {
+      "epoch": 2.9189817805918077,
+      "grad_norm": 0.6696395874023438,
+      "learning_rate": 8.992327404872825e-08,
+      "loss": 0.3409,
+      "num_input_tokens_seen": 185057280,
+      "step": 22590
+    },
+    {
+      "epoch": 2.9202739372011886,
+      "grad_norm": 1.176025629043579,
+      "learning_rate": 8.707943484553838e-08,
+      "loss": 0.5035,
+      "num_input_tokens_seen": 185139200,
+      "step": 22600
+    },
+    {
+      "epoch": 2.92156609381057,
+      "grad_norm": 0.6432631015777588,
+      "learning_rate": 8.428121110874154e-08,
+      "loss": 0.7041,
+      "num_input_tokens_seen": 185221120,
+      "step": 22610
+    },
+    {
+      "epoch": 2.9228582504199507,
+      "grad_norm": 0.8045111298561096,
+      "learning_rate": 8.152860796187545e-08,
+      "loss": 0.8394,
+      "num_input_tokens_seen": 185303040,
+      "step": 22620
+    },
+    {
+      "epoch": 2.924150407029332,
+      "grad_norm": 0.2948504686355591,
+      "learning_rate": 7.882163044494462e-08,
+      "loss": 0.3898,
+      "num_input_tokens_seen": 185384960,
+      "step": 22630
+    },
+    {
+      "epoch": 2.925442563638713,
+      "grad_norm": 1.1298701763153076,
+      "learning_rate": 7.616028351441484e-08,
+      "loss": 0.7981,
+      "num_input_tokens_seen": 185466880,
+      "step": 22640
+    },
+    {
+      "epoch": 2.9267347202480942,
+      "grad_norm": 0.6935545802116394,
+      "learning_rate": 7.354457204320486e-08,
+      "loss": 0.8605,
+      "num_input_tokens_seen": 185548800,
+      "step": 22650
+    },
+    {
+      "epoch": 2.928026876857475,
+      "grad_norm": 0.624178409576416,
+      "learning_rate": 7.097450082066969e-08,
+      "loss": 0.687,
+      "num_input_tokens_seen": 185630720,
+      "step": 22660
+    },
+    {
+      "epoch": 2.9293190334668564,
+      "grad_norm": 0.7468099594116211,
+      "learning_rate": 6.845007455260343e-08,
+      "loss": 0.6511,
+      "num_input_tokens_seen": 185712640,
+      "step": 22670
+    },
+    {
+      "epoch": 2.9306111900762373,
+      "grad_norm": 0.20860843360424042,
+      "learning_rate": 6.59712978612198e-08,
+      "loss": 0.4383,
+      "num_input_tokens_seen": 185794560,
+      "step": 22680
+    },
+    {
+      "epoch": 2.931903346685618,
+      "grad_norm": 0.3760750889778137,
+      "learning_rate": 6.353817528514938e-08,
+      "loss": 0.4383,
+      "num_input_tokens_seen": 185876480,
+      "step": 22690
+    },
+    {
+      "epoch": 2.9331955032949995,
+      "grad_norm": 0.42980486154556274,
+      "learning_rate": 6.11507112794285e-08,
+      "loss": 0.6782,
+      "num_input_tokens_seen": 185958400,
+      "step": 22700
+    },
+    {
+      "epoch": 2.9344876599043803,
+      "grad_norm": 1.127774715423584,
+      "learning_rate": 5.880891021549928e-08,
+      "loss": 0.676,
+      "num_input_tokens_seen": 186040320,
+      "step": 22710
+    },
+    {
+      "epoch": 2.9357798165137616,
+      "grad_norm": 0.8025173544883728,
+      "learning_rate": 5.6512776381192903e-08,
+      "loss": 0.7285,
+      "num_input_tokens_seen": 186122240,
+      "step": 22720
+    },
+    {
+      "epoch": 2.9370719731231425,
+      "grad_norm": 0.6310427784919739,
+      "learning_rate": 5.426231398071302e-08,
+      "loss": 0.6244,
+      "num_input_tokens_seen": 186204160,
+      "step": 22730
+    },
+    {
+      "epoch": 2.9383641297325234,
+      "grad_norm": 0.8183490037918091,
+      "learning_rate": 5.205752713465794e-08,
+      "loss": 0.8171,
+      "num_input_tokens_seen": 186286080,
+      "step": 22740
+    },
+    {
+      "epoch": 2.9396562863419047,
+      "grad_norm": 1.129209041595459,
+      "learning_rate": 4.989841987997901e-08,
+      "loss": 0.9641,
+      "num_input_tokens_seen": 186368000,
+      "step": 22750
+    },
+    {
+      "epoch": 2.9409484429512855,
+      "grad_norm": 0.7792519927024841,
+      "learning_rate": 4.778499616999166e-08,
+      "loss": 0.7594,
+      "num_input_tokens_seen": 186449920,
+      "step": 22760
+    },
+    {
+      "epoch": 2.942240599560667,
+      "grad_norm": 0.7968661189079285,
+      "learning_rate": 4.57172598743727e-08,
+      "loss": 0.9113,
+      "num_input_tokens_seen": 186531840,
+      "step": 22770
+    },
+    {
+      "epoch": 2.9435327561700477,
+      "grad_norm": 0.6901931166648865,
+      "learning_rate": 4.369521477913529e-08,
+      "loss": 0.8022,
+      "num_input_tokens_seen": 186613760,
+      "step": 22780
+    },
+    {
+      "epoch": 2.9448249127794286,
+      "grad_norm": 0.7363210916519165,
+      "learning_rate": 4.171886458664009e-08,
+      "loss": 0.7282,
+      "num_input_tokens_seen": 186695680,
+      "step": 22790
+    },
+    {
+      "epoch": 2.94611706938881,
+      "grad_norm": 0.35474592447280884,
+      "learning_rate": 3.9788212915573e-08,
+      "loss": 0.6295,
+      "num_input_tokens_seen": 186777600,
+      "step": 22800
+    },
+    {
+      "epoch": 2.947409225998191,
+      "grad_norm": 0.9408590197563171,
+      "learning_rate": 3.7903263300956285e-08,
+      "loss": 0.6325,
+      "num_input_tokens_seen": 186859520,
+      "step": 22810
+    },
+    {
+      "epoch": 2.948701382607572,
+      "grad_norm": 0.9621270895004272,
+      "learning_rate": 3.606401919411806e-08,
+      "loss": 0.8424,
+      "num_input_tokens_seen": 186941440,
+      "step": 22820
+    },
+    {
+      "epoch": 2.949993539216953,
+      "grad_norm": 0.752377450466156,
+      "learning_rate": 3.427048396271171e-08,
+      "loss": 1.1133,
+      "num_input_tokens_seen": 187023360,
+      "step": 22830
+    },
+    {
+      "epoch": 2.9512856958263343,
+      "grad_norm": 0.8131755590438843,
+      "learning_rate": 3.252266089069367e-08,
+      "loss": 0.6982,
+      "num_input_tokens_seen": 187105280,
+      "step": 22840
+    },
+    {
+      "epoch": 2.952577852435715,
+      "grad_norm": 0.7799323797225952,
+      "learning_rate": 3.0820553178320667e-08,
+      "loss": 0.9222,
+      "num_input_tokens_seen": 187187200,
+      "step": 22850
+    },
+    {
+      "epoch": 2.9538700090450964,
+      "grad_norm": 0.4065955877304077,
+      "learning_rate": 2.9164163942146937e-08,
+      "loss": 0.5809,
+      "num_input_tokens_seen": 187269120,
+      "step": 22860
+    },
+    {
+      "epoch": 2.9551621656544773,
+      "grad_norm": 0.6465580463409424,
+      "learning_rate": 2.7553496215015907e-08,
+      "loss": 0.7016,
+      "num_input_tokens_seen": 187351040,
+      "step": 22870
+    },
+    {
+      "epoch": 2.956454322263858,
+      "grad_norm": 0.608161985874176,
+      "learning_rate": 2.5988552946051848e-08,
+      "loss": 0.5702,
+      "num_input_tokens_seen": 187432960,
+      "step": 22880
+    },
+    {
+      "epoch": 2.9577464788732395,
+      "grad_norm": 0.7066859602928162,
+      "learning_rate": 2.44693370006599e-08,
+      "loss": 0.6657,
+      "num_input_tokens_seen": 187514880,
+      "step": 22890
+    },
+    {
+      "epoch": 2.9590386354826204,
+      "grad_norm": 0.31225040555000305,
+      "learning_rate": 2.2995851160520498e-08,
+      "loss": 0.6522,
+      "num_input_tokens_seen": 187596800,
+      "step": 22900
+    },
+    {
+      "epoch": 2.9603307920920017,
+      "grad_norm": 0.7561403512954712,
+      "learning_rate": 2.156809812358107e-08,
+      "loss": 0.8856,
+      "num_input_tokens_seen": 187678720,
+      "step": 22910
+    },
+    {
+      "epoch": 2.9616229487013825,
+      "grad_norm": 1.5660463571548462,
+      "learning_rate": 2.0186080504050466e-08,
+      "loss": 0.718,
+      "num_input_tokens_seen": 187760640,
+      "step": 22920
+    },
+    {
+      "epoch": 2.9629151053107634,
+      "grad_norm": 0.6079100370407104,
+      "learning_rate": 1.8849800832401733e-08,
+      "loss": 0.7686,
+      "num_input_tokens_seen": 187842560,
+      "step": 22930
+    },
+    {
+      "epoch": 2.9642072619201447,
+      "grad_norm": 0.6655999422073364,
+      "learning_rate": 1.75592615553527e-08,
+      "loss": 0.8922,
+      "num_input_tokens_seen": 187924480,
+      "step": 22940
+    },
+    {
+      "epoch": 2.965499418529526,
+      "grad_norm": 0.7710443139076233,
+      "learning_rate": 1.6314465035879855e-08,
+      "loss": 0.5413,
+      "num_input_tokens_seen": 188006400,
+      "step": 22950
+    },
+    {
+      "epoch": 2.966791575138907,
+      "grad_norm": 1.0660786628723145,
+      "learning_rate": 1.5115413553201674e-08,
+      "loss": 0.4049,
+      "num_input_tokens_seen": 188088320,
+      "step": 22960
+    },
+    {
+      "epoch": 2.9680837317482878,
+      "grad_norm": 0.6799715161323547,
+      "learning_rate": 1.3962109302773085e-08,
+      "loss": 1.021,
+      "num_input_tokens_seen": 188170240,
+      "step": 22970
+    },
+    {
+      "epoch": 2.969375888357669,
+      "grad_norm": 1.0657424926757812,
+      "learning_rate": 1.2854554396291018e-08,
+      "loss": 0.4626,
+      "num_input_tokens_seen": 188252160,
+      "step": 22980
+    },
+    {
+      "epoch": 2.97066804496705,
+      "grad_norm": 1.7538788318634033,
+      "learning_rate": 1.1792750861686074e-08,
+      "loss": 0.9079,
+      "num_input_tokens_seen": 188334080,
+      "step": 22990
+    },
+    {
+      "epoch": 2.9719602015764313,
+      "grad_norm": 0.40748029947280884,
+      "learning_rate": 1.0776700643116976e-08,
+      "loss": 0.8018,
+      "num_input_tokens_seen": 188416000,
+      "step": 23000
+    },
+    {
+      "epoch": 2.973252358185812,
+      "grad_norm": 0.6334804892539978,
+      "learning_rate": 9.806405600967794e-09,
+      "loss": 0.6943,
+      "num_input_tokens_seen": 188497920,
+      "step": 23010
+    },
+    {
+      "epoch": 2.974544514795193,
+      "grad_norm": 0.8970416188240051,
+      "learning_rate": 8.881867511845166e-09,
+      "loss": 1.0121,
+      "num_input_tokens_seen": 188579840,
+      "step": 23020
+    },
+    {
+      "epoch": 2.9758366714045743,
+      "grad_norm": 0.32017982006073,
+      "learning_rate": 8.00308806857275e-09,
+      "loss": 1.0772,
+      "num_input_tokens_seen": 188661760,
+      "step": 23030
+    },
+    {
+      "epoch": 2.977128828013955,
+      "grad_norm": 0.7373923659324646,
+      "learning_rate": 7.1700688801940034e-09,
+      "loss": 0.7207,
+      "num_input_tokens_seen": 188743680,
+      "step": 23040
+    },
+    {
+      "epoch": 2.9784209846233365,
+      "grad_norm": 0.6554174423217773,
+      "learning_rate": 6.382811471963846e-09,
+      "loss": 0.589,
+      "num_input_tokens_seen": 188825600,
+      "step": 23050
+    },
+    {
+      "epoch": 2.9797131412327174,
+      "grad_norm": 0.23197783529758453,
+      "learning_rate": 5.6413172853486685e-09,
+      "loss": 0.5539,
+      "num_input_tokens_seen": 188907520,
+      "step": 23060
+    },
+    {
+      "epoch": 2.9810052978420982,
+      "grad_norm": 0.8123525977134705,
+      "learning_rate": 4.94558767802078e-09,
+      "loss": 1.0715,
+      "num_input_tokens_seen": 188989440,
+      "step": 23070
+    },
+    {
+      "epoch": 2.9822974544514795,
+      "grad_norm": 0.8512336015701294,
+      "learning_rate": 4.295623923858405e-09,
+      "loss": 0.9075,
+      "num_input_tokens_seen": 189071360,
+      "step": 23080
+    },
+    {
+      "epoch": 2.983589611060861,
+      "grad_norm": 0.5837238430976868,
+      "learning_rate": 3.6914272129429106e-09,
+      "loss": 0.9787,
+      "num_input_tokens_seen": 189153280,
+      "step": 23090
+    },
+    {
+      "epoch": 2.9848817676702417,
+      "grad_norm": 0.7648143172264099,
+      "learning_rate": 3.1329986515560295e-09,
+      "loss": 0.397,
+      "num_input_tokens_seen": 189235200,
+      "step": 23100
+    },
+    {
+      "epoch": 2.9861739242796226,
+      "grad_norm": 0.7528153657913208,
+      "learning_rate": 2.6203392621798605e-09,
+      "loss": 0.8219,
+      "num_input_tokens_seen": 189317120,
+      "step": 23110
+    },
+    {
+      "epoch": 2.987466080889004,
+      "grad_norm": 0.9112277626991272,
+      "learning_rate": 2.153449983491318e-09,
+      "loss": 0.93,
+      "num_input_tokens_seen": 189399040,
+      "step": 23120
+    },
+    {
+      "epoch": 2.9887582374983848,
+      "grad_norm": 0.6468605995178223,
+      "learning_rate": 1.7323316703621305e-09,
+      "loss": 1.1683,
+      "num_input_tokens_seen": 189480960,
+      "step": 23130
+    },
+    {
+      "epoch": 2.990050394107766,
+      "grad_norm": 0.8173196315765381,
+      "learning_rate": 1.356985093856067e-09,
+      "loss": 0.5163,
+      "num_input_tokens_seen": 189562880,
+      "step": 23140
+    },
+    {
+      "epoch": 2.991342550717147,
+      "grad_norm": 0.9355853796005249,
+      "learning_rate": 1.0274109412372613e-09,
+      "loss": 1.0414,
+      "num_input_tokens_seen": 189644800,
+      "step": 23150
+    },
+    {
+      "epoch": 2.992634707326528,
+      "grad_norm": 0.6136227250099182,
+      "learning_rate": 7.436098159480099e-10,
+      "loss": 0.743,
+      "num_input_tokens_seen": 189726720,
+      "step": 23160
+    },
+    {
+      "epoch": 2.993926863935909,
+      "grad_norm": 0.31119829416275024,
+      "learning_rate": 5.055822376337505e-10,
+      "loss": 0.941,
+      "num_input_tokens_seen": 189808640,
+      "step": 23170
+    },
+    {
+      "epoch": 2.99521902054529,
+      "grad_norm": 0.6250201463699341,
+      "learning_rate": 3.1332864211808254e-10,
+      "loss": 0.6753,
+      "num_input_tokens_seen": 189890560,
+      "step": 23180
+    },
+    {
+      "epoch": 2.9965111771546713,
+      "grad_norm": 0.5410284996032715,
+      "learning_rate": 1.6684938141664498e-10,
+      "loss": 0.8739,
+      "num_input_tokens_seen": 189972480,
+      "step": 23190
+    },
+    {
+      "epoch": 2.997803333764052,
+      "grad_norm": 0.7503390908241272,
+      "learning_rate": 6.614472373434044e-11,
+      "loss": 0.851,
+      "num_input_tokens_seen": 190054400,
+      "step": 23200
+    },
+    {
+      "epoch": 2.999095490373433,
+      "grad_norm": 0.6340755224227905,
+      "learning_rate": 1.1214853459784457e-11,
+      "loss": 0.7338,
+      "num_input_tokens_seen": 190136320,
+      "step": 23210
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 23217,
+  "num_input_tokens_seen": 190193664,
+  "num_train_epochs": 3,
+  "save_steps": 1000,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 8.660062333952852e+18,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}