diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,13163 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.9999466719486831,
+  "eval_steps": 500,
+  "global_step": 16407,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0006094634436208223,
+      "grad_norm": 9.033647537231445,
+      "learning_rate": 9.7442143727162e-07,
+      "loss": 6.6451,
+      "num_input_tokens_seen": 1401772,
+      "step": 10
+    },
+    {
+      "epoch": 0.0012189268872416447,
+      "grad_norm": 8.373307228088379,
+      "learning_rate": 1.94884287454324e-06,
+      "loss": 6.4199,
+      "num_input_tokens_seen": 2825364,
+      "step": 20
+    },
+    {
+      "epoch": 0.001828390330862467,
+      "grad_norm": 8.205723762512207,
+      "learning_rate": 2.92326431181486e-06,
+      "loss": 5.9935,
+      "num_input_tokens_seen": 4259836,
+      "step": 30
+    },
+    {
+      "epoch": 0.0024378537744832894,
+      "grad_norm": 8.017410278320312,
+      "learning_rate": 3.89768574908648e-06,
+      "loss": 5.8172,
+      "num_input_tokens_seen": 5653988,
+      "step": 40
+    },
+    {
+      "epoch": 0.0030473172181041115,
+      "grad_norm": 8.248656272888184,
+      "learning_rate": 4.8721071863581e-06,
+      "loss": 5.4176,
+      "num_input_tokens_seen": 7026620,
+      "step": 50
+    },
+    {
+      "epoch": 0.003656780661724934,
+      "grad_norm": 6.726569175720215,
+      "learning_rate": 5.84652862362972e-06,
+      "loss": 5.0672,
+      "num_input_tokens_seen": 8474412,
+      "step": 60
+    },
+    {
+      "epoch": 0.004266244105345756,
+      "grad_norm": 7.437803268432617,
+      "learning_rate": 6.8209500609013406e-06,
+      "loss": 4.6725,
+      "num_input_tokens_seen": 9888436,
+      "step": 70
+    },
+    {
+      "epoch": 0.004875707548966579,
+      "grad_norm": 7.783193111419678,
+      "learning_rate": 7.79537149817296e-06,
+      "loss": 4.0473,
+      "num_input_tokens_seen": 11322860,
+      "step": 80
+    },
+    {
+      "epoch": 0.0054851709925874004,
+      "grad_norm": 1.6730059385299683,
+      "learning_rate": 8.769792935444581e-06,
+      "loss": 3.3354,
+      "num_input_tokens_seen": 12764288,
+      "step": 90
+    },
+    {
+      "epoch": 0.006094634436208223,
+      "grad_norm": 1.6392016410827637,
+      "learning_rate": 9.7442143727162e-06,
+      "loss": 2.9862,
+      "num_input_tokens_seen": 14208968,
+      "step": 100
+    },
+    {
+      "epoch": 0.0067040978798290456,
+      "grad_norm": 1.278815507888794,
+      "learning_rate": 1.071863580998782e-05,
+      "loss": 2.7488,
+      "num_input_tokens_seen": 15575528,
+      "step": 110
+    },
+    {
+      "epoch": 0.007313561323449868,
+      "grad_norm": 1.2898283004760742,
+      "learning_rate": 1.169305724725944e-05,
+      "loss": 2.6214,
+      "num_input_tokens_seen": 16993668,
+      "step": 120
+    },
+    {
+      "epoch": 0.00792302476707069,
+      "grad_norm": 1.0845414400100708,
+      "learning_rate": 1.266747868453106e-05,
+      "loss": 2.5408,
+      "num_input_tokens_seen": 18424664,
+      "step": 130
+    },
+    {
+      "epoch": 0.008532488210691512,
+      "grad_norm": 1.1549726724624634,
+      "learning_rate": 1.3641900121802681e-05,
+      "loss": 2.4494,
+      "num_input_tokens_seen": 19842364,
+      "step": 140
+    },
+    {
+      "epoch": 0.009141951654312334,
+      "grad_norm": 1.2677397727966309,
+      "learning_rate": 1.46163215590743e-05,
+      "loss": 2.4496,
+      "num_input_tokens_seen": 21287212,
+      "step": 150
+    },
+    {
+      "epoch": 0.009751415097933157,
+      "grad_norm": 1.2440944910049438,
+      "learning_rate": 1.559074299634592e-05,
+      "loss": 2.4503,
+      "num_input_tokens_seen": 22706312,
+      "step": 160
+    },
+    {
+      "epoch": 0.01036087854155398,
+      "grad_norm": 1.2314107418060303,
+      "learning_rate": 1.656516443361754e-05,
+      "loss": 2.3764,
+      "num_input_tokens_seen": 24127816,
+      "step": 170
+    },
+    {
+      "epoch": 0.010970341985174801,
+      "grad_norm": 1.0929701328277588,
+      "learning_rate": 1.7539585870889162e-05,
+      "loss": 2.4311,
+      "num_input_tokens_seen": 25495676,
+      "step": 180
+    },
+    {
+      "epoch": 0.011579805428795624,
+      "grad_norm": 1.0920017957687378,
+      "learning_rate": 1.8514007308160783e-05,
+      "loss": 2.3656,
+      "num_input_tokens_seen": 26913848,
+      "step": 190
+    },
+    {
+      "epoch": 0.012189268872416446,
+      "grad_norm": 1.259750485420227,
+      "learning_rate": 1.94884287454324e-05,
+      "loss": 2.2739,
+      "num_input_tokens_seen": 28326716,
+      "step": 200
+    },
+    {
+      "epoch": 0.01279873231603727,
+      "grad_norm": 1.0931645631790161,
+      "learning_rate": 2.046285018270402e-05,
+      "loss": 2.231,
+      "num_input_tokens_seen": 29699424,
+      "step": 210
+    },
+    {
+      "epoch": 0.013408195759658091,
+      "grad_norm": 1.0463930368423462,
+      "learning_rate": 2.143727161997564e-05,
+      "loss": 2.216,
+      "num_input_tokens_seen": 31109904,
+      "step": 220
+    },
+    {
+      "epoch": 0.014017659203278913,
+      "grad_norm": 1.0597470998764038,
+      "learning_rate": 2.241169305724726e-05,
+      "loss": 2.1803,
+      "num_input_tokens_seen": 32547736,
+      "step": 230
+    },
+    {
+      "epoch": 0.014627122646899736,
+      "grad_norm": 1.1559436321258545,
+      "learning_rate": 2.338611449451888e-05,
+      "loss": 2.1441,
+      "num_input_tokens_seen": 33972620,
+      "step": 240
+    },
+    {
+      "epoch": 0.015236586090520558,
+      "grad_norm": 1.2502082586288452,
+      "learning_rate": 2.4360535931790504e-05,
+      "loss": 2.1537,
+      "num_input_tokens_seen": 35404252,
+      "step": 250
+    },
+    {
+      "epoch": 0.01584604953414138,
+      "grad_norm": 1.096165418624878,
+      "learning_rate": 2.533495736906212e-05,
+      "loss": 2.1193,
+      "num_input_tokens_seen": 36804296,
+      "step": 260
+    },
+    {
+      "epoch": 0.016455512977762203,
+      "grad_norm": 1.0119866132736206,
+      "learning_rate": 2.630937880633374e-05,
+      "loss": 2.1115,
+      "num_input_tokens_seen": 38185184,
+      "step": 270
+    },
+    {
+      "epoch": 0.017064976421383025,
+      "grad_norm": 1.1104751825332642,
+      "learning_rate": 2.7283800243605362e-05,
+      "loss": 2.2292,
+      "num_input_tokens_seen": 39600904,
+      "step": 280
+    },
+    {
+      "epoch": 0.017674439865003846,
+      "grad_norm": 0.9693301320075989,
+      "learning_rate": 2.8258221680876983e-05,
+      "loss": 2.1193,
+      "num_input_tokens_seen": 41014344,
+      "step": 290
+    },
+    {
+      "epoch": 0.018283903308624668,
+      "grad_norm": 0.9671382308006287,
+      "learning_rate": 2.92326431181486e-05,
+      "loss": 2.052,
+      "num_input_tokens_seen": 42441088,
+      "step": 300
+    },
+    {
+      "epoch": 0.018893366752245493,
+      "grad_norm": 1.053604006767273,
+      "learning_rate": 3.020706455542022e-05,
+      "loss": 2.081,
+      "num_input_tokens_seen": 43835948,
+      "step": 310
+    },
+    {
+      "epoch": 0.019502830195866315,
+      "grad_norm": 0.9839246869087219,
+      "learning_rate": 3.118148599269184e-05,
+      "loss": 2.0446,
+      "num_input_tokens_seen": 45267668,
+      "step": 320
+    },
+    {
+      "epoch": 0.020112293639487137,
+      "grad_norm": 1.0461212396621704,
+      "learning_rate": 3.215590742996346e-05,
+      "loss": 2.0046,
+      "num_input_tokens_seen": 46734856,
+      "step": 330
+    },
+    {
+      "epoch": 0.02072175708310796,
+      "grad_norm": 0.9702940583229065,
+      "learning_rate": 3.313032886723508e-05,
+      "loss": 2.0476,
+      "num_input_tokens_seen": 48137552,
+      "step": 340
+    },
+    {
+      "epoch": 0.02133122052672878,
+      "grad_norm": 1.0886714458465576,
+      "learning_rate": 3.41047503045067e-05,
+      "loss": 2.0536,
+      "num_input_tokens_seen": 49555456,
+      "step": 350
+    },
+    {
+      "epoch": 0.021940683970349602,
+      "grad_norm": 1.0513168573379517,
+      "learning_rate": 3.5079171741778324e-05,
+      "loss": 2.0185,
+      "num_input_tokens_seen": 50949976,
+      "step": 360
+    },
+    {
+      "epoch": 0.022550147413970427,
+      "grad_norm": 1.0875654220581055,
+      "learning_rate": 3.605359317904994e-05,
+      "loss": 1.9939,
+      "num_input_tokens_seen": 52330820,
+      "step": 370
+    },
+    {
+      "epoch": 0.02315961085759125,
+      "grad_norm": 0.9643621444702148,
+      "learning_rate": 3.7028014616321566e-05,
+      "loss": 2.0392,
+      "num_input_tokens_seen": 53764024,
+      "step": 380
+    },
+    {
+      "epoch": 0.02376907430121207,
+      "grad_norm": 1.0301045179367065,
+      "learning_rate": 3.800243605359318e-05,
+      "loss": 2.0448,
+      "num_input_tokens_seen": 55201816,
+      "step": 390
+    },
+    {
+      "epoch": 0.024378537744832892,
+      "grad_norm": 1.0076828002929688,
+      "learning_rate": 3.89768574908648e-05,
+      "loss": 1.9415,
+      "num_input_tokens_seen": 56622004,
+      "step": 400
+    },
+    {
+      "epoch": 0.024988001188453714,
+      "grad_norm": 0.9446476101875305,
+      "learning_rate": 3.9951278928136424e-05,
+      "loss": 1.9752,
+      "num_input_tokens_seen": 58038112,
+      "step": 410
+    },
+    {
+      "epoch": 0.02559746463207454,
+      "grad_norm": 0.952552318572998,
+      "learning_rate": 4.092570036540804e-05,
+      "loss": 1.9806,
+      "num_input_tokens_seen": 59484616,
+      "step": 420
+    },
+    {
+      "epoch": 0.02620692807569536,
+      "grad_norm": 0.8900318145751953,
+      "learning_rate": 4.1900121802679666e-05,
+      "loss": 1.9547,
+      "num_input_tokens_seen": 60889256,
+      "step": 430
+    },
+    {
+      "epoch": 0.026816391519316182,
+      "grad_norm": 0.9296801090240479,
+      "learning_rate": 4.287454323995128e-05,
+      "loss": 2.0001,
+      "num_input_tokens_seen": 62274740,
+      "step": 440
+    },
+    {
+      "epoch": 0.027425854962937004,
+      "grad_norm": 0.9372000098228455,
+      "learning_rate": 4.38489646772229e-05,
+      "loss": 2.001,
+      "num_input_tokens_seen": 63719728,
+      "step": 450
+    },
+    {
+      "epoch": 0.028035318406557826,
+      "grad_norm": 1.1481891870498657,
+      "learning_rate": 4.482338611449452e-05,
+      "loss": 2.0047,
+      "num_input_tokens_seen": 65120052,
+      "step": 460
+    },
+    {
+      "epoch": 0.028644781850178647,
+      "grad_norm": 0.9526330232620239,
+      "learning_rate": 4.579780755176614e-05,
+      "loss": 1.8862,
+      "num_input_tokens_seen": 66516240,
+      "step": 470
+    },
+    {
+      "epoch": 0.029254245293799472,
+      "grad_norm": 0.9846788048744202,
+      "learning_rate": 4.677222898903776e-05,
+      "loss": 1.8632,
+      "num_input_tokens_seen": 67966636,
+      "step": 480
+    },
+    {
+      "epoch": 0.029863708737420294,
+      "grad_norm": 0.9888688921928406,
+      "learning_rate": 4.774665042630938e-05,
+      "loss": 1.7964,
+      "num_input_tokens_seen": 69405656,
+      "step": 490
+    },
+    {
+      "epoch": 0.030473172181041116,
+      "grad_norm": 0.9095290303230286,
+      "learning_rate": 4.872107186358101e-05,
+      "loss": 1.9531,
+      "num_input_tokens_seen": 70820452,
+      "step": 500
+    },
+    {
+      "epoch": 0.031082635624661938,
+      "grad_norm": 0.8255355358123779,
+      "learning_rate": 4.9695493300852625e-05,
+      "loss": 1.951,
+      "num_input_tokens_seen": 72275948,
+      "step": 510
+    },
+    {
+      "epoch": 0.03169209906828276,
+      "grad_norm": 0.8973667025566101,
+      "learning_rate": 5.066991473812424e-05,
+      "loss": 1.9297,
+      "num_input_tokens_seen": 73697960,
+      "step": 520
+    },
+    {
+      "epoch": 0.03230156251190358,
+      "grad_norm": 1.0261414051055908,
+      "learning_rate": 5.164433617539586e-05,
+      "loss": 1.8723,
+      "num_input_tokens_seen": 75117608,
+      "step": 530
+    },
+    {
+      "epoch": 0.032911025955524406,
+      "grad_norm": 0.8425599336624146,
+      "learning_rate": 5.261875761266748e-05,
+      "loss": 1.8414,
+      "num_input_tokens_seen": 76461224,
+      "step": 540
+    },
+    {
+      "epoch": 0.033520489399145224,
+      "grad_norm": 0.9196986556053162,
+      "learning_rate": 5.35931790499391e-05,
+      "loss": 1.9135,
+      "num_input_tokens_seen": 77904064,
+      "step": 550
+    },
+    {
+      "epoch": 0.03412995284276605,
+      "grad_norm": 0.8252116441726685,
+      "learning_rate": 5.4567600487210725e-05,
+      "loss": 1.8992,
+      "num_input_tokens_seen": 79294400,
+      "step": 560
+    },
+    {
+      "epoch": 0.034739416286386875,
+      "grad_norm": 1.140308141708374,
+      "learning_rate": 5.554202192448234e-05,
+      "loss": 1.8564,
+      "num_input_tokens_seen": 80694984,
+      "step": 570
+    },
+    {
+      "epoch": 0.03534887973000769,
+      "grad_norm": 0.772463321685791,
+      "learning_rate": 5.6516443361753966e-05,
+      "loss": 1.9372,
+      "num_input_tokens_seen": 82116012,
+      "step": 580
+    },
+    {
+      "epoch": 0.03595834317362852,
+      "grad_norm": 0.8610614538192749,
+      "learning_rate": 5.749086479902558e-05,
+      "loss": 1.8236,
+      "num_input_tokens_seen": 83500604,
+      "step": 590
+    },
+    {
+      "epoch": 0.036567806617249336,
+      "grad_norm": 1.0011142492294312,
+      "learning_rate": 5.84652862362972e-05,
+      "loss": 1.875,
+      "num_input_tokens_seen": 84936208,
+      "step": 600
+    },
+    {
+      "epoch": 0.03717727006087016,
+      "grad_norm": 0.9461501836776733,
+      "learning_rate": 5.9439707673568825e-05,
+      "loss": 1.8397,
+      "num_input_tokens_seen": 86347300,
+      "step": 610
+    },
+    {
+      "epoch": 0.03778673350449099,
+      "grad_norm": 0.8287586569786072,
+      "learning_rate": 6.041412911084044e-05,
+      "loss": 1.8283,
+      "num_input_tokens_seen": 87787620,
+      "step": 620
+    },
+    {
+      "epoch": 0.038396196948111805,
+      "grad_norm": 0.9248697757720947,
+      "learning_rate": 6.138855054811207e-05,
+      "loss": 1.8192,
+      "num_input_tokens_seen": 89217528,
+      "step": 630
+    },
+    {
+      "epoch": 0.03900566039173263,
+      "grad_norm": 0.8168230056762695,
+      "learning_rate": 6.236297198538368e-05,
+      "loss": 1.8232,
+      "num_input_tokens_seen": 90602460,
+      "step": 640
+    },
+    {
+      "epoch": 0.03961512383535345,
+      "grad_norm": 0.8372480273246765,
+      "learning_rate": 6.33373934226553e-05,
+      "loss": 1.9267,
+      "num_input_tokens_seen": 91995380,
+      "step": 650
+    },
+    {
+      "epoch": 0.04022458727897427,
+      "grad_norm": 1.0177308320999146,
+      "learning_rate": 6.431181485992692e-05,
+      "loss": 1.8319,
+      "num_input_tokens_seen": 93396324,
+      "step": 660
+    },
+    {
+      "epoch": 0.0408340507225951,
+      "grad_norm": 0.9811750054359436,
+      "learning_rate": 6.528623629719854e-05,
+      "loss": 1.8442,
+      "num_input_tokens_seen": 94835448,
+      "step": 670
+    },
+    {
+      "epoch": 0.04144351416621592,
+      "grad_norm": 0.8578136563301086,
+      "learning_rate": 6.626065773447017e-05,
+      "loss": 1.8373,
+      "num_input_tokens_seen": 96247680,
+      "step": 680
+    },
+    {
+      "epoch": 0.04205297760983674,
+      "grad_norm": 0.7805215120315552,
+      "learning_rate": 6.723507917174178e-05,
+      "loss": 1.8717,
+      "num_input_tokens_seen": 97649748,
+      "step": 690
+    },
+    {
+      "epoch": 0.04266244105345756,
+      "grad_norm": 0.7658279538154602,
+      "learning_rate": 6.82095006090134e-05,
+      "loss": 1.8215,
+      "num_input_tokens_seen": 99072716,
+      "step": 700
+    },
+    {
+      "epoch": 0.043271904497078385,
+      "grad_norm": 0.8456059098243713,
+      "learning_rate": 6.918392204628502e-05,
+      "loss": 1.9152,
+      "num_input_tokens_seen": 100460184,
+      "step": 710
+    },
+    {
+      "epoch": 0.043881367940699204,
+      "grad_norm": 0.7689954042434692,
+      "learning_rate": 7.015834348355665e-05,
+      "loss": 1.8299,
+      "num_input_tokens_seen": 101833788,
+      "step": 720
+    },
+    {
+      "epoch": 0.04449083138432003,
+      "grad_norm": 0.769614577293396,
+      "learning_rate": 7.113276492082827e-05,
+      "loss": 1.7905,
+      "num_input_tokens_seen": 103233860,
+      "step": 730
+    },
+    {
+      "epoch": 0.045100294827940854,
+      "grad_norm": 0.984970211982727,
+      "learning_rate": 7.210718635809988e-05,
+      "loss": 1.7542,
+      "num_input_tokens_seen": 104649404,
+      "step": 740
+    },
+    {
+      "epoch": 0.04570975827156167,
+      "grad_norm": 0.898064136505127,
+      "learning_rate": 7.30816077953715e-05,
+      "loss": 1.8333,
+      "num_input_tokens_seen": 106019108,
+      "step": 750
+    },
+    {
+      "epoch": 0.0463192217151825,
+      "grad_norm": 0.8178762197494507,
+      "learning_rate": 7.405602923264313e-05,
+      "loss": 1.7696,
+      "num_input_tokens_seen": 107412140,
+      "step": 760
+    },
+    {
+      "epoch": 0.046928685158803315,
+      "grad_norm": 0.8419963121414185,
+      "learning_rate": 7.503045066991475e-05,
+      "loss": 1.8582,
+      "num_input_tokens_seen": 108814020,
+      "step": 770
+    },
+    {
+      "epoch": 0.04753814860242414,
+      "grad_norm": 0.8736382126808167,
+      "learning_rate": 7.600487210718637e-05,
+      "loss": 1.8505,
+      "num_input_tokens_seen": 110243636,
+      "step": 780
+    },
+    {
+      "epoch": 0.048147612046044966,
+      "grad_norm": 0.8330615758895874,
+      "learning_rate": 7.697929354445798e-05,
+      "loss": 1.8439,
+      "num_input_tokens_seen": 111661300,
+      "step": 790
+    },
+    {
+      "epoch": 0.048757075489665784,
+      "grad_norm": 0.8429380059242249,
+      "learning_rate": 7.79537149817296e-05,
+      "loss": 1.7612,
+      "num_input_tokens_seen": 113084304,
+      "step": 800
+    },
+    {
+      "epoch": 0.04936653893328661,
+      "grad_norm": 0.7914267182350159,
+      "learning_rate": 7.892813641900122e-05,
+      "loss": 1.8355,
+      "num_input_tokens_seen": 114514328,
+      "step": 810
+    },
+    {
+      "epoch": 0.04997600237690743,
+      "grad_norm": 0.8499755263328552,
+      "learning_rate": 7.990255785627285e-05,
+      "loss": 1.762,
+      "num_input_tokens_seen": 115939680,
+      "step": 820
+    },
+    {
+      "epoch": 0.05058546582052825,
+      "grad_norm": 0.8349440693855286,
+      "learning_rate": 8e-05,
+      "loss": 1.7935,
+      "num_input_tokens_seen": 117339404,
+      "step": 830
+    },
+    {
+      "epoch": 0.05119492926414908,
+      "grad_norm": 0.8505790829658508,
+      "learning_rate": 8e-05,
+      "loss": 1.8398,
+      "num_input_tokens_seen": 118763072,
+      "step": 840
+    },
+    {
+      "epoch": 0.051804392707769896,
+      "grad_norm": 0.761175274848938,
+      "learning_rate": 8e-05,
+      "loss": 1.7851,
+      "num_input_tokens_seen": 120156740,
+      "step": 850
+    },
+    {
+      "epoch": 0.05241385615139072,
+      "grad_norm": 0.8605666756629944,
+      "learning_rate": 8e-05,
+      "loss": 1.8358,
+      "num_input_tokens_seen": 121570328,
+      "step": 860
+    },
+    {
+      "epoch": 0.05302331959501154,
+      "grad_norm": 0.648381769657135,
+      "learning_rate": 8e-05,
+      "loss": 1.7502,
+      "num_input_tokens_seen": 122939696,
+      "step": 870
+    },
+    {
+      "epoch": 0.053632783038632365,
+      "grad_norm": 0.9061549305915833,
+      "learning_rate": 8e-05,
+      "loss": 1.7837,
+      "num_input_tokens_seen": 124365264,
+      "step": 880
+    },
+    {
+      "epoch": 0.05424224648225319,
+      "grad_norm": 0.8057026267051697,
+      "learning_rate": 8e-05,
+      "loss": 1.7609,
+      "num_input_tokens_seen": 125746168,
+      "step": 890
+    },
+    {
+      "epoch": 0.05485170992587401,
+      "grad_norm": 0.8674312829971313,
+      "learning_rate": 8e-05,
+      "loss": 1.7415,
+      "num_input_tokens_seen": 127137780,
+      "step": 900
+    },
+    {
+      "epoch": 0.05546117336949483,
+      "grad_norm": 0.8366326689720154,
+      "learning_rate": 8e-05,
+      "loss": 1.8147,
+      "num_input_tokens_seen": 128546252,
+      "step": 910
+    },
+    {
+      "epoch": 0.05607063681311565,
+      "grad_norm": 0.8231328129768372,
+      "learning_rate": 8e-05,
+      "loss": 1.8652,
+      "num_input_tokens_seen": 129950036,
+      "step": 920
+    },
+    {
+      "epoch": 0.056680100256736476,
+      "grad_norm": 0.7696998715400696,
+      "learning_rate": 8e-05,
+      "loss": 1.8445,
+      "num_input_tokens_seen": 131396708,
+      "step": 930
+    },
+    {
+      "epoch": 0.057289563700357295,
+      "grad_norm": 0.8545131683349609,
+      "learning_rate": 8e-05,
+      "loss": 1.8091,
+      "num_input_tokens_seen": 132834744,
+      "step": 940
+    },
+    {
+      "epoch": 0.05789902714397812,
+      "grad_norm": 0.8772911429405212,
+      "learning_rate": 8e-05,
+      "loss": 1.822,
+      "num_input_tokens_seen": 134197864,
+      "step": 950
+    },
+    {
+      "epoch": 0.058508490587598945,
+      "grad_norm": 0.7780118584632874,
+      "learning_rate": 8e-05,
+      "loss": 1.6964,
+      "num_input_tokens_seen": 135587736,
+      "step": 960
+    },
+    {
+      "epoch": 0.05911795403121976,
+      "grad_norm": 0.8001295328140259,
+      "learning_rate": 8e-05,
+      "loss": 1.6976,
+      "num_input_tokens_seen": 136998024,
+      "step": 970
+    },
+    {
+      "epoch": 0.05972741747484059,
+      "grad_norm": 0.8727300763130188,
+      "learning_rate": 8e-05,
+      "loss": 1.6902,
+      "num_input_tokens_seen": 138396964,
+      "step": 980
+    },
+    {
+      "epoch": 0.06033688091846141,
+      "grad_norm": 0.9872994422912598,
+      "learning_rate": 8e-05,
+      "loss": 1.7319,
+      "num_input_tokens_seen": 139815224,
+      "step": 990
+    },
+    {
+      "epoch": 0.06094634436208223,
+      "grad_norm": 0.8009439706802368,
+      "learning_rate": 8e-05,
+      "loss": 1.6803,
+      "num_input_tokens_seen": 141251384,
+      "step": 1000
+    },
+    {
+      "epoch": 0.06155580780570306,
+      "grad_norm": 0.7269445657730103,
+      "learning_rate": 8e-05,
+      "loss": 1.7466,
+      "num_input_tokens_seen": 142666948,
+      "step": 1010
+    },
+    {
+      "epoch": 0.062165271249323875,
+      "grad_norm": 0.7160496711730957,
+      "learning_rate": 8e-05,
+      "loss": 1.7092,
+      "num_input_tokens_seen": 144099664,
+      "step": 1020
+    },
+    {
+      "epoch": 0.0627747346929447,
+      "grad_norm": 0.8215118646621704,
+      "learning_rate": 8e-05,
+      "loss": 1.8191,
+      "num_input_tokens_seen": 145505712,
+      "step": 1030
+    },
+    {
+      "epoch": 0.06338419813656553,
+      "grad_norm": 0.7651961445808411,
+      "learning_rate": 8e-05,
+      "loss": 1.7255,
+      "num_input_tokens_seen": 146921148,
+      "step": 1040
+    },
+    {
+      "epoch": 0.06399366158018634,
+      "grad_norm": 0.8512997031211853,
+      "learning_rate": 8e-05,
+      "loss": 1.7603,
+      "num_input_tokens_seen": 148324952,
+      "step": 1050
+    },
+    {
+      "epoch": 0.06460312502380716,
+      "grad_norm": 0.7710238695144653,
+      "learning_rate": 8e-05,
+      "loss": 1.7784,
+      "num_input_tokens_seen": 149742540,
+      "step": 1060
+    },
+    {
+      "epoch": 0.06521258846742799,
+      "grad_norm": 0.7975362539291382,
+      "learning_rate": 8e-05,
+      "loss": 1.7519,
+      "num_input_tokens_seen": 151140136,
+      "step": 1070
+    },
+    {
+      "epoch": 0.06582205191104881,
+      "grad_norm": 0.8341178894042969,
+      "learning_rate": 8e-05,
+      "loss": 1.7708,
+      "num_input_tokens_seen": 152574668,
+      "step": 1080
+    },
+    {
+      "epoch": 0.06643151535466964,
+      "grad_norm": 0.81528240442276,
+      "learning_rate": 8e-05,
+      "loss": 1.7906,
+      "num_input_tokens_seen": 153945316,
+      "step": 1090
+    },
+    {
+      "epoch": 0.06704097879829045,
+      "grad_norm": 0.7310307621955872,
+      "learning_rate": 8e-05,
+      "loss": 1.7695,
+      "num_input_tokens_seen": 155334976,
+      "step": 1100
+    },
+    {
+      "epoch": 0.06765044224191127,
+      "grad_norm": 0.7746317982673645,
+      "learning_rate": 8e-05,
+      "loss": 1.7351,
+      "num_input_tokens_seen": 156703560,
+      "step": 1110
+    },
+    {
+      "epoch": 0.0682599056855321,
+      "grad_norm": 0.7563662528991699,
+      "learning_rate": 8e-05,
+      "loss": 1.8099,
+      "num_input_tokens_seen": 158068384,
+      "step": 1120
+    },
+    {
+      "epoch": 0.06886936912915292,
+      "grad_norm": 0.7843930721282959,
+      "learning_rate": 8e-05,
+      "loss": 1.6761,
+      "num_input_tokens_seen": 159493532,
+      "step": 1130
+    },
+    {
+      "epoch": 0.06947883257277375,
+      "grad_norm": 0.8035304546356201,
+      "learning_rate": 8e-05,
+      "loss": 1.6405,
+      "num_input_tokens_seen": 160906460,
+      "step": 1140
+    },
+    {
+      "epoch": 0.07008829601639456,
+      "grad_norm": 0.6847174763679504,
+      "learning_rate": 8e-05,
+      "loss": 1.7033,
+      "num_input_tokens_seen": 162330960,
+      "step": 1150
+    },
+    {
+      "epoch": 0.07069775946001539,
+      "grad_norm": 0.7156793475151062,
+      "learning_rate": 8e-05,
+      "loss": 1.7531,
+      "num_input_tokens_seen": 163763200,
+      "step": 1160
+    },
+    {
+      "epoch": 0.07130722290363621,
+      "grad_norm": 0.7055004239082336,
+      "learning_rate": 8e-05,
+      "loss": 1.6826,
+      "num_input_tokens_seen": 165128772,
+      "step": 1170
+    },
+    {
+      "epoch": 0.07191668634725704,
+      "grad_norm": 0.689820408821106,
+      "learning_rate": 8e-05,
+      "loss": 1.7022,
+      "num_input_tokens_seen": 166506640,
+      "step": 1180
+    },
+    {
+      "epoch": 0.07252614979087786,
+      "grad_norm": 0.7453446388244629,
+      "learning_rate": 8e-05,
+      "loss": 1.7654,
+      "num_input_tokens_seen": 167908440,
+      "step": 1190
+    },
+    {
+      "epoch": 0.07313561323449867,
+      "grad_norm": 0.768242597579956,
+      "learning_rate": 8e-05,
+      "loss": 1.732,
+      "num_input_tokens_seen": 169276732,
+      "step": 1200
+    },
+    {
+      "epoch": 0.0737450766781195,
+      "grad_norm": 0.7208901047706604,
+      "learning_rate": 8e-05,
+      "loss": 1.754,
+      "num_input_tokens_seen": 170678632,
+      "step": 1210
+    },
+    {
+      "epoch": 0.07435454012174032,
+      "grad_norm": 0.7076956629753113,
+      "learning_rate": 8e-05,
+      "loss": 1.6955,
+      "num_input_tokens_seen": 172093484,
+      "step": 1220
+    },
+    {
+      "epoch": 0.07496400356536115,
+      "grad_norm": 0.7701781392097473,
+      "learning_rate": 8e-05,
+      "loss": 1.7475,
+      "num_input_tokens_seen": 173482152,
+      "step": 1230
+    },
+    {
+      "epoch": 0.07557346700898197,
+      "grad_norm": 0.6883390545845032,
+      "learning_rate": 8e-05,
+      "loss": 1.7302,
+      "num_input_tokens_seen": 174939192,
+      "step": 1240
+    },
+    {
+      "epoch": 0.07618293045260278,
+      "grad_norm": 0.6942645311355591,
+      "learning_rate": 8e-05,
+      "loss": 1.771,
+      "num_input_tokens_seen": 176393408,
+      "step": 1250
+    },
+    {
+      "epoch": 0.07679239389622361,
+      "grad_norm": 0.6907592415809631,
+      "learning_rate": 8e-05,
+      "loss": 1.8247,
+      "num_input_tokens_seen": 177820192,
+      "step": 1260
+    },
+    {
+      "epoch": 0.07740185733984443,
+      "grad_norm": 0.6885057687759399,
+      "learning_rate": 8e-05,
+      "loss": 1.7403,
+      "num_input_tokens_seen": 179198048,
+      "step": 1270
+    },
+    {
+      "epoch": 0.07801132078346526,
+      "grad_norm": 0.7358853816986084,
+      "learning_rate": 8e-05,
+      "loss": 1.716,
+      "num_input_tokens_seen": 180638000,
+      "step": 1280
+    },
+    {
+      "epoch": 0.07862078422708609,
+      "grad_norm": 0.8144451379776001,
+      "learning_rate": 8e-05,
+      "loss": 1.6946,
+      "num_input_tokens_seen": 182017752,
+      "step": 1290
+    },
+    {
+      "epoch": 0.0792302476707069,
+      "grad_norm": 0.7160412669181824,
+      "learning_rate": 8e-05,
+      "loss": 1.7141,
+      "num_input_tokens_seen": 183405184,
+      "step": 1300
+    },
+    {
+      "epoch": 0.07983971111432772,
+      "grad_norm": 0.8351532220840454,
+      "learning_rate": 8e-05,
+      "loss": 1.6821,
+      "num_input_tokens_seen": 184819236,
+      "step": 1310
+    },
+    {
+      "epoch": 0.08044917455794855,
+      "grad_norm": 0.6285978555679321,
+      "learning_rate": 8e-05,
+      "loss": 1.7392,
+      "num_input_tokens_seen": 186241748,
+      "step": 1320
+    },
+    {
+      "epoch": 0.08105863800156937,
+      "grad_norm": 0.6861914992332458,
+      "learning_rate": 8e-05,
+      "loss": 1.6637,
+      "num_input_tokens_seen": 187651796,
+      "step": 1330
+    },
+    {
+      "epoch": 0.0816681014451902,
+      "grad_norm": 0.7372764945030212,
+      "learning_rate": 8e-05,
+      "loss": 1.6151,
+      "num_input_tokens_seen": 189052704,
+      "step": 1340
+    },
+    {
+      "epoch": 0.08227756488881101,
+      "grad_norm": 0.754943311214447,
+      "learning_rate": 8e-05,
+      "loss": 1.6771,
+      "num_input_tokens_seen": 190457440,
+      "step": 1350
+    },
+    {
+      "epoch": 0.08288702833243183,
+      "grad_norm": 0.7047508358955383,
+      "learning_rate": 8e-05,
+      "loss": 1.7488,
+      "num_input_tokens_seen": 191855300,
+      "step": 1360
+    },
+    {
+      "epoch": 0.08349649177605266,
+      "grad_norm": 0.7752687931060791,
+      "learning_rate": 8e-05,
+      "loss": 1.7385,
+      "num_input_tokens_seen": 193237128,
+      "step": 1370
+    },
+    {
+      "epoch": 0.08410595521967348,
+      "grad_norm": 0.7014668583869934,
+      "learning_rate": 8e-05,
+      "loss": 1.767,
+      "num_input_tokens_seen": 194589060,
+      "step": 1380
+    },
+    {
+      "epoch": 0.08471541866329431,
+      "grad_norm": 0.6854328513145447,
+      "learning_rate": 8e-05,
+      "loss": 1.7614,
+      "num_input_tokens_seen": 196034960,
+      "step": 1390
+    },
+    {
+      "epoch": 0.08532488210691512,
+      "grad_norm": 0.7649319171905518,
+      "learning_rate": 8e-05,
+      "loss": 1.7496,
+      "num_input_tokens_seen": 197441788,
+      "step": 1400
+    },
+    {
+      "epoch": 0.08593434555053595,
+      "grad_norm": 0.6990752220153809,
+      "learning_rate": 8e-05,
+      "loss": 1.7276,
+      "num_input_tokens_seen": 198816188,
+      "step": 1410
+    },
+    {
+      "epoch": 0.08654380899415677,
+      "grad_norm": 0.6721024513244629,
+      "learning_rate": 8e-05,
+      "loss": 1.675,
+      "num_input_tokens_seen": 200227720,
+      "step": 1420
+    },
+    {
+      "epoch": 0.0871532724377776,
+      "grad_norm": 0.7723363637924194,
+      "learning_rate": 8e-05,
+      "loss": 1.6968,
+      "num_input_tokens_seen": 201640184,
+      "step": 1430
+    },
+    {
+      "epoch": 0.08776273588139841,
+      "grad_norm": 0.7399781942367554,
+      "learning_rate": 8e-05,
+      "loss": 1.6968,
+      "num_input_tokens_seen": 203082784,
+      "step": 1440
+    },
+    {
+      "epoch": 0.08837219932501923,
+      "grad_norm": 0.6369470953941345,
+      "learning_rate": 8e-05,
+      "loss": 1.7369,
+      "num_input_tokens_seen": 204507764,
+      "step": 1450
+    },
+    {
+      "epoch": 0.08898166276864006,
+      "grad_norm": 0.6783959865570068,
+      "learning_rate": 8e-05,
+      "loss": 1.7148,
+      "num_input_tokens_seen": 205926692,
+      "step": 1460
+    },
+    {
+      "epoch": 0.08959112621226088,
+      "grad_norm": 0.6937606930732727,
+      "learning_rate": 8e-05,
+      "loss": 1.6742,
+      "num_input_tokens_seen": 207292676,
+      "step": 1470
+    },
+    {
+      "epoch": 0.09020058965588171,
+      "grad_norm": 0.7870405316352844,
+      "learning_rate": 8e-05,
+      "loss": 1.6477,
+      "num_input_tokens_seen": 208696852,
+      "step": 1480
+    },
+    {
+      "epoch": 0.09081005309950252,
+      "grad_norm": 0.7263059020042419,
+      "learning_rate": 8e-05,
+      "loss": 1.6267,
+      "num_input_tokens_seen": 210121428,
+      "step": 1490
+    },
+    {
+      "epoch": 0.09141951654312334,
+      "grad_norm": 0.753978967666626,
+      "learning_rate": 8e-05,
+      "loss": 1.6423,
+      "num_input_tokens_seen": 211550144,
+      "step": 1500
+    },
+    {
+      "epoch": 0.09202897998674417,
+      "grad_norm": 0.6854983568191528,
+      "learning_rate": 8e-05,
+      "loss": 1.6713,
+      "num_input_tokens_seen": 212992572,
+      "step": 1510
+    },
+    {
+      "epoch": 0.092638443430365,
+      "grad_norm": 0.7292976975440979,
+      "learning_rate": 8e-05,
+      "loss": 1.6481,
+      "num_input_tokens_seen": 214390412,
+      "step": 1520
+    },
+    {
+      "epoch": 0.09324790687398582,
+      "grad_norm": 0.6837400197982788,
+      "learning_rate": 8e-05,
+      "loss": 1.5692,
+      "num_input_tokens_seen": 215806552,
+      "step": 1530
+    },
+    {
+      "epoch": 0.09385737031760663,
+      "grad_norm": 0.7413853406906128,
+      "learning_rate": 8e-05,
+      "loss": 1.6803,
+      "num_input_tokens_seen": 217210216,
+      "step": 1540
+    },
+    {
+      "epoch": 0.09446683376122746,
+      "grad_norm": 0.7433369755744934,
+      "learning_rate": 8e-05,
+      "loss": 1.6934,
+      "num_input_tokens_seen": 218598456,
+      "step": 1550
+    },
+    {
+      "epoch": 0.09507629720484828,
+      "grad_norm": 0.7748942375183105,
+      "learning_rate": 8e-05,
+      "loss": 1.6741,
+      "num_input_tokens_seen": 219976860,
+      "step": 1560
+    },
+    {
+      "epoch": 0.0956857606484691,
+      "grad_norm": 0.7293086647987366,
+      "learning_rate": 8e-05,
+      "loss": 1.7087,
+      "num_input_tokens_seen": 221355396,
+      "step": 1570
+    },
+    {
+      "epoch": 0.09629522409208993,
+      "grad_norm": 0.862250566482544,
+      "learning_rate": 8e-05,
+      "loss": 1.6832,
+      "num_input_tokens_seen": 222791252,
+      "step": 1580
+    },
+    {
+      "epoch": 0.09690468753571074,
+      "grad_norm": 0.7801169157028198,
+      "learning_rate": 8e-05,
+      "loss": 1.5818,
+      "num_input_tokens_seen": 224182892,
+      "step": 1590
+    },
+    {
+      "epoch": 0.09751415097933157,
+      "grad_norm": 0.7509076595306396,
+      "learning_rate": 8e-05,
+      "loss": 1.6964,
+      "num_input_tokens_seen": 225596820,
+      "step": 1600
+    },
+    {
+      "epoch": 0.0981236144229524,
+      "grad_norm": 0.6174609661102295,
+      "learning_rate": 8e-05,
+      "loss": 1.6795,
+      "num_input_tokens_seen": 227002192,
+      "step": 1610
+    },
+    {
+      "epoch": 0.09873307786657322,
+      "grad_norm": 0.7319700717926025,
+      "learning_rate": 8e-05,
+      "loss": 1.5898,
+      "num_input_tokens_seen": 228388248,
+      "step": 1620
+    },
+    {
+      "epoch": 0.09934254131019404,
+      "grad_norm": 0.7365676164627075,
+      "learning_rate": 8e-05,
+      "loss": 1.6146,
+      "num_input_tokens_seen": 229791408,
+      "step": 1630
+    },
+    {
+      "epoch": 0.09995200475381485,
+      "grad_norm": 0.7724565267562866,
+      "learning_rate": 8e-05,
+      "loss": 1.638,
+      "num_input_tokens_seen": 231187424,
+      "step": 1640
+    },
+    {
+      "epoch": 0.10056146819743568,
+      "grad_norm": 0.7245753407478333,
+      "learning_rate": 8e-05,
+      "loss": 1.712,
+      "num_input_tokens_seen": 232600616,
+      "step": 1650
+    },
+    {
+      "epoch": 0.1011709316410565,
+      "grad_norm": 0.6499654650688171,
+      "learning_rate": 8e-05,
+      "loss": 1.6265,
+      "num_input_tokens_seen": 234006344,
+      "step": 1660
+    },
+    {
+      "epoch": 0.10178039508467733,
+      "grad_norm": 0.6563583612442017,
+      "learning_rate": 8e-05,
+      "loss": 1.664,
+      "num_input_tokens_seen": 235416748,
+      "step": 1670
+    },
+    {
+      "epoch": 0.10238985852829816,
+      "grad_norm": 0.7042690515518188,
+      "learning_rate": 8e-05,
+      "loss": 1.6468,
+      "num_input_tokens_seen": 236830896,
+      "step": 1680
+    },
+    {
+      "epoch": 0.10299932197191897,
+      "grad_norm": 0.7631860375404358,
+      "learning_rate": 8e-05,
+      "loss": 1.6325,
+      "num_input_tokens_seen": 238262140,
+      "step": 1690
+    },
+    {
+      "epoch": 0.10360878541553979,
+      "grad_norm": 0.6432304382324219,
+      "learning_rate": 8e-05,
+      "loss": 1.6756,
+      "num_input_tokens_seen": 239674040,
+      "step": 1700
+    },
+    {
+      "epoch": 0.10421824885916062,
+      "grad_norm": 0.7273523807525635,
+      "learning_rate": 8e-05,
+      "loss": 1.6661,
+      "num_input_tokens_seen": 241093432,
+      "step": 1710
+    },
+    {
+      "epoch": 0.10482771230278144,
+      "grad_norm": 0.781891942024231,
+      "learning_rate": 8e-05,
+      "loss": 1.7187,
+      "num_input_tokens_seen": 242530300,
+      "step": 1720
+    },
+    {
+      "epoch": 0.10543717574640227,
+      "grad_norm": 0.6699258089065552,
+      "learning_rate": 8e-05,
+      "loss": 1.6269,
+      "num_input_tokens_seen": 243920956,
+      "step": 1730
+    },
+    {
+      "epoch": 0.10604663919002308,
+      "grad_norm": 0.7054083943367004,
+      "learning_rate": 8e-05,
+      "loss": 1.6728,
+      "num_input_tokens_seen": 245343376,
+      "step": 1740
+    },
+    {
+      "epoch": 0.1066561026336439,
+      "grad_norm": 0.7309428453445435,
+      "learning_rate": 8e-05,
+      "loss": 1.6168,
+      "num_input_tokens_seen": 246742064,
+      "step": 1750
+    },
+    {
+      "epoch": 0.10726556607726473,
+      "grad_norm": 0.6241644620895386,
+      "learning_rate": 8e-05,
+      "loss": 1.6413,
+      "num_input_tokens_seen": 248121468,
+      "step": 1760
+    },
+    {
+      "epoch": 0.10787502952088555,
+      "grad_norm": 0.7994136214256287,
+      "learning_rate": 8e-05,
+      "loss": 1.6438,
+      "num_input_tokens_seen": 249536668,
+      "step": 1770
+    },
+    {
+      "epoch": 0.10848449296450638,
+      "grad_norm": 0.6934658885002136,
+      "learning_rate": 8e-05,
+      "loss": 1.7127,
+      "num_input_tokens_seen": 250943296,
+      "step": 1780
+    },
+    {
+      "epoch": 0.10909395640812719,
+      "grad_norm": 0.7179552912712097,
+      "learning_rate": 8e-05,
+      "loss": 1.6016,
+      "num_input_tokens_seen": 252370756,
+      "step": 1790
+    },
+    {
+      "epoch": 0.10970341985174802,
+      "grad_norm": 0.7161432504653931,
+      "learning_rate": 8e-05,
+      "loss": 1.6472,
+      "num_input_tokens_seen": 253768292,
+      "step": 1800
+    },
+    {
+      "epoch": 0.11031288329536884,
+      "grad_norm": 0.749668300151825,
+      "learning_rate": 8e-05,
+      "loss": 1.645,
+      "num_input_tokens_seen": 255228292,
+      "step": 1810
+    },
+    {
+      "epoch": 0.11092234673898967,
+      "grad_norm": 0.6765767335891724,
+      "learning_rate": 8e-05,
+      "loss": 1.6538,
+      "num_input_tokens_seen": 256611280,
+      "step": 1820
+    },
+    {
+      "epoch": 0.11153181018261048,
+      "grad_norm": 0.6850834488868713,
+      "learning_rate": 8e-05,
+      "loss": 1.7177,
+      "num_input_tokens_seen": 258027728,
+      "step": 1830
+    },
+    {
+      "epoch": 0.1121412736262313,
+      "grad_norm": 0.7252477407455444,
+      "learning_rate": 8e-05,
+      "loss": 1.6964,
+      "num_input_tokens_seen": 259428460,
+      "step": 1840
+    },
+    {
+      "epoch": 0.11275073706985213,
+      "grad_norm": 0.7121752500534058,
+      "learning_rate": 8e-05,
+      "loss": 1.6462,
+      "num_input_tokens_seen": 260851868,
+      "step": 1850
+    },
+    {
+      "epoch": 0.11336020051347295,
+      "grad_norm": 0.6859217882156372,
+      "learning_rate": 8e-05,
+      "loss": 1.6794,
+      "num_input_tokens_seen": 262264644,
+      "step": 1860
+    },
+    {
+      "epoch": 0.11396966395709378,
+      "grad_norm": 0.8705602884292603,
+      "learning_rate": 8e-05,
+      "loss": 1.6131,
+      "num_input_tokens_seen": 263652388,
+      "step": 1870
+    },
+    {
+      "epoch": 0.11457912740071459,
+      "grad_norm": 0.6738168001174927,
+      "learning_rate": 8e-05,
+      "loss": 1.6295,
+      "num_input_tokens_seen": 265074652,
+      "step": 1880
+    },
+    {
+      "epoch": 0.11518859084433541,
+      "grad_norm": 0.7672312259674072,
+      "learning_rate": 8e-05,
+      "loss": 1.7278,
+      "num_input_tokens_seen": 266487960,
+      "step": 1890
+    },
+    {
+      "epoch": 0.11579805428795624,
+      "grad_norm": 0.7259830236434937,
+      "learning_rate": 8e-05,
+      "loss": 1.5629,
+      "num_input_tokens_seen": 267907564,
+      "step": 1900
+    },
+    {
+      "epoch": 0.11640751773157706,
+      "grad_norm": 0.7508255243301392,
+      "learning_rate": 8e-05,
+      "loss": 1.6847,
+      "num_input_tokens_seen": 269339832,
+      "step": 1910
+    },
+    {
+      "epoch": 0.11701698117519789,
+      "grad_norm": 0.5813190340995789,
+      "learning_rate": 8e-05,
+      "loss": 1.6564,
+      "num_input_tokens_seen": 270751000,
+      "step": 1920
+    },
+    {
+      "epoch": 0.1176264446188187,
+      "grad_norm": 0.7213466167449951,
+      "learning_rate": 8e-05,
+      "loss": 1.6302,
+      "num_input_tokens_seen": 272212216,
+      "step": 1930
+    },
+    {
+      "epoch": 0.11823590806243953,
+      "grad_norm": 0.6883914470672607,
+      "learning_rate": 8e-05,
+      "loss": 1.5737,
+      "num_input_tokens_seen": 273691760,
+      "step": 1940
+    },
+    {
+      "epoch": 0.11884537150606035,
+      "grad_norm": 0.7095319628715515,
+      "learning_rate": 8e-05,
+      "loss": 1.6719,
+      "num_input_tokens_seen": 275114348,
+      "step": 1950
+    },
+    {
+      "epoch": 0.11945483494968118,
+      "grad_norm": 0.6590485572814941,
+      "learning_rate": 8e-05,
+      "loss": 1.6855,
+      "num_input_tokens_seen": 276525856,
+      "step": 1960
+    },
+    {
+      "epoch": 0.120064298393302,
+      "grad_norm": 0.7270433306694031,
+      "learning_rate": 8e-05,
+      "loss": 1.6401,
+      "num_input_tokens_seen": 277930536,
+      "step": 1970
+    },
+    {
+      "epoch": 0.12067376183692281,
+      "grad_norm": 0.6123723387718201,
+      "learning_rate": 8e-05,
+      "loss": 1.6706,
+      "num_input_tokens_seen": 279346748,
+      "step": 1980
+    },
+    {
+      "epoch": 0.12128322528054364,
+      "grad_norm": 0.7456077337265015,
+      "learning_rate": 8e-05,
+      "loss": 1.7432,
+      "num_input_tokens_seen": 280734376,
+      "step": 1990
+    },
+    {
+      "epoch": 0.12189268872416446,
+      "grad_norm": 0.6831845641136169,
+      "learning_rate": 8e-05,
+      "loss": 1.6291,
+      "num_input_tokens_seen": 282132752,
+      "step": 2000
+    },
+    {
+      "epoch": 0.12250215216778529,
+      "grad_norm": 0.6901050209999084,
+      "learning_rate": 8e-05,
+      "loss": 1.6807,
+      "num_input_tokens_seen": 283584816,
+      "step": 2010
+    },
+    {
+      "epoch": 0.12311161561140611,
+      "grad_norm": 0.7861285209655762,
+      "learning_rate": 8e-05,
+      "loss": 1.7389,
+      "num_input_tokens_seen": 285064504,
+      "step": 2020
+    },
+    {
+      "epoch": 0.12372107905502693,
+      "grad_norm": 0.5943942070007324,
+      "learning_rate": 8e-05,
+      "loss": 1.6045,
+      "num_input_tokens_seen": 286464552,
+      "step": 2030
+    },
+    {
+      "epoch": 0.12433054249864775,
+      "grad_norm": 0.7200583815574646,
+      "learning_rate": 8e-05,
+      "loss": 1.6617,
+      "num_input_tokens_seen": 287842704,
+      "step": 2040
+    },
+    {
+      "epoch": 0.12494000594226858,
+      "grad_norm": 0.6980604529380798,
+      "learning_rate": 8e-05,
+      "loss": 1.5976,
+      "num_input_tokens_seen": 289280200,
+      "step": 2050
+    },
+    {
+      "epoch": 0.1255494693858894,
+      "grad_norm": 0.6354398727416992,
+      "learning_rate": 8e-05,
+      "loss": 1.6191,
+      "num_input_tokens_seen": 290665088,
+      "step": 2060
+    },
+    {
+      "epoch": 0.1261589328295102,
+      "grad_norm": 0.654530942440033,
+      "learning_rate": 8e-05,
+      "loss": 1.6089,
+      "num_input_tokens_seen": 292110316,
+      "step": 2070
+    },
+    {
+      "epoch": 0.12676839627313105,
+      "grad_norm": 0.6449682712554932,
+      "learning_rate": 8e-05,
+      "loss": 1.6043,
+      "num_input_tokens_seen": 293537852,
+      "step": 2080
+    },
+    {
+      "epoch": 0.12737785971675186,
+      "grad_norm": 0.7298957705497742,
+      "learning_rate": 8e-05,
+      "loss": 1.6753,
+      "num_input_tokens_seen": 294958232,
+      "step": 2090
+    },
+    {
+      "epoch": 0.12798732316037267,
+      "grad_norm": 0.7615550756454468,
+      "learning_rate": 8e-05,
+      "loss": 1.5889,
+      "num_input_tokens_seen": 296341104,
+      "step": 2100
+    },
+    {
+      "epoch": 0.1285967866039935,
+      "grad_norm": 0.6716745495796204,
+      "learning_rate": 8e-05,
+      "loss": 1.6452,
+      "num_input_tokens_seen": 297751192,
+      "step": 2110
+    },
+    {
+      "epoch": 0.12920625004761432,
+      "grad_norm": 0.6735673546791077,
+      "learning_rate": 8e-05,
+      "loss": 1.5333,
+      "num_input_tokens_seen": 299157160,
+      "step": 2120
+    },
+    {
+      "epoch": 0.12981571349123516,
+      "grad_norm": 0.6383569836616516,
+      "learning_rate": 8e-05,
+      "loss": 1.7071,
+      "num_input_tokens_seen": 300556440,
+      "step": 2130
+    },
+    {
+      "epoch": 0.13042517693485597,
+      "grad_norm": 0.6580716967582703,
+      "learning_rate": 8e-05,
+      "loss": 1.5882,
+      "num_input_tokens_seen": 301950588,
+      "step": 2140
+    },
+    {
+      "epoch": 0.13103464037847679,
+      "grad_norm": 0.7268714904785156,
+      "learning_rate": 8e-05,
+      "loss": 1.6112,
+      "num_input_tokens_seen": 303350336,
+      "step": 2150
+    },
+    {
+      "epoch": 0.13164410382209762,
+      "grad_norm": 0.6901399493217468,
+      "learning_rate": 8e-05,
+      "loss": 1.572,
+      "num_input_tokens_seen": 304781136,
+      "step": 2160
+    },
+    {
+      "epoch": 0.13225356726571844,
+      "grad_norm": 0.6628725528717041,
+      "learning_rate": 8e-05,
+      "loss": 1.6983,
+      "num_input_tokens_seen": 306176884,
+      "step": 2170
+    },
+    {
+      "epoch": 0.13286303070933927,
+      "grad_norm": 0.688841700553894,
+      "learning_rate": 8e-05,
+      "loss": 1.6409,
+      "num_input_tokens_seen": 307601844,
+      "step": 2180
+    },
+    {
+      "epoch": 0.1334724941529601,
+      "grad_norm": 0.6499971747398376,
+      "learning_rate": 8e-05,
+      "loss": 1.5889,
+      "num_input_tokens_seen": 308995452,
+      "step": 2190
+    },
+    {
+      "epoch": 0.1340819575965809,
+      "grad_norm": 0.6463174223899841,
+      "learning_rate": 8e-05,
+      "loss": 1.6411,
+      "num_input_tokens_seen": 310417804,
+      "step": 2200
+    },
+    {
+      "epoch": 0.13469142104020174,
+      "grad_norm": 0.6206100583076477,
+      "learning_rate": 8e-05,
+      "loss": 1.59,
+      "num_input_tokens_seen": 311846032,
+      "step": 2210
+    },
+    {
+      "epoch": 0.13530088448382255,
+      "grad_norm": 0.6116955876350403,
+      "learning_rate": 8e-05,
+      "loss": 1.6057,
+      "num_input_tokens_seen": 313234256,
+      "step": 2220
+    },
+    {
+      "epoch": 0.1359103479274434,
+      "grad_norm": 0.7880204916000366,
+      "learning_rate": 8e-05,
+      "loss": 1.6058,
+      "num_input_tokens_seen": 314639172,
+      "step": 2230
+    },
+    {
+      "epoch": 0.1365198113710642,
+      "grad_norm": 0.6306136846542358,
+      "learning_rate": 8e-05,
+      "loss": 1.665,
+      "num_input_tokens_seen": 316031736,
+      "step": 2240
+    },
+    {
+      "epoch": 0.137129274814685,
+      "grad_norm": 0.6810283660888672,
+      "learning_rate": 8e-05,
+      "loss": 1.6544,
+      "num_input_tokens_seen": 317438412,
+      "step": 2250
+    },
+    {
+      "epoch": 0.13773873825830585,
+      "grad_norm": 0.671161949634552,
+      "learning_rate": 8e-05,
+      "loss": 1.6091,
+      "num_input_tokens_seen": 318863568,
+      "step": 2260
+    },
+    {
+      "epoch": 0.13834820170192666,
+      "grad_norm": 0.6195483803749084,
+      "learning_rate": 8e-05,
+      "loss": 1.5676,
+      "num_input_tokens_seen": 320256312,
+      "step": 2270
+    },
+    {
+      "epoch": 0.1389576651455475,
+      "grad_norm": 0.6664822101593018,
+      "learning_rate": 8e-05,
+      "loss": 1.6302,
+      "num_input_tokens_seen": 321670092,
+      "step": 2280
+    },
+    {
+      "epoch": 0.1395671285891683,
+      "grad_norm": 0.6218630075454712,
+      "learning_rate": 8e-05,
+      "loss": 1.6288,
+      "num_input_tokens_seen": 323080412,
+      "step": 2290
+    },
+    {
+      "epoch": 0.14017659203278912,
+      "grad_norm": 0.676080048084259,
+      "learning_rate": 8e-05,
+      "loss": 1.5945,
+      "num_input_tokens_seen": 324469600,
+      "step": 2300
+    },
+    {
+      "epoch": 0.14078605547640996,
+      "grad_norm": 0.5576323866844177,
+      "learning_rate": 8e-05,
+      "loss": 1.5904,
+      "num_input_tokens_seen": 325857624,
+      "step": 2310
+    },
+    {
+      "epoch": 0.14139551892003077,
+      "grad_norm": 0.6566000580787659,
+      "learning_rate": 8e-05,
+      "loss": 1.6492,
+      "num_input_tokens_seen": 327258636,
+      "step": 2320
+    },
+    {
+      "epoch": 0.1420049823636516,
+      "grad_norm": 0.6909166574478149,
+      "learning_rate": 8e-05,
+      "loss": 1.6071,
+      "num_input_tokens_seen": 328702764,
+      "step": 2330
+    },
+    {
+      "epoch": 0.14261444580727242,
+      "grad_norm": 0.647297739982605,
+      "learning_rate": 8e-05,
+      "loss": 1.5577,
+      "num_input_tokens_seen": 330100304,
+      "step": 2340
+    },
+    {
+      "epoch": 0.14322390925089323,
+      "grad_norm": 0.6550387740135193,
+      "learning_rate": 8e-05,
+      "loss": 1.6879,
+      "num_input_tokens_seen": 331549276,
+      "step": 2350
+    },
+    {
+      "epoch": 0.14383337269451407,
+      "grad_norm": 0.6743698716163635,
+      "learning_rate": 8e-05,
+      "loss": 1.7019,
+      "num_input_tokens_seen": 332916584,
+      "step": 2360
+    },
+    {
+      "epoch": 0.14444283613813488,
+      "grad_norm": 0.581924557685852,
+      "learning_rate": 8e-05,
+      "loss": 1.6591,
+      "num_input_tokens_seen": 334299004,
+      "step": 2370
+    },
+    {
+      "epoch": 0.14505229958175572,
+      "grad_norm": 0.6334020495414734,
+      "learning_rate": 8e-05,
+      "loss": 1.6316,
+      "num_input_tokens_seen": 335702312,
+      "step": 2380
+    },
+    {
+      "epoch": 0.14566176302537653,
+      "grad_norm": 0.5923337340354919,
+      "learning_rate": 8e-05,
+      "loss": 1.5888,
+      "num_input_tokens_seen": 337127304,
+      "step": 2390
+    },
+    {
+      "epoch": 0.14627122646899735,
+      "grad_norm": 0.6918947696685791,
+      "learning_rate": 8e-05,
+      "loss": 1.6433,
+      "num_input_tokens_seen": 338550764,
+      "step": 2400
+    },
+    {
+      "epoch": 0.14688068991261818,
+      "grad_norm": 0.7656240463256836,
+      "learning_rate": 8e-05,
+      "loss": 1.5949,
+      "num_input_tokens_seen": 339962352,
+      "step": 2410
+    },
+    {
+      "epoch": 0.147490153356239,
+      "grad_norm": 0.7092509269714355,
+      "learning_rate": 8e-05,
+      "loss": 1.6059,
+      "num_input_tokens_seen": 341363788,
+      "step": 2420
+    },
+    {
+      "epoch": 0.14809961679985983,
+      "grad_norm": 0.6121082901954651,
+      "learning_rate": 8e-05,
+      "loss": 1.5637,
+      "num_input_tokens_seen": 342800568,
+      "step": 2430
+    },
+    {
+      "epoch": 0.14870908024348065,
+      "grad_norm": 0.5944322347640991,
+      "learning_rate": 8e-05,
+      "loss": 1.6384,
+      "num_input_tokens_seen": 344226896,
+      "step": 2440
+    },
+    {
+      "epoch": 0.14931854368710146,
+      "grad_norm": 0.6050378084182739,
+      "learning_rate": 8e-05,
+      "loss": 1.6027,
+      "num_input_tokens_seen": 345639348,
+      "step": 2450
+    },
+    {
+      "epoch": 0.1499280071307223,
+      "grad_norm": 0.7982641458511353,
+      "learning_rate": 8e-05,
+      "loss": 1.5957,
+      "num_input_tokens_seen": 347053896,
+      "step": 2460
+    },
+    {
+      "epoch": 0.1505374705743431,
+      "grad_norm": 0.6125765442848206,
+      "learning_rate": 8e-05,
+      "loss": 1.576,
+      "num_input_tokens_seen": 348438848,
+      "step": 2470
+    },
+    {
+      "epoch": 0.15114693401796395,
+      "grad_norm": 0.728951096534729,
+      "learning_rate": 8e-05,
+      "loss": 1.6711,
+      "num_input_tokens_seen": 349823624,
+      "step": 2480
+    },
+    {
+      "epoch": 0.15175639746158476,
+      "grad_norm": 0.5719558596611023,
+      "learning_rate": 8e-05,
+      "loss": 1.6289,
+      "num_input_tokens_seen": 351265596,
+      "step": 2490
+    },
+    {
+      "epoch": 0.15236586090520557,
+      "grad_norm": 0.5965576767921448,
+      "learning_rate": 8e-05,
+      "loss": 1.6114,
+      "num_input_tokens_seen": 352677476,
+      "step": 2500
+    },
+    {
+      "epoch": 0.1529753243488264,
+      "grad_norm": 0.6597477197647095,
+      "learning_rate": 8e-05,
+      "loss": 1.5929,
+      "num_input_tokens_seen": 354079432,
+      "step": 2510
+    },
+    {
+      "epoch": 0.15358478779244722,
+      "grad_norm": 0.633777379989624,
+      "learning_rate": 8e-05,
+      "loss": 1.6062,
+      "num_input_tokens_seen": 355500376,
+      "step": 2520
+    },
+    {
+      "epoch": 0.15419425123606806,
+      "grad_norm": 0.6986291408538818,
+      "learning_rate": 8e-05,
+      "loss": 1.5994,
+      "num_input_tokens_seen": 356938480,
+      "step": 2530
+    },
+    {
+      "epoch": 0.15480371467968887,
+      "grad_norm": 0.6283817291259766,
+      "learning_rate": 8e-05,
+      "loss": 1.6403,
+      "num_input_tokens_seen": 358347276,
+      "step": 2540
+    },
+    {
+      "epoch": 0.15541317812330968,
+      "grad_norm": 0.6436857581138611,
+      "learning_rate": 8e-05,
+      "loss": 1.5707,
+      "num_input_tokens_seen": 359743248,
+      "step": 2550
+    },
+    {
+      "epoch": 0.15602264156693052,
+      "grad_norm": 0.6602137684822083,
+      "learning_rate": 8e-05,
+      "loss": 1.6227,
+      "num_input_tokens_seen": 361133912,
+      "step": 2560
+    },
+    {
+      "epoch": 0.15663210501055133,
+      "grad_norm": 0.7311964631080627,
+      "learning_rate": 8e-05,
+      "loss": 1.5496,
+      "num_input_tokens_seen": 362551344,
+      "step": 2570
+    },
+    {
+      "epoch": 0.15724156845417217,
+      "grad_norm": 0.6504777073860168,
+      "learning_rate": 8e-05,
+      "loss": 1.595,
+      "num_input_tokens_seen": 363943344,
+      "step": 2580
+    },
+    {
+      "epoch": 0.15785103189779298,
+      "grad_norm": 0.6566677093505859,
+      "learning_rate": 8e-05,
+      "loss": 1.5807,
+      "num_input_tokens_seen": 365347512,
+      "step": 2590
+    },
+    {
+      "epoch": 0.1584604953414138,
+      "grad_norm": 0.6459395885467529,
+      "learning_rate": 8e-05,
+      "loss": 1.6358,
+      "num_input_tokens_seen": 366720924,
+      "step": 2600
+    },
+    {
+      "epoch": 0.15906995878503463,
+      "grad_norm": 0.6944287419319153,
+      "learning_rate": 8e-05,
+      "loss": 1.586,
+      "num_input_tokens_seen": 368150964,
+      "step": 2610
+    },
+    {
+      "epoch": 0.15967942222865544,
+      "grad_norm": 0.5620209574699402,
+      "learning_rate": 8e-05,
+      "loss": 1.5993,
+      "num_input_tokens_seen": 369511972,
+      "step": 2620
+    },
+    {
+      "epoch": 0.16028888567227628,
+      "grad_norm": 0.6273583173751831,
+      "learning_rate": 8e-05,
+      "loss": 1.5833,
+      "num_input_tokens_seen": 370942616,
+      "step": 2630
+    },
+    {
+      "epoch": 0.1608983491158971,
+      "grad_norm": 0.669262707233429,
+      "learning_rate": 8e-05,
+      "loss": 1.5427,
+      "num_input_tokens_seen": 372347592,
+      "step": 2640
+    },
+    {
+      "epoch": 0.1615078125595179,
+      "grad_norm": 0.5964462161064148,
+      "learning_rate": 8e-05,
+      "loss": 1.6693,
+      "num_input_tokens_seen": 373729360,
+      "step": 2650
+    },
+    {
+      "epoch": 0.16211727600313874,
+      "grad_norm": 0.608135461807251,
+      "learning_rate": 8e-05,
+      "loss": 1.5437,
+      "num_input_tokens_seen": 375181280,
+      "step": 2660
+    },
+    {
+      "epoch": 0.16272673944675956,
+      "grad_norm": 0.6479006409645081,
+      "learning_rate": 8e-05,
+      "loss": 1.6328,
+      "num_input_tokens_seen": 376596372,
+      "step": 2670
+    },
+    {
+      "epoch": 0.1633362028903804,
+      "grad_norm": 0.6235690712928772,
+      "learning_rate": 8e-05,
+      "loss": 1.6596,
+      "num_input_tokens_seen": 378020808,
+      "step": 2680
+    },
+    {
+      "epoch": 0.1639456663340012,
+      "grad_norm": 0.6381850838661194,
+      "learning_rate": 8e-05,
+      "loss": 1.5794,
+      "num_input_tokens_seen": 379380024,
+      "step": 2690
+    },
+    {
+      "epoch": 0.16455512977762202,
+      "grad_norm": 0.6151752471923828,
+      "learning_rate": 8e-05,
+      "loss": 1.5792,
+      "num_input_tokens_seen": 380803184,
+      "step": 2700
+    },
+    {
+      "epoch": 0.16516459322124286,
+      "grad_norm": 0.6571025252342224,
+      "learning_rate": 8e-05,
+      "loss": 1.5779,
+      "num_input_tokens_seen": 382188904,
+      "step": 2710
+    },
+    {
+      "epoch": 0.16577405666486367,
+      "grad_norm": 0.694426417350769,
+      "learning_rate": 8e-05,
+      "loss": 1.6054,
+      "num_input_tokens_seen": 383612672,
+      "step": 2720
+    },
+    {
+      "epoch": 0.1663835201084845,
+      "grad_norm": 0.6907577514648438,
+      "learning_rate": 8e-05,
+      "loss": 1.5726,
+      "num_input_tokens_seen": 385033960,
+      "step": 2730
+    },
+    {
+      "epoch": 0.16699298355210532,
+      "grad_norm": 0.7059677243232727,
+      "learning_rate": 8e-05,
+      "loss": 1.5798,
+      "num_input_tokens_seen": 386416964,
+      "step": 2740
+    },
+    {
+      "epoch": 0.16760244699572613,
+      "grad_norm": 0.6744717359542847,
+      "learning_rate": 8e-05,
+      "loss": 1.6229,
+      "num_input_tokens_seen": 387838080,
+      "step": 2750
+    },
+    {
+      "epoch": 0.16821191043934697,
+      "grad_norm": 0.6473004817962646,
+      "learning_rate": 8e-05,
+      "loss": 1.5853,
+      "num_input_tokens_seen": 389250364,
+      "step": 2760
+    },
+    {
+      "epoch": 0.16882137388296778,
+      "grad_norm": 0.7390977144241333,
+      "learning_rate": 8e-05,
+      "loss": 1.5684,
+      "num_input_tokens_seen": 390679648,
+      "step": 2770
+    },
+    {
+      "epoch": 0.16943083732658862,
+      "grad_norm": 0.7005506753921509,
+      "learning_rate": 8e-05,
+      "loss": 1.6613,
+      "num_input_tokens_seen": 392061336,
+      "step": 2780
+    },
+    {
+      "epoch": 0.17004030077020943,
+      "grad_norm": 0.6911351084709167,
+      "learning_rate": 8e-05,
+      "loss": 1.5572,
+      "num_input_tokens_seen": 393461980,
+      "step": 2790
+    },
+    {
+      "epoch": 0.17064976421383024,
+      "grad_norm": 0.5468894243240356,
+      "learning_rate": 8e-05,
+      "loss": 1.589,
+      "num_input_tokens_seen": 394847452,
+      "step": 2800
+    },
+    {
+      "epoch": 0.17125922765745108,
+      "grad_norm": 0.5998040437698364,
+      "learning_rate": 8e-05,
+      "loss": 1.5748,
+      "num_input_tokens_seen": 396233508,
+      "step": 2810
+    },
+    {
+      "epoch": 0.1718686911010719,
+      "grad_norm": 0.7814628481864929,
+      "learning_rate": 8e-05,
+      "loss": 1.5206,
+      "num_input_tokens_seen": 397650544,
+      "step": 2820
+    },
+    {
+      "epoch": 0.17247815454469273,
+      "grad_norm": 0.6909236311912537,
+      "learning_rate": 8e-05,
+      "loss": 1.5474,
+      "num_input_tokens_seen": 399046560,
+      "step": 2830
+    },
+    {
+      "epoch": 0.17308761798831354,
+      "grad_norm": 0.7126919627189636,
+      "learning_rate": 8e-05,
+      "loss": 1.5615,
+      "num_input_tokens_seen": 400408700,
+      "step": 2840
+    },
+    {
+      "epoch": 0.17369708143193435,
+      "grad_norm": 0.6222878098487854,
+      "learning_rate": 8e-05,
+      "loss": 1.5642,
+      "num_input_tokens_seen": 401806840,
+      "step": 2850
+    },
+    {
+      "epoch": 0.1743065448755552,
+      "grad_norm": 0.6794285774230957,
+      "learning_rate": 8e-05,
+      "loss": 1.5793,
+      "num_input_tokens_seen": 403203952,
+      "step": 2860
+    },
+    {
+      "epoch": 0.174916008319176,
+      "grad_norm": 0.609735906124115,
+      "learning_rate": 8e-05,
+      "loss": 1.6068,
+      "num_input_tokens_seen": 404591240,
+      "step": 2870
+    },
+    {
+      "epoch": 0.17552547176279681,
+      "grad_norm": 0.6210402250289917,
+      "learning_rate": 8e-05,
+      "loss": 1.6045,
+      "num_input_tokens_seen": 405971452,
+      "step": 2880
+    },
+    {
+      "epoch": 0.17613493520641765,
+      "grad_norm": 0.6104692816734314,
+      "learning_rate": 8e-05,
+      "loss": 1.5263,
+      "num_input_tokens_seen": 407388256,
+      "step": 2890
+    },
+    {
+      "epoch": 0.17674439865003846,
+      "grad_norm": 0.6604185700416565,
+      "learning_rate": 8e-05,
+      "loss": 1.6198,
+      "num_input_tokens_seen": 408813112,
+      "step": 2900
+    },
+    {
+      "epoch": 0.1773538620936593,
+      "grad_norm": 0.6400834321975708,
+      "learning_rate": 8e-05,
+      "loss": 1.5672,
+      "num_input_tokens_seen": 410235500,
+      "step": 2910
+    },
+    {
+      "epoch": 0.17796332553728011,
+      "grad_norm": 0.6400529742240906,
+      "learning_rate": 8e-05,
+      "loss": 1.5761,
+      "num_input_tokens_seen": 411639812,
+      "step": 2920
+    },
+    {
+      "epoch": 0.17857278898090093,
+      "grad_norm": 0.603792130947113,
+      "learning_rate": 8e-05,
+      "loss": 1.6047,
+      "num_input_tokens_seen": 413089172,
+      "step": 2930
+    },
+    {
+      "epoch": 0.17918225242452177,
+      "grad_norm": 0.6779669523239136,
+      "learning_rate": 8e-05,
+      "loss": 1.5808,
+      "num_input_tokens_seen": 414491152,
+      "step": 2940
+    },
+    {
+      "epoch": 0.17979171586814258,
+      "grad_norm": 0.6752368807792664,
+      "learning_rate": 8e-05,
+      "loss": 1.5736,
+      "num_input_tokens_seen": 415907296,
+      "step": 2950
+    },
+    {
+      "epoch": 0.18040117931176342,
+      "grad_norm": 0.6246203184127808,
+      "learning_rate": 8e-05,
+      "loss": 1.5875,
+      "num_input_tokens_seen": 417303208,
+      "step": 2960
+    },
+    {
+      "epoch": 0.18101064275538423,
+      "grad_norm": 0.6521744728088379,
+      "learning_rate": 8e-05,
+      "loss": 1.586,
+      "num_input_tokens_seen": 418730556,
+      "step": 2970
+    },
+    {
+      "epoch": 0.18162010619900504,
+      "grad_norm": 0.6168652176856995,
+      "learning_rate": 8e-05,
+      "loss": 1.541,
+      "num_input_tokens_seen": 420143632,
+      "step": 2980
+    },
+    {
+      "epoch": 0.18222956964262588,
+      "grad_norm": 0.6491835713386536,
+      "learning_rate": 8e-05,
+      "loss": 1.551,
+      "num_input_tokens_seen": 421602604,
+      "step": 2990
+    },
+    {
+      "epoch": 0.1828390330862467,
+      "grad_norm": 0.8011573553085327,
+      "learning_rate": 8e-05,
+      "loss": 1.6447,
+      "num_input_tokens_seen": 423013584,
+      "step": 3000
+    },
+    {
+      "epoch": 0.18344849652986753,
+      "grad_norm": 0.641477644443512,
+      "learning_rate": 8e-05,
+      "loss": 1.5983,
+      "num_input_tokens_seen": 424408696,
+      "step": 3010
+    },
+    {
+      "epoch": 0.18405795997348834,
+      "grad_norm": 0.6406200528144836,
+      "learning_rate": 8e-05,
+      "loss": 1.6416,
+      "num_input_tokens_seen": 425827168,
+      "step": 3020
+    },
+    {
+      "epoch": 0.18466742341710915,
+      "grad_norm": 0.6057642102241516,
+      "learning_rate": 8e-05,
+      "loss": 1.5256,
+      "num_input_tokens_seen": 427245472,
+      "step": 3030
+    },
+    {
+      "epoch": 0.18527688686073,
+      "grad_norm": 0.6526055932044983,
+      "learning_rate": 8e-05,
+      "loss": 1.4959,
+      "num_input_tokens_seen": 428628168,
+      "step": 3040
+    },
+    {
+      "epoch": 0.1858863503043508,
+      "grad_norm": 0.6561000943183899,
+      "learning_rate": 8e-05,
+      "loss": 1.5155,
+      "num_input_tokens_seen": 430033608,
+      "step": 3050
+    },
+    {
+      "epoch": 0.18649581374797164,
+      "grad_norm": 0.6516884565353394,
+      "learning_rate": 8e-05,
+      "loss": 1.6234,
+      "num_input_tokens_seen": 431432912,
+      "step": 3060
+    },
+    {
+      "epoch": 0.18710527719159245,
+      "grad_norm": 0.6904422044754028,
+      "learning_rate": 8e-05,
+      "loss": 1.4835,
+      "num_input_tokens_seen": 432800096,
+      "step": 3070
+    },
+    {
+      "epoch": 0.18771474063521326,
+      "grad_norm": 0.5921774506568909,
+      "learning_rate": 8e-05,
+      "loss": 1.5685,
+      "num_input_tokens_seen": 434190712,
+      "step": 3080
+    },
+    {
+      "epoch": 0.1883242040788341,
+      "grad_norm": 0.5747608542442322,
+      "learning_rate": 8e-05,
+      "loss": 1.6084,
+      "num_input_tokens_seen": 435587468,
+      "step": 3090
+    },
+    {
+      "epoch": 0.1889336675224549,
+      "grad_norm": 0.6614587306976318,
+      "learning_rate": 8e-05,
+      "loss": 1.5626,
+      "num_input_tokens_seen": 437003988,
+      "step": 3100
+    },
+    {
+      "epoch": 0.18954313096607575,
+      "grad_norm": 0.5542994141578674,
+      "learning_rate": 8e-05,
+      "loss": 1.6015,
+      "num_input_tokens_seen": 438403756,
+      "step": 3110
+    },
+    {
+      "epoch": 0.19015259440969656,
+      "grad_norm": 0.6427204608917236,
+      "learning_rate": 8e-05,
+      "loss": 1.529,
+      "num_input_tokens_seen": 439826144,
+      "step": 3120
+    },
+    {
+      "epoch": 0.19076205785331737,
+      "grad_norm": 0.560806930065155,
+      "learning_rate": 8e-05,
+      "loss": 1.5518,
+      "num_input_tokens_seen": 441217644,
+      "step": 3130
+    },
+    {
+      "epoch": 0.1913715212969382,
+      "grad_norm": 0.6509169340133667,
+      "learning_rate": 8e-05,
+      "loss": 1.5269,
+      "num_input_tokens_seen": 442617344,
+      "step": 3140
+    },
+    {
+      "epoch": 0.19198098474055902,
+      "grad_norm": 0.6311179995536804,
+      "learning_rate": 8e-05,
+      "loss": 1.5414,
+      "num_input_tokens_seen": 444028916,
+      "step": 3150
+    },
+    {
+      "epoch": 0.19259044818417986,
+      "grad_norm": 0.6070147752761841,
+      "learning_rate": 8e-05,
+      "loss": 1.5521,
+      "num_input_tokens_seen": 445413608,
+      "step": 3160
+    },
+    {
+      "epoch": 0.19319991162780067,
+      "grad_norm": 0.6126248836517334,
+      "learning_rate": 8e-05,
+      "loss": 1.6203,
+      "num_input_tokens_seen": 446824748,
+      "step": 3170
+    },
+    {
+      "epoch": 0.19380937507142149,
+      "grad_norm": 0.6572065949440002,
+      "learning_rate": 8e-05,
+      "loss": 1.56,
+      "num_input_tokens_seen": 448217780,
+      "step": 3180
+    },
+    {
+      "epoch": 0.19441883851504232,
+      "grad_norm": 0.6731168627738953,
+      "learning_rate": 8e-05,
+      "loss": 1.6071,
+      "num_input_tokens_seen": 449669388,
+      "step": 3190
+    },
+    {
+      "epoch": 0.19502830195866314,
+      "grad_norm": 0.5746667385101318,
+      "learning_rate": 8e-05,
+      "loss": 1.6396,
+      "num_input_tokens_seen": 451076256,
+      "step": 3200
+    },
+    {
+      "epoch": 0.19563776540228398,
+      "grad_norm": 0.5883861184120178,
+      "learning_rate": 8e-05,
+      "loss": 1.5441,
+      "num_input_tokens_seen": 452433176,
+      "step": 3210
+    },
+    {
+      "epoch": 0.1962472288459048,
+      "grad_norm": 0.6213635206222534,
+      "learning_rate": 8e-05,
+      "loss": 1.5408,
+      "num_input_tokens_seen": 453864744,
+      "step": 3220
+    },
+    {
+      "epoch": 0.1968566922895256,
+      "grad_norm": 0.5844454169273376,
+      "learning_rate": 8e-05,
+      "loss": 1.5892,
+      "num_input_tokens_seen": 455266916,
+      "step": 3230
+    },
+    {
+      "epoch": 0.19746615573314644,
+      "grad_norm": 0.7327582240104675,
+      "learning_rate": 8e-05,
+      "loss": 1.5541,
+      "num_input_tokens_seen": 456655360,
+      "step": 3240
+    },
+    {
+      "epoch": 0.19807561917676725,
+      "grad_norm": 0.7098453044891357,
+      "learning_rate": 8e-05,
+      "loss": 1.5219,
+      "num_input_tokens_seen": 458100268,
+      "step": 3250
+    },
+    {
+      "epoch": 0.1986850826203881,
+      "grad_norm": 0.6221525073051453,
+      "learning_rate": 8e-05,
+      "loss": 1.5587,
+      "num_input_tokens_seen": 459523316,
+      "step": 3260
+    },
+    {
+      "epoch": 0.1992945460640089,
+      "grad_norm": 0.646511971950531,
+      "learning_rate": 8e-05,
+      "loss": 1.6495,
+      "num_input_tokens_seen": 460960464,
+      "step": 3270
+    },
+    {
+      "epoch": 0.1999040095076297,
+      "grad_norm": 0.6622752547264099,
+      "learning_rate": 8e-05,
+      "loss": 1.5559,
+      "num_input_tokens_seen": 462371980,
+      "step": 3280
+    },
+    {
+      "epoch": 0.20051347295125055,
+      "grad_norm": 0.5687728524208069,
+      "learning_rate": 8e-05,
+      "loss": 1.5691,
+      "num_input_tokens_seen": 463756352,
+      "step": 3290
+    },
+    {
+      "epoch": 0.20112293639487136,
+      "grad_norm": 0.6098371148109436,
+      "learning_rate": 8e-05,
+      "loss": 1.5524,
+      "num_input_tokens_seen": 465193884,
+      "step": 3300
+    },
+    {
+      "epoch": 0.2017323998384922,
+      "grad_norm": 0.681311309337616,
+      "learning_rate": 8e-05,
+      "loss": 1.5381,
+      "num_input_tokens_seen": 466639236,
+      "step": 3310
+    },
+    {
+      "epoch": 0.202341863282113,
+      "grad_norm": 0.5747373104095459,
+      "learning_rate": 8e-05,
+      "loss": 1.5177,
+      "num_input_tokens_seen": 468072688,
+      "step": 3320
+    },
+    {
+      "epoch": 0.20295132672573382,
+      "grad_norm": 0.5559400320053101,
+      "learning_rate": 8e-05,
+      "loss": 1.5873,
+      "num_input_tokens_seen": 469442532,
+      "step": 3330
+    },
+    {
+      "epoch": 0.20356079016935466,
+      "grad_norm": 0.6499312520027161,
+      "learning_rate": 8e-05,
+      "loss": 1.6033,
+      "num_input_tokens_seen": 470904716,
+      "step": 3340
+    },
+    {
+      "epoch": 0.20417025361297547,
+      "grad_norm": 0.5671247839927673,
+      "learning_rate": 8e-05,
+      "loss": 1.546,
+      "num_input_tokens_seen": 472320820,
+      "step": 3350
+    },
+    {
+      "epoch": 0.2047797170565963,
+      "grad_norm": 0.605981171131134,
+      "learning_rate": 8e-05,
+      "loss": 1.5692,
+      "num_input_tokens_seen": 473731260,
+      "step": 3360
+    },
+    {
+      "epoch": 0.20538918050021712,
+      "grad_norm": 0.7082146406173706,
+      "learning_rate": 8e-05,
+      "loss": 1.502,
+      "num_input_tokens_seen": 475164164,
+      "step": 3370
+    },
+    {
+      "epoch": 0.20599864394383793,
+      "grad_norm": 0.6158855557441711,
+      "learning_rate": 8e-05,
+      "loss": 1.5276,
+      "num_input_tokens_seen": 476603328,
+      "step": 3380
+    },
+    {
+      "epoch": 0.20660810738745877,
+      "grad_norm": 0.701566755771637,
+      "learning_rate": 8e-05,
+      "loss": 1.451,
+      "num_input_tokens_seen": 477996612,
+      "step": 3390
+    },
+    {
+      "epoch": 0.20721757083107958,
+      "grad_norm": 0.7635865211486816,
+      "learning_rate": 8e-05,
+      "loss": 1.5975,
+      "num_input_tokens_seen": 479446800,
+      "step": 3400
+    },
+    {
+      "epoch": 0.20782703427470042,
+      "grad_norm": 0.5621991753578186,
+      "learning_rate": 8e-05,
+      "loss": 1.5216,
+      "num_input_tokens_seen": 480863616,
+      "step": 3410
+    },
+    {
+      "epoch": 0.20843649771832123,
+      "grad_norm": 0.6029215455055237,
+      "learning_rate": 8e-05,
+      "loss": 1.671,
+      "num_input_tokens_seen": 482273428,
+      "step": 3420
+    },
+    {
+      "epoch": 0.20904596116194205,
+      "grad_norm": 0.6263468265533447,
+      "learning_rate": 8e-05,
+      "loss": 1.6108,
+      "num_input_tokens_seen": 483698556,
+      "step": 3430
+    },
+    {
+      "epoch": 0.20965542460556288,
+      "grad_norm": 0.6057296991348267,
+      "learning_rate": 8e-05,
+      "loss": 1.5797,
+      "num_input_tokens_seen": 485164152,
+      "step": 3440
+    },
+    {
+      "epoch": 0.2102648880491837,
+      "grad_norm": 0.639839231967926,
+      "learning_rate": 8e-05,
+      "loss": 1.5568,
+      "num_input_tokens_seen": 486564516,
+      "step": 3450
+    },
+    {
+      "epoch": 0.21087435149280453,
+      "grad_norm": 0.6368933320045471,
+      "learning_rate": 8e-05,
+      "loss": 1.5162,
+      "num_input_tokens_seen": 487982336,
+      "step": 3460
+    },
+    {
+      "epoch": 0.21148381493642535,
+      "grad_norm": 0.6618413925170898,
+      "learning_rate": 8e-05,
+      "loss": 1.5737,
+      "num_input_tokens_seen": 489402684,
+      "step": 3470
+    },
+    {
+      "epoch": 0.21209327838004616,
+      "grad_norm": 0.7187027335166931,
+      "learning_rate": 8e-05,
+      "loss": 1.5436,
+      "num_input_tokens_seen": 490822360,
+      "step": 3480
+    },
+    {
+      "epoch": 0.212702741823667,
+      "grad_norm": 0.6799234747886658,
+      "learning_rate": 8e-05,
+      "loss": 1.6655,
+      "num_input_tokens_seen": 492270836,
+      "step": 3490
+    },
+    {
+      "epoch": 0.2133122052672878,
+      "grad_norm": 0.6822348833084106,
+      "learning_rate": 8e-05,
+      "loss": 1.5637,
+      "num_input_tokens_seen": 493667632,
+      "step": 3500
+    },
+    {
+      "epoch": 0.21392166871090865,
+      "grad_norm": 0.5559425950050354,
+      "learning_rate": 8e-05,
+      "loss": 1.6041,
+      "num_input_tokens_seen": 495092028,
+      "step": 3510
+    },
+    {
+      "epoch": 0.21453113215452946,
+      "grad_norm": 0.6507359147071838,
+      "learning_rate": 8e-05,
+      "loss": 1.5193,
+      "num_input_tokens_seen": 496485460,
+      "step": 3520
+    },
+    {
+      "epoch": 0.21514059559815027,
+      "grad_norm": 0.6291355490684509,
+      "learning_rate": 8e-05,
+      "loss": 1.5885,
+      "num_input_tokens_seen": 497903368,
+      "step": 3530
+    },
+    {
+      "epoch": 0.2157500590417711,
+      "grad_norm": 0.6117660999298096,
+      "learning_rate": 8e-05,
+      "loss": 1.583,
+      "num_input_tokens_seen": 499318444,
+      "step": 3540
+    },
+    {
+      "epoch": 0.21635952248539192,
+      "grad_norm": 0.6362354159355164,
+      "learning_rate": 8e-05,
+      "loss": 1.5588,
+      "num_input_tokens_seen": 500692116,
+      "step": 3550
+    },
+    {
+      "epoch": 0.21696898592901276,
+      "grad_norm": 0.7075550556182861,
+      "learning_rate": 8e-05,
+      "loss": 1.6182,
+      "num_input_tokens_seen": 502070960,
+      "step": 3560
+    },
+    {
+      "epoch": 0.21757844937263357,
+      "grad_norm": 0.5841058492660522,
+      "learning_rate": 8e-05,
+      "loss": 1.5713,
+      "num_input_tokens_seen": 503481416,
+      "step": 3570
+    },
+    {
+      "epoch": 0.21818791281625438,
+      "grad_norm": 0.6521745920181274,
+      "learning_rate": 8e-05,
+      "loss": 1.5641,
+      "num_input_tokens_seen": 504920860,
+      "step": 3580
+    },
+    {
+      "epoch": 0.21879737625987522,
+      "grad_norm": 0.527927041053772,
+      "learning_rate": 8e-05,
+      "loss": 1.5431,
+      "num_input_tokens_seen": 506389836,
+      "step": 3590
+    },
+    {
+      "epoch": 0.21940683970349603,
+      "grad_norm": 0.6047568321228027,
+      "learning_rate": 8e-05,
+      "loss": 1.5653,
+      "num_input_tokens_seen": 507827788,
+      "step": 3600
+    },
+    {
+      "epoch": 0.22001630314711687,
+      "grad_norm": 0.5986194610595703,
+      "learning_rate": 8e-05,
+      "loss": 1.5669,
+      "num_input_tokens_seen": 509254352,
+      "step": 3610
+    },
+    {
+      "epoch": 0.22062576659073768,
+      "grad_norm": 0.7015334367752075,
+      "learning_rate": 8e-05,
+      "loss": 1.5154,
+      "num_input_tokens_seen": 510665164,
+      "step": 3620
+    },
+    {
+      "epoch": 0.2212352300343585,
+      "grad_norm": 0.6172800064086914,
+      "learning_rate": 8e-05,
+      "loss": 1.5364,
+      "num_input_tokens_seen": 512088312,
+      "step": 3630
+    },
+    {
+      "epoch": 0.22184469347797933,
+      "grad_norm": 0.7089115381240845,
+      "learning_rate": 8e-05,
+      "loss": 1.644,
+      "num_input_tokens_seen": 513489328,
+      "step": 3640
+    },
+    {
+      "epoch": 0.22245415692160014,
+      "grad_norm": 0.5689125657081604,
+      "learning_rate": 8e-05,
+      "loss": 1.5437,
+      "num_input_tokens_seen": 514882072,
+      "step": 3650
+    },
+    {
+      "epoch": 0.22306362036522095,
+      "grad_norm": 0.610406756401062,
+      "learning_rate": 8e-05,
+      "loss": 1.5451,
+      "num_input_tokens_seen": 516281980,
+      "step": 3660
+    },
+    {
+      "epoch": 0.2236730838088418,
+      "grad_norm": 0.5259641408920288,
+      "learning_rate": 8e-05,
+      "loss": 1.5613,
+      "num_input_tokens_seen": 517696272,
+      "step": 3670
+    },
+    {
+      "epoch": 0.2242825472524626,
+      "grad_norm": 0.6053969264030457,
+      "learning_rate": 8e-05,
+      "loss": 1.5655,
+      "num_input_tokens_seen": 519099176,
+      "step": 3680
+    },
+    {
+      "epoch": 0.22489201069608344,
+      "grad_norm": 0.6376533508300781,
+      "learning_rate": 8e-05,
+      "loss": 1.5481,
+      "num_input_tokens_seen": 520548316,
+      "step": 3690
+    },
+    {
+      "epoch": 0.22550147413970426,
+      "grad_norm": 0.6387814879417419,
+      "learning_rate": 8e-05,
+      "loss": 1.5541,
+      "num_input_tokens_seen": 521977404,
+      "step": 3700
+    },
+    {
+      "epoch": 0.22611093758332507,
+      "grad_norm": 0.5563859343528748,
+      "learning_rate": 8e-05,
+      "loss": 1.5862,
+      "num_input_tokens_seen": 523431332,
+      "step": 3710
+    },
+    {
+      "epoch": 0.2267204010269459,
+      "grad_norm": 0.5293943285942078,
+      "learning_rate": 8e-05,
+      "loss": 1.5249,
+      "num_input_tokens_seen": 524857860,
+      "step": 3720
+    },
+    {
+      "epoch": 0.22732986447056672,
+      "grad_norm": 0.6297414898872375,
+      "learning_rate": 8e-05,
+      "loss": 1.5081,
+      "num_input_tokens_seen": 526236604,
+      "step": 3730
+    },
+    {
+      "epoch": 0.22793932791418756,
+      "grad_norm": 0.7299327850341797,
+      "learning_rate": 8e-05,
+      "loss": 1.5876,
+      "num_input_tokens_seen": 527654036,
+      "step": 3740
+    },
+    {
+      "epoch": 0.22854879135780837,
+      "grad_norm": 0.6514659523963928,
+      "learning_rate": 8e-05,
+      "loss": 1.5381,
+      "num_input_tokens_seen": 529042780,
+      "step": 3750
+    },
+    {
+      "epoch": 0.22915825480142918,
+      "grad_norm": 0.5427886247634888,
+      "learning_rate": 8e-05,
+      "loss": 1.5652,
+      "num_input_tokens_seen": 530471304,
+      "step": 3760
+    },
+    {
+      "epoch": 0.22976771824505002,
+      "grad_norm": 0.6196932196617126,
+      "learning_rate": 8e-05,
+      "loss": 1.5239,
+      "num_input_tokens_seen": 531891692,
+      "step": 3770
+    },
+    {
+      "epoch": 0.23037718168867083,
+      "grad_norm": 0.628355860710144,
+      "learning_rate": 8e-05,
+      "loss": 1.4701,
+      "num_input_tokens_seen": 533308800,
+      "step": 3780
+    },
+    {
+      "epoch": 0.23098664513229167,
+      "grad_norm": 0.5529937744140625,
+      "learning_rate": 8e-05,
+      "loss": 1.5185,
+      "num_input_tokens_seen": 534696500,
+      "step": 3790
+    },
+    {
+      "epoch": 0.23159610857591248,
+      "grad_norm": 0.5507948994636536,
+      "learning_rate": 8e-05,
+      "loss": 1.5384,
+      "num_input_tokens_seen": 536107284,
+      "step": 3800
+    },
+    {
+      "epoch": 0.2322055720195333,
+      "grad_norm": 0.6283483505249023,
+      "learning_rate": 8e-05,
+      "loss": 1.6055,
+      "num_input_tokens_seen": 537501972,
+      "step": 3810
+    },
+    {
+      "epoch": 0.23281503546315413,
+      "grad_norm": 0.6029003858566284,
+      "learning_rate": 8e-05,
+      "loss": 1.4906,
+      "num_input_tokens_seen": 538873180,
+      "step": 3820
+    },
+    {
+      "epoch": 0.23342449890677494,
+      "grad_norm": 0.6179420351982117,
+      "learning_rate": 8e-05,
+      "loss": 1.5218,
+      "num_input_tokens_seen": 540283024,
+      "step": 3830
+    },
+    {
+      "epoch": 0.23403396235039578,
+      "grad_norm": 0.6198109984397888,
+      "learning_rate": 8e-05,
+      "loss": 1.5806,
+      "num_input_tokens_seen": 541699176,
+      "step": 3840
+    },
+    {
+      "epoch": 0.2346434257940166,
+      "grad_norm": 0.6719751954078674,
+      "learning_rate": 8e-05,
+      "loss": 1.4983,
+      "num_input_tokens_seen": 543113424,
+      "step": 3850
+    },
+    {
+      "epoch": 0.2352528892376374,
+      "grad_norm": 0.7113502025604248,
+      "learning_rate": 8e-05,
+      "loss": 1.5353,
+      "num_input_tokens_seen": 544510612,
+      "step": 3860
+    },
+    {
+      "epoch": 0.23586235268125824,
+      "grad_norm": 0.6690021753311157,
+      "learning_rate": 8e-05,
+      "loss": 1.5478,
+      "num_input_tokens_seen": 545973208,
+      "step": 3870
+    },
+    {
+      "epoch": 0.23647181612487905,
+      "grad_norm": 0.5268478989601135,
+      "learning_rate": 8e-05,
+      "loss": 1.5472,
+      "num_input_tokens_seen": 547339972,
+      "step": 3880
+    },
+    {
+      "epoch": 0.2370812795684999,
+      "grad_norm": 0.5470036864280701,
+      "learning_rate": 8e-05,
+      "loss": 1.6029,
+      "num_input_tokens_seen": 548743084,
+      "step": 3890
+    },
+    {
+      "epoch": 0.2376907430121207,
+      "grad_norm": 0.5641947388648987,
+      "learning_rate": 8e-05,
+      "loss": 1.5503,
+      "num_input_tokens_seen": 550127048,
+      "step": 3900
+    },
+    {
+      "epoch": 0.23830020645574151,
+      "grad_norm": 0.6231464147567749,
+      "learning_rate": 8e-05,
+      "loss": 1.518,
+      "num_input_tokens_seen": 551531300,
+      "step": 3910
+    },
+    {
+      "epoch": 0.23890966989936235,
+      "grad_norm": 0.5082821846008301,
+      "learning_rate": 8e-05,
+      "loss": 1.4749,
+      "num_input_tokens_seen": 552910292,
+      "step": 3920
+    },
+    {
+      "epoch": 0.23951913334298316,
+      "grad_norm": 0.7179940938949585,
+      "learning_rate": 8e-05,
+      "loss": 1.511,
+      "num_input_tokens_seen": 554318712,
+      "step": 3930
+    },
+    {
+      "epoch": 0.240128596786604,
+      "grad_norm": 0.6112212538719177,
+      "learning_rate": 8e-05,
+      "loss": 1.4962,
+      "num_input_tokens_seen": 555780652,
+      "step": 3940
+    },
+    {
+      "epoch": 0.24073806023022482,
+      "grad_norm": 0.7789644598960876,
+      "learning_rate": 8e-05,
+      "loss": 1.5523,
+      "num_input_tokens_seen": 557205392,
+      "step": 3950
+    },
+    {
+      "epoch": 0.24134752367384563,
+      "grad_norm": 0.559933066368103,
+      "learning_rate": 8e-05,
+      "loss": 1.5004,
+      "num_input_tokens_seen": 558642004,
+      "step": 3960
+    },
+    {
+      "epoch": 0.24195698711746647,
+      "grad_norm": 0.6398128867149353,
+      "learning_rate": 8e-05,
+      "loss": 1.5336,
+      "num_input_tokens_seen": 560061460,
+      "step": 3970
+    },
+    {
+      "epoch": 0.24256645056108728,
+      "grad_norm": 0.5948666334152222,
+      "learning_rate": 8e-05,
+      "loss": 1.6063,
+      "num_input_tokens_seen": 561476852,
+      "step": 3980
+    },
+    {
+      "epoch": 0.24317591400470812,
+      "grad_norm": 0.5978219509124756,
+      "learning_rate": 8e-05,
+      "loss": 1.535,
+      "num_input_tokens_seen": 562885480,
+      "step": 3990
+    },
+    {
+      "epoch": 0.24378537744832893,
+      "grad_norm": 0.5732299089431763,
+      "learning_rate": 8e-05,
+      "loss": 1.6014,
+      "num_input_tokens_seen": 564296708,
+      "step": 4000
+    },
+    {
+      "epoch": 0.24439484089194974,
+      "grad_norm": 0.5508894324302673,
+      "learning_rate": 8e-05,
+      "loss": 1.5665,
+      "num_input_tokens_seen": 565716316,
+      "step": 4010
+    },
+    {
+      "epoch": 0.24500430433557058,
+      "grad_norm": 0.6127896904945374,
+      "learning_rate": 8e-05,
+      "loss": 1.5016,
+      "num_input_tokens_seen": 567148104,
+      "step": 4020
+    },
+    {
+      "epoch": 0.2456137677791914,
+      "grad_norm": 0.6429306864738464,
+      "learning_rate": 8e-05,
+      "loss": 1.5325,
+      "num_input_tokens_seen": 568546676,
+      "step": 4030
+    },
+    {
+      "epoch": 0.24622323122281223,
+      "grad_norm": 0.5604584813117981,
+      "learning_rate": 8e-05,
+      "loss": 1.5726,
+      "num_input_tokens_seen": 569955020,
+      "step": 4040
+    },
+    {
+      "epoch": 0.24683269466643304,
+      "grad_norm": 0.5747659206390381,
+      "learning_rate": 8e-05,
+      "loss": 1.613,
+      "num_input_tokens_seen": 571360824,
+      "step": 4050
+    },
+    {
+      "epoch": 0.24744215811005385,
+      "grad_norm": 0.6003303527832031,
+      "learning_rate": 8e-05,
+      "loss": 1.4661,
+      "num_input_tokens_seen": 572783056,
+      "step": 4060
+    },
+    {
+      "epoch": 0.2480516215536747,
+      "grad_norm": 0.5869039297103882,
+      "learning_rate": 8e-05,
+      "loss": 1.5553,
+      "num_input_tokens_seen": 574182888,
+      "step": 4070
+    },
+    {
+      "epoch": 0.2486610849972955,
+      "grad_norm": 0.6359077095985413,
+      "learning_rate": 8e-05,
+      "loss": 1.5035,
+      "num_input_tokens_seen": 575590724,
+      "step": 4080
+    },
+    {
+      "epoch": 0.24927054844091634,
+      "grad_norm": 0.6442738175392151,
+      "learning_rate": 8e-05,
+      "loss": 1.5434,
+      "num_input_tokens_seen": 577002740,
+      "step": 4090
+    },
+    {
+      "epoch": 0.24988001188453715,
+      "grad_norm": 0.5803224444389343,
+      "learning_rate": 8e-05,
+      "loss": 1.5408,
+      "num_input_tokens_seen": 578408172,
+      "step": 4100
+    },
+    {
+      "epoch": 0.25048947532815796,
+      "grad_norm": 0.5301817059516907,
+      "learning_rate": 8e-05,
+      "loss": 1.5587,
+      "num_input_tokens_seen": 579811960,
+      "step": 4110
+    },
+    {
+      "epoch": 0.2510989387717788,
+      "grad_norm": 0.5798521637916565,
+      "learning_rate": 8e-05,
+      "loss": 1.5368,
+      "num_input_tokens_seen": 581216492,
+      "step": 4120
+    },
+    {
+      "epoch": 0.25170840221539964,
+      "grad_norm": 0.6303583979606628,
+      "learning_rate": 8e-05,
+      "loss": 1.5264,
+      "num_input_tokens_seen": 582601144,
+      "step": 4130
+    },
+    {
+      "epoch": 0.2523178656590204,
+      "grad_norm": 0.5813618898391724,
+      "learning_rate": 8e-05,
+      "loss": 1.5848,
+      "num_input_tokens_seen": 583981308,
+      "step": 4140
+    },
+    {
+      "epoch": 0.25292732910264126,
+      "grad_norm": 0.6459015607833862,
+      "learning_rate": 8e-05,
+      "loss": 1.5018,
+      "num_input_tokens_seen": 585375584,
+      "step": 4150
+    },
+    {
+      "epoch": 0.2535367925462621,
+      "grad_norm": 0.5455829501152039,
+      "learning_rate": 8e-05,
+      "loss": 1.5464,
+      "num_input_tokens_seen": 586802916,
+      "step": 4160
+    },
+    {
+      "epoch": 0.2541462559898829,
+      "grad_norm": 0.6944795250892639,
+      "learning_rate": 8e-05,
+      "loss": 1.548,
+      "num_input_tokens_seen": 588224368,
+      "step": 4170
+    },
+    {
+      "epoch": 0.2547557194335037,
+      "grad_norm": 0.49821239709854126,
+      "learning_rate": 8e-05,
+      "loss": 1.5018,
+      "num_input_tokens_seen": 589678916,
+      "step": 4180
+    },
+    {
+      "epoch": 0.25536518287712456,
+      "grad_norm": 0.6219947338104248,
+      "learning_rate": 8e-05,
+      "loss": 1.4944,
+      "num_input_tokens_seen": 591107332,
+      "step": 4190
+    },
+    {
+      "epoch": 0.25597464632074535,
+      "grad_norm": 0.6208930611610413,
+      "learning_rate": 8e-05,
+      "loss": 1.5005,
+      "num_input_tokens_seen": 592503952,
+      "step": 4200
+    },
+    {
+      "epoch": 0.2565841097643662,
+      "grad_norm": 0.5741164684295654,
+      "learning_rate": 8e-05,
+      "loss": 1.551,
+      "num_input_tokens_seen": 593961892,
+      "step": 4210
+    },
+    {
+      "epoch": 0.257193573207987,
+      "grad_norm": 0.6542494893074036,
+      "learning_rate": 8e-05,
+      "loss": 1.4703,
+      "num_input_tokens_seen": 595372500,
+      "step": 4220
+    },
+    {
+      "epoch": 0.25780303665160786,
+      "grad_norm": 0.5730968117713928,
+      "learning_rate": 8e-05,
+      "loss": 1.5682,
+      "num_input_tokens_seen": 596779600,
+      "step": 4230
+    },
+    {
+      "epoch": 0.25841250009522865,
+      "grad_norm": 0.683051347732544,
+      "learning_rate": 8e-05,
+      "loss": 1.4536,
+      "num_input_tokens_seen": 598184840,
+      "step": 4240
+    },
+    {
+      "epoch": 0.2590219635388495,
+      "grad_norm": 0.7352629899978638,
+      "learning_rate": 8e-05,
+      "loss": 1.5573,
+      "num_input_tokens_seen": 599586716,
+      "step": 4250
+    },
+    {
+      "epoch": 0.2596314269824703,
+      "grad_norm": 0.5975140929222107,
+      "learning_rate": 8e-05,
+      "loss": 1.4868,
+      "num_input_tokens_seen": 600969532,
+      "step": 4260
+    },
+    {
+      "epoch": 0.2602408904260911,
+      "grad_norm": 0.5328396558761597,
+      "learning_rate": 8e-05,
+      "loss": 1.4971,
+      "num_input_tokens_seen": 602381612,
+      "step": 4270
+    },
+    {
+      "epoch": 0.26085035386971195,
+      "grad_norm": 0.832878053188324,
+      "learning_rate": 8e-05,
+      "loss": 1.5396,
+      "num_input_tokens_seen": 603796776,
+      "step": 4280
+    },
+    {
+      "epoch": 0.2614598173133328,
+      "grad_norm": 0.5859492421150208,
+      "learning_rate": 8e-05,
+      "loss": 1.5363,
+      "num_input_tokens_seen": 605221644,
+      "step": 4290
+    },
+    {
+      "epoch": 0.26206928075695357,
+      "grad_norm": 0.5248042345046997,
+      "learning_rate": 8e-05,
+      "loss": 1.5086,
+      "num_input_tokens_seen": 606642144,
+      "step": 4300
+    },
+    {
+      "epoch": 0.2626787442005744,
+      "grad_norm": 0.5584684014320374,
+      "learning_rate": 8e-05,
+      "loss": 1.5288,
+      "num_input_tokens_seen": 608083892,
+      "step": 4310
+    },
+    {
+      "epoch": 0.26328820764419525,
+      "grad_norm": 0.5786579251289368,
+      "learning_rate": 8e-05,
+      "loss": 1.5388,
+      "num_input_tokens_seen": 609490492,
+      "step": 4320
+    },
+    {
+      "epoch": 0.2638976710878161,
+      "grad_norm": 0.6442510485649109,
+      "learning_rate": 8e-05,
+      "loss": 1.5189,
+      "num_input_tokens_seen": 610881576,
+      "step": 4330
+    },
+    {
+      "epoch": 0.26450713453143687,
+      "grad_norm": 0.5844077467918396,
+      "learning_rate": 8e-05,
+      "loss": 1.4962,
+      "num_input_tokens_seen": 612271752,
+      "step": 4340
+    },
+    {
+      "epoch": 0.2651165979750577,
+      "grad_norm": 0.5478509664535522,
+      "learning_rate": 8e-05,
+      "loss": 1.479,
+      "num_input_tokens_seen": 613667596,
+      "step": 4350
+    },
+    {
+      "epoch": 0.26572606141867855,
+      "grad_norm": 0.5829759836196899,
+      "learning_rate": 8e-05,
+      "loss": 1.5768,
+      "num_input_tokens_seen": 615067476,
+      "step": 4360
+    },
+    {
+      "epoch": 0.26633552486229933,
+      "grad_norm": 0.6037372350692749,
+      "learning_rate": 8e-05,
+      "loss": 1.5158,
+      "num_input_tokens_seen": 616456876,
+      "step": 4370
+    },
+    {
+      "epoch": 0.2669449883059202,
+      "grad_norm": 0.7143189311027527,
+      "learning_rate": 8e-05,
+      "loss": 1.5373,
+      "num_input_tokens_seen": 617885136,
+      "step": 4380
+    },
+    {
+      "epoch": 0.267554451749541,
+      "grad_norm": 0.5982239246368408,
+      "learning_rate": 8e-05,
+      "loss": 1.5993,
+      "num_input_tokens_seen": 619310072,
+      "step": 4390
+    },
+    {
+      "epoch": 0.2681639151931618,
+      "grad_norm": 0.5947147011756897,
+      "learning_rate": 8e-05,
+      "loss": 1.5218,
+      "num_input_tokens_seen": 620729588,
+      "step": 4400
+    },
+    {
+      "epoch": 0.26877337863678263,
+      "grad_norm": 0.6012521386146545,
+      "learning_rate": 8e-05,
+      "loss": 1.4877,
+      "num_input_tokens_seen": 622107468,
+      "step": 4410
+    },
+    {
+      "epoch": 0.2693828420804035,
+      "grad_norm": 0.8092216849327087,
+      "learning_rate": 8e-05,
+      "loss": 1.5039,
+      "num_input_tokens_seen": 623481492,
+      "step": 4420
+    },
+    {
+      "epoch": 0.2699923055240243,
+      "grad_norm": 0.5891050100326538,
+      "learning_rate": 8e-05,
+      "loss": 1.6217,
+      "num_input_tokens_seen": 624886240,
+      "step": 4430
+    },
+    {
+      "epoch": 0.2706017689676451,
+      "grad_norm": 0.5771660804748535,
+      "learning_rate": 8e-05,
+      "loss": 1.464,
+      "num_input_tokens_seen": 626297120,
+      "step": 4440
+    },
+    {
+      "epoch": 0.27121123241126593,
+      "grad_norm": 0.6003730893135071,
+      "learning_rate": 8e-05,
+      "loss": 1.5165,
+      "num_input_tokens_seen": 627731736,
+      "step": 4450
+    },
+    {
+      "epoch": 0.2718206958548868,
+      "grad_norm": 0.545914351940155,
+      "learning_rate": 8e-05,
+      "loss": 1.5915,
+      "num_input_tokens_seen": 629175788,
+      "step": 4460
+    },
+    {
+      "epoch": 0.27243015929850756,
+      "grad_norm": 0.5131253004074097,
+      "learning_rate": 8e-05,
+      "loss": 1.519,
+      "num_input_tokens_seen": 630590732,
+      "step": 4470
+    },
+    {
+      "epoch": 0.2730396227421284,
+      "grad_norm": 0.6029266715049744,
+      "learning_rate": 8e-05,
+      "loss": 1.5594,
+      "num_input_tokens_seen": 632034684,
+      "step": 4480
+    },
+    {
+      "epoch": 0.27364908618574924,
+      "grad_norm": 0.7176979780197144,
+      "learning_rate": 8e-05,
+      "loss": 1.4626,
+      "num_input_tokens_seen": 633439316,
+      "step": 4490
+    },
+    {
+      "epoch": 0.27425854962937,
+      "grad_norm": 0.5998932719230652,
+      "learning_rate": 8e-05,
+      "loss": 1.4572,
+      "num_input_tokens_seen": 634866124,
+      "step": 4500
+    },
+    {
+      "epoch": 0.27486801307299086,
+      "grad_norm": 0.6117346882820129,
+      "learning_rate": 8e-05,
+      "loss": 1.4799,
+      "num_input_tokens_seen": 636287180,
+      "step": 4510
+    },
+    {
+      "epoch": 0.2754774765166117,
+      "grad_norm": 0.6476706862449646,
+      "learning_rate": 8e-05,
+      "loss": 1.4876,
+      "num_input_tokens_seen": 637660604,
+      "step": 4520
+    },
+    {
+      "epoch": 0.27608693996023254,
+      "grad_norm": 0.5890971422195435,
+      "learning_rate": 8e-05,
+      "loss": 1.497,
+      "num_input_tokens_seen": 639105088,
+      "step": 4530
+    },
+    {
+      "epoch": 0.2766964034038533,
+      "grad_norm": 0.5373267531394958,
+      "learning_rate": 8e-05,
+      "loss": 1.4828,
+      "num_input_tokens_seen": 640515580,
+      "step": 4540
+    },
+    {
+      "epoch": 0.27730586684747416,
+      "grad_norm": 0.5229721665382385,
+      "learning_rate": 8e-05,
+      "loss": 1.4927,
+      "num_input_tokens_seen": 641960740,
+      "step": 4550
+    },
+    {
+      "epoch": 0.277915330291095,
+      "grad_norm": 0.5390836000442505,
+      "learning_rate": 8e-05,
+      "loss": 1.5461,
+      "num_input_tokens_seen": 643367736,
+      "step": 4560
+    },
+    {
+      "epoch": 0.2785247937347158,
+      "grad_norm": 0.5476319789886475,
+      "learning_rate": 8e-05,
+      "loss": 1.5791,
+      "num_input_tokens_seen": 644815132,
+      "step": 4570
+    },
+    {
+      "epoch": 0.2791342571783366,
+      "grad_norm": 0.5691061615943909,
+      "learning_rate": 8e-05,
+      "loss": 1.5169,
+      "num_input_tokens_seen": 646189852,
+      "step": 4580
+    },
+    {
+      "epoch": 0.27974372062195746,
+      "grad_norm": 0.5501654744148254,
+      "learning_rate": 8e-05,
+      "loss": 1.4543,
+      "num_input_tokens_seen": 647603012,
+      "step": 4590
+    },
+    {
+      "epoch": 0.28035318406557824,
+      "grad_norm": 0.6424432992935181,
+      "learning_rate": 8e-05,
+      "loss": 1.5923,
+      "num_input_tokens_seen": 649042496,
+      "step": 4600
+    },
+    {
+      "epoch": 0.2809626475091991,
+      "grad_norm": 0.5200062394142151,
+      "learning_rate": 8e-05,
+      "loss": 1.5062,
+      "num_input_tokens_seen": 650458192,
+      "step": 4610
+    },
+    {
+      "epoch": 0.2815721109528199,
+      "grad_norm": 0.6190493702888489,
+      "learning_rate": 8e-05,
+      "loss": 1.5041,
+      "num_input_tokens_seen": 651891436,
+      "step": 4620
+    },
+    {
+      "epoch": 0.28218157439644076,
+      "grad_norm": 0.5610544681549072,
+      "learning_rate": 8e-05,
+      "loss": 1.4763,
+      "num_input_tokens_seen": 653287132,
+      "step": 4630
+    },
+    {
+      "epoch": 0.28279103784006154,
+      "grad_norm": 0.6020642518997192,
+      "learning_rate": 8e-05,
+      "loss": 1.4961,
+      "num_input_tokens_seen": 654715988,
+      "step": 4640
+    },
+    {
+      "epoch": 0.2834005012836824,
+      "grad_norm": 0.5459628701210022,
+      "learning_rate": 8e-05,
+      "loss": 1.5022,
+      "num_input_tokens_seen": 656116916,
+      "step": 4650
+    },
+    {
+      "epoch": 0.2840099647273032,
+      "grad_norm": 0.53467857837677,
+      "learning_rate": 8e-05,
+      "loss": 1.5723,
+      "num_input_tokens_seen": 657533764,
+      "step": 4660
+    },
+    {
+      "epoch": 0.284619428170924,
+      "grad_norm": 0.5662572383880615,
+      "learning_rate": 8e-05,
+      "loss": 1.4852,
+      "num_input_tokens_seen": 658933892,
+      "step": 4670
+    },
+    {
+      "epoch": 0.28522889161454484,
+      "grad_norm": 0.6344622373580933,
+      "learning_rate": 8e-05,
+      "loss": 1.5427,
+      "num_input_tokens_seen": 660343924,
+      "step": 4680
+    },
+    {
+      "epoch": 0.2858383550581657,
+      "grad_norm": 0.6415901184082031,
+      "learning_rate": 8e-05,
+      "loss": 1.4782,
+      "num_input_tokens_seen": 661722560,
+      "step": 4690
+    },
+    {
+      "epoch": 0.28644781850178647,
+      "grad_norm": 0.6006565690040588,
+      "learning_rate": 8e-05,
+      "loss": 1.5184,
+      "num_input_tokens_seen": 663126892,
+      "step": 4700
+    },
+    {
+      "epoch": 0.2870572819454073,
+      "grad_norm": 0.5979477763175964,
+      "learning_rate": 8e-05,
+      "loss": 1.4963,
+      "num_input_tokens_seen": 664549752,
+      "step": 4710
+    },
+    {
+      "epoch": 0.28766674538902814,
+      "grad_norm": 0.560646653175354,
+      "learning_rate": 8e-05,
+      "loss": 1.4522,
+      "num_input_tokens_seen": 665976108,
+      "step": 4720
+    },
+    {
+      "epoch": 0.288276208832649,
+      "grad_norm": 0.4898284673690796,
+      "learning_rate": 8e-05,
+      "loss": 1.4917,
+      "num_input_tokens_seen": 667423832,
+      "step": 4730
+    },
+    {
+      "epoch": 0.28888567227626977,
+      "grad_norm": 0.5207868218421936,
+      "learning_rate": 8e-05,
+      "loss": 1.4298,
+      "num_input_tokens_seen": 668840704,
+      "step": 4740
+    },
+    {
+      "epoch": 0.2894951357198906,
+      "grad_norm": 0.5953548550605774,
+      "learning_rate": 8e-05,
+      "loss": 1.5316,
+      "num_input_tokens_seen": 670225524,
+      "step": 4750
+    },
+    {
+      "epoch": 0.29010459916351145,
+      "grad_norm": 0.6386962532997131,
+      "learning_rate": 8e-05,
+      "loss": 1.5264,
+      "num_input_tokens_seen": 671644060,
+      "step": 4760
+    },
+    {
+      "epoch": 0.29071406260713223,
+      "grad_norm": 0.581168532371521,
+      "learning_rate": 8e-05,
+      "loss": 1.4726,
+      "num_input_tokens_seen": 673055808,
+      "step": 4770
+    },
+    {
+      "epoch": 0.29132352605075307,
+      "grad_norm": 0.7128572463989258,
+      "learning_rate": 8e-05,
+      "loss": 1.4414,
+      "num_input_tokens_seen": 674498056,
+      "step": 4780
+    },
+    {
+      "epoch": 0.2919329894943739,
+      "grad_norm": 0.5856772065162659,
+      "learning_rate": 8e-05,
+      "loss": 1.507,
+      "num_input_tokens_seen": 675874572,
+      "step": 4790
+    },
+    {
+      "epoch": 0.2925424529379947,
+      "grad_norm": 0.6448741555213928,
+      "learning_rate": 8e-05,
+      "loss": 1.488,
+      "num_input_tokens_seen": 677325828,
+      "step": 4800
+    },
+    {
+      "epoch": 0.29315191638161553,
+      "grad_norm": 0.6047000885009766,
+      "learning_rate": 8e-05,
+      "loss": 1.4665,
+      "num_input_tokens_seen": 678720324,
+      "step": 4810
+    },
+    {
+      "epoch": 0.29376137982523637,
+      "grad_norm": 0.5227587819099426,
+      "learning_rate": 8e-05,
+      "loss": 1.5083,
+      "num_input_tokens_seen": 680144452,
+      "step": 4820
+    },
+    {
+      "epoch": 0.2943708432688572,
+      "grad_norm": 0.5058688521385193,
+      "learning_rate": 8e-05,
+      "loss": 1.5108,
+      "num_input_tokens_seen": 681537236,
+      "step": 4830
+    },
+    {
+      "epoch": 0.294980306712478,
+      "grad_norm": 0.6823422312736511,
+      "learning_rate": 8e-05,
+      "loss": 1.5768,
+      "num_input_tokens_seen": 682917996,
+      "step": 4840
+    },
+    {
+      "epoch": 0.29558977015609883,
+      "grad_norm": 0.6330751180648804,
+      "learning_rate": 8e-05,
+      "loss": 1.5083,
+      "num_input_tokens_seen": 684353484,
+      "step": 4850
+    },
+    {
+      "epoch": 0.29619923359971967,
+      "grad_norm": 0.6086271405220032,
+      "learning_rate": 8e-05,
+      "loss": 1.5022,
+      "num_input_tokens_seen": 685747296,
+      "step": 4860
+    },
+    {
+      "epoch": 0.29680869704334045,
+      "grad_norm": 0.5951144695281982,
+      "learning_rate": 8e-05,
+      "loss": 1.583,
+      "num_input_tokens_seen": 687170156,
+      "step": 4870
+    },
+    {
+      "epoch": 0.2974181604869613,
+      "grad_norm": 0.5558424592018127,
+      "learning_rate": 8e-05,
+      "loss": 1.4283,
+      "num_input_tokens_seen": 688574796,
+      "step": 4880
+    },
+    {
+      "epoch": 0.29802762393058213,
+      "grad_norm": 0.6256549954414368,
+      "learning_rate": 8e-05,
+      "loss": 1.6313,
+      "num_input_tokens_seen": 689998376,
+      "step": 4890
+    },
+    {
+      "epoch": 0.2986370873742029,
+      "grad_norm": 0.5750453472137451,
+      "learning_rate": 8e-05,
+      "loss": 1.4544,
+      "num_input_tokens_seen": 691383556,
+      "step": 4900
+    },
+    {
+      "epoch": 0.29924655081782375,
+      "grad_norm": 0.69186931848526,
+      "learning_rate": 8e-05,
+      "loss": 1.5238,
+      "num_input_tokens_seen": 692797940,
+      "step": 4910
+    },
+    {
+      "epoch": 0.2998560142614446,
+      "grad_norm": 0.5753540992736816,
+      "learning_rate": 8e-05,
+      "loss": 1.5077,
+      "num_input_tokens_seen": 694209928,
+      "step": 4920
+    },
+    {
+      "epoch": 0.3004654777050654,
+      "grad_norm": 0.6008949279785156,
+      "learning_rate": 8e-05,
+      "loss": 1.5407,
+      "num_input_tokens_seen": 695637656,
+      "step": 4930
+    },
+    {
+      "epoch": 0.3010749411486862,
+      "grad_norm": 0.5696120262145996,
+      "learning_rate": 8e-05,
+      "loss": 1.4826,
+      "num_input_tokens_seen": 697072516,
+      "step": 4940
+    },
+    {
+      "epoch": 0.30168440459230705,
+      "grad_norm": 0.5745052099227905,
+      "learning_rate": 8e-05,
+      "loss": 1.5266,
+      "num_input_tokens_seen": 698505536,
+      "step": 4950
+    },
+    {
+      "epoch": 0.3022938680359279,
+      "grad_norm": 0.5735147595405579,
+      "learning_rate": 8e-05,
+      "loss": 1.4663,
+      "num_input_tokens_seen": 699890632,
+      "step": 4960
+    },
+    {
+      "epoch": 0.3029033314795487,
+      "grad_norm": 0.5933482050895691,
+      "learning_rate": 8e-05,
+      "loss": 1.521,
+      "num_input_tokens_seen": 701311308,
+      "step": 4970
+    },
+    {
+      "epoch": 0.3035127949231695,
+      "grad_norm": 0.5017425417900085,
+      "learning_rate": 8e-05,
+      "loss": 1.4664,
+      "num_input_tokens_seen": 702715200,
+      "step": 4980
+    },
+    {
+      "epoch": 0.30412225836679035,
+      "grad_norm": 0.6305922865867615,
+      "learning_rate": 8e-05,
+      "loss": 1.5193,
+      "num_input_tokens_seen": 704117168,
+      "step": 4990
+    },
+    {
+      "epoch": 0.30473172181041114,
+      "grad_norm": 0.5459288954734802,
+      "learning_rate": 8e-05,
+      "loss": 1.4912,
+      "num_input_tokens_seen": 705551000,
+      "step": 5000
+    },
+    {
+      "epoch": 0.305341185254032,
+      "grad_norm": 0.5966220498085022,
+      "learning_rate": 8e-05,
+      "loss": 1.5046,
+      "num_input_tokens_seen": 706973412,
+      "step": 5010
+    },
+    {
+      "epoch": 0.3059506486976528,
+      "grad_norm": 0.5664955377578735,
+      "learning_rate": 8e-05,
+      "loss": 1.5611,
+      "num_input_tokens_seen": 708316344,
+      "step": 5020
+    },
+    {
+      "epoch": 0.3065601121412736,
+      "grad_norm": 0.5635570883750916,
+      "learning_rate": 8e-05,
+      "loss": 1.4977,
+      "num_input_tokens_seen": 709764736,
+      "step": 5030
+    },
+    {
+      "epoch": 0.30716957558489444,
+      "grad_norm": 0.555467963218689,
+      "learning_rate": 8e-05,
+      "loss": 1.5142,
+      "num_input_tokens_seen": 711149252,
+      "step": 5040
+    },
+    {
+      "epoch": 0.3077790390285153,
+      "grad_norm": 0.5750370025634766,
+      "learning_rate": 8e-05,
+      "loss": 1.4936,
+      "num_input_tokens_seen": 712543440,
+      "step": 5050
+    },
+    {
+      "epoch": 0.3083885024721361,
+      "grad_norm": 0.4970649778842926,
+      "learning_rate": 8e-05,
+      "loss": 1.4638,
+      "num_input_tokens_seen": 713913020,
+      "step": 5060
+    },
+    {
+      "epoch": 0.3089979659157569,
+      "grad_norm": 0.510879635810852,
+      "learning_rate": 8e-05,
+      "loss": 1.455,
+      "num_input_tokens_seen": 715350768,
+      "step": 5070
+    },
+    {
+      "epoch": 0.30960742935937774,
+      "grad_norm": 0.5770237445831299,
+      "learning_rate": 8e-05,
+      "loss": 1.5248,
+      "num_input_tokens_seen": 716774184,
+      "step": 5080
+    },
+    {
+      "epoch": 0.3102168928029986,
+      "grad_norm": 0.508195698261261,
+      "learning_rate": 8e-05,
+      "loss": 1.4908,
+      "num_input_tokens_seen": 718209544,
+      "step": 5090
+    },
+    {
+      "epoch": 0.31082635624661936,
+      "grad_norm": 0.501331627368927,
+      "learning_rate": 8e-05,
+      "loss": 1.4317,
+      "num_input_tokens_seen": 719606848,
+      "step": 5100
+    },
+    {
+      "epoch": 0.3114358196902402,
+      "grad_norm": 0.5919586420059204,
+      "learning_rate": 8e-05,
+      "loss": 1.4905,
+      "num_input_tokens_seen": 721020264,
+      "step": 5110
+    },
+    {
+      "epoch": 0.31204528313386104,
+      "grad_norm": 0.51059490442276,
+      "learning_rate": 8e-05,
+      "loss": 1.4769,
+      "num_input_tokens_seen": 722433924,
+      "step": 5120
+    },
+    {
+      "epoch": 0.3126547465774818,
+      "grad_norm": 0.5880390405654907,
+      "learning_rate": 8e-05,
+      "loss": 1.4865,
+      "num_input_tokens_seen": 723841812,
+      "step": 5130
+    },
+    {
+      "epoch": 0.31326421002110266,
+      "grad_norm": 0.5752800107002258,
+      "learning_rate": 8e-05,
+      "loss": 1.5344,
+      "num_input_tokens_seen": 725237092,
+      "step": 5140
+    },
+    {
+      "epoch": 0.3138736734647235,
+      "grad_norm": 0.5985324382781982,
+      "learning_rate": 8e-05,
+      "loss": 1.4905,
+      "num_input_tokens_seen": 726645824,
+      "step": 5150
+    },
+    {
+      "epoch": 0.31448313690834434,
+      "grad_norm": 0.5916955471038818,
+      "learning_rate": 8e-05,
+      "loss": 1.4696,
+      "num_input_tokens_seen": 728082536,
+      "step": 5160
+    },
+    {
+      "epoch": 0.3150926003519651,
+      "grad_norm": 0.5135784149169922,
+      "learning_rate": 8e-05,
+      "loss": 1.4112,
+      "num_input_tokens_seen": 729471752,
+      "step": 5170
+    },
+    {
+      "epoch": 0.31570206379558596,
+      "grad_norm": 0.597804069519043,
+      "learning_rate": 8e-05,
+      "loss": 1.4754,
+      "num_input_tokens_seen": 730906264,
+      "step": 5180
+    },
+    {
+      "epoch": 0.3163115272392068,
+      "grad_norm": 0.48136067390441895,
+      "learning_rate": 8e-05,
+      "loss": 1.5421,
+      "num_input_tokens_seen": 732311764,
+      "step": 5190
+    },
+    {
+      "epoch": 0.3169209906828276,
+      "grad_norm": 0.5294743180274963,
+      "learning_rate": 8e-05,
+      "loss": 1.42,
+      "num_input_tokens_seen": 733705960,
+      "step": 5200
+    },
+    {
+      "epoch": 0.3175304541264484,
+      "grad_norm": 0.49918803572654724,
+      "learning_rate": 8e-05,
+      "loss": 1.5201,
+      "num_input_tokens_seen": 735136208,
+      "step": 5210
+    },
+    {
+      "epoch": 0.31813991757006926,
+      "grad_norm": 0.6225274205207825,
+      "learning_rate": 8e-05,
+      "loss": 1.4671,
+      "num_input_tokens_seen": 736548572,
+      "step": 5220
+    },
+    {
+      "epoch": 0.31874938101369005,
+      "grad_norm": 0.6126905083656311,
+      "learning_rate": 8e-05,
+      "loss": 1.5022,
+      "num_input_tokens_seen": 737948492,
+      "step": 5230
+    },
+    {
+      "epoch": 0.3193588444573109,
+      "grad_norm": 0.6735352873802185,
+      "learning_rate": 8e-05,
+      "loss": 1.4824,
+      "num_input_tokens_seen": 739360472,
+      "step": 5240
+    },
+    {
+      "epoch": 0.3199683079009317,
+      "grad_norm": 0.5282402634620667,
+      "learning_rate": 8e-05,
+      "loss": 1.3916,
+      "num_input_tokens_seen": 740751936,
+      "step": 5250
+    },
+    {
+      "epoch": 0.32057777134455256,
+      "grad_norm": 0.5800075531005859,
+      "learning_rate": 8e-05,
+      "loss": 1.4639,
+      "num_input_tokens_seen": 742134900,
+      "step": 5260
+    },
+    {
+      "epoch": 0.32118723478817335,
+      "grad_norm": 0.7085527777671814,
+      "learning_rate": 8e-05,
+      "loss": 1.5392,
+      "num_input_tokens_seen": 743507672,
+      "step": 5270
+    },
+    {
+      "epoch": 0.3217966982317942,
+      "grad_norm": 0.5233438611030579,
+      "learning_rate": 8e-05,
+      "loss": 1.4817,
+      "num_input_tokens_seen": 744911012,
+      "step": 5280
+    },
+    {
+      "epoch": 0.322406161675415,
+      "grad_norm": 0.5503348112106323,
+      "learning_rate": 8e-05,
+      "loss": 1.4695,
+      "num_input_tokens_seen": 746348456,
+      "step": 5290
+    },
+    {
+      "epoch": 0.3230156251190358,
+      "grad_norm": 0.5066059827804565,
+      "learning_rate": 8e-05,
+      "loss": 1.5024,
+      "num_input_tokens_seen": 747760896,
+      "step": 5300
+    },
+    {
+      "epoch": 0.32362508856265665,
+      "grad_norm": 0.6413135528564453,
+      "learning_rate": 8e-05,
+      "loss": 1.4737,
+      "num_input_tokens_seen": 749182700,
+      "step": 5310
+    },
+    {
+      "epoch": 0.3242345520062775,
+      "grad_norm": 0.6394685506820679,
+      "learning_rate": 8e-05,
+      "loss": 1.497,
+      "num_input_tokens_seen": 750569816,
+      "step": 5320
+    },
+    {
+      "epoch": 0.32484401544989827,
+      "grad_norm": 0.6730980277061462,
+      "learning_rate": 8e-05,
+      "loss": 1.5137,
+      "num_input_tokens_seen": 752049528,
+      "step": 5330
+    },
+    {
+      "epoch": 0.3254534788935191,
+      "grad_norm": 0.5059208273887634,
+      "learning_rate": 8e-05,
+      "loss": 1.4047,
+      "num_input_tokens_seen": 753469828,
+      "step": 5340
+    },
+    {
+      "epoch": 0.32606294233713995,
+      "grad_norm": 0.6344226002693176,
+      "learning_rate": 8e-05,
+      "loss": 1.4388,
+      "num_input_tokens_seen": 754893840,
+      "step": 5350
+    },
+    {
+      "epoch": 0.3266724057807608,
+      "grad_norm": 0.5653851628303528,
+      "learning_rate": 8e-05,
+      "loss": 1.5341,
+      "num_input_tokens_seen": 756345148,
+      "step": 5360
+    },
+    {
+      "epoch": 0.32728186922438157,
+      "grad_norm": 0.5219058394432068,
+      "learning_rate": 8e-05,
+      "loss": 1.4945,
+      "num_input_tokens_seen": 757782128,
+      "step": 5370
+    },
+    {
+      "epoch": 0.3278913326680024,
+      "grad_norm": 0.6265976428985596,
+      "learning_rate": 8e-05,
+      "loss": 1.5015,
+      "num_input_tokens_seen": 759205500,
+      "step": 5380
+    },
+    {
+      "epoch": 0.32850079611162325,
+      "grad_norm": 0.5684333443641663,
+      "learning_rate": 8e-05,
+      "loss": 1.4409,
+      "num_input_tokens_seen": 760595412,
+      "step": 5390
+    },
+    {
+      "epoch": 0.32911025955524403,
+      "grad_norm": 0.5176586508750916,
+      "learning_rate": 8e-05,
+      "loss": 1.4624,
+      "num_input_tokens_seen": 761988752,
+      "step": 5400
+    },
+    {
+      "epoch": 0.3297197229988649,
+      "grad_norm": 0.6107752323150635,
+      "learning_rate": 8e-05,
+      "loss": 1.4213,
+      "num_input_tokens_seen": 763379640,
+      "step": 5410
+    },
+    {
+      "epoch": 0.3303291864424857,
+      "grad_norm": 0.7825304865837097,
+      "learning_rate": 8e-05,
+      "loss": 1.5461,
+      "num_input_tokens_seen": 764777136,
+      "step": 5420
+    },
+    {
+      "epoch": 0.3309386498861065,
+      "grad_norm": 0.5260487794876099,
+      "learning_rate": 8e-05,
+      "loss": 1.4863,
+      "num_input_tokens_seen": 766185628,
+      "step": 5430
+    },
+    {
+      "epoch": 0.33154811332972733,
+      "grad_norm": 0.5938482284545898,
+      "learning_rate": 8e-05,
+      "loss": 1.5117,
+      "num_input_tokens_seen": 767555088,
+      "step": 5440
+    },
+    {
+      "epoch": 0.3321575767733482,
+      "grad_norm": 0.5161702632904053,
+      "learning_rate": 8e-05,
+      "loss": 1.5179,
+      "num_input_tokens_seen": 768994976,
+      "step": 5450
+    },
+    {
+      "epoch": 0.332767040216969,
+      "grad_norm": 0.5811363458633423,
+      "learning_rate": 8e-05,
+      "loss": 1.496,
+      "num_input_tokens_seen": 770367576,
+      "step": 5460
+    },
+    {
+      "epoch": 0.3333765036605898,
+      "grad_norm": 0.5560015439987183,
+      "learning_rate": 8e-05,
+      "loss": 1.5018,
+      "num_input_tokens_seen": 771776868,
+      "step": 5470
+    },
+    {
+      "epoch": 0.33398596710421063,
+      "grad_norm": 0.5517913103103638,
+      "learning_rate": 8e-05,
+      "loss": 1.5197,
+      "num_input_tokens_seen": 773199948,
+      "step": 5480
+    },
+    {
+      "epoch": 0.3345954305478315,
+      "grad_norm": 0.5763364434242249,
+      "learning_rate": 8e-05,
+      "loss": 1.4531,
+      "num_input_tokens_seen": 774635788,
+      "step": 5490
+    },
+    {
+      "epoch": 0.33520489399145226,
+      "grad_norm": 0.5679718852043152,
+      "learning_rate": 8e-05,
+      "loss": 1.5311,
+      "num_input_tokens_seen": 776038608,
+      "step": 5500
+    },
+    {
+      "epoch": 0.3358143574350731,
+      "grad_norm": 0.6076738238334656,
+      "learning_rate": 8e-05,
+      "loss": 1.5176,
+      "num_input_tokens_seen": 777447108,
+      "step": 5510
+    },
+    {
+      "epoch": 0.33642382087869394,
+      "grad_norm": 0.5834559798240662,
+      "learning_rate": 8e-05,
+      "loss": 1.5009,
+      "num_input_tokens_seen": 778861076,
+      "step": 5520
+    },
+    {
+      "epoch": 0.3370332843223147,
+      "grad_norm": 0.5051541328430176,
+      "learning_rate": 8e-05,
+      "loss": 1.4895,
+      "num_input_tokens_seen": 780263956,
+      "step": 5530
+    },
+    {
+      "epoch": 0.33764274776593556,
+      "grad_norm": 0.5240879058837891,
+      "learning_rate": 8e-05,
+      "loss": 1.5555,
+      "num_input_tokens_seen": 781682280,
+      "step": 5540
+    },
+    {
+      "epoch": 0.3382522112095564,
+      "grad_norm": 0.4911370873451233,
+      "learning_rate": 8e-05,
+      "loss": 1.4297,
+      "num_input_tokens_seen": 783125984,
+      "step": 5550
+    },
+    {
+      "epoch": 0.33886167465317724,
+      "grad_norm": 0.5864706635475159,
+      "learning_rate": 8e-05,
+      "loss": 1.5062,
+      "num_input_tokens_seen": 784568312,
+      "step": 5560
+    },
+    {
+      "epoch": 0.339471138096798,
+      "grad_norm": 0.5688824653625488,
+      "learning_rate": 8e-05,
+      "loss": 1.4459,
+      "num_input_tokens_seen": 785985256,
+      "step": 5570
+    },
+    {
+      "epoch": 0.34008060154041886,
+      "grad_norm": 0.521834671497345,
+      "learning_rate": 8e-05,
+      "loss": 1.4781,
+      "num_input_tokens_seen": 787388484,
+      "step": 5580
+    },
+    {
+      "epoch": 0.3406900649840397,
+      "grad_norm": 0.5668113827705383,
+      "learning_rate": 8e-05,
+      "loss": 1.4792,
+      "num_input_tokens_seen": 788815088,
+      "step": 5590
+    },
+    {
+      "epoch": 0.3412995284276605,
+      "grad_norm": 0.5735951662063599,
+      "learning_rate": 8e-05,
+      "loss": 1.482,
+      "num_input_tokens_seen": 790242832,
+      "step": 5600
+    },
+    {
+      "epoch": 0.3419089918712813,
+      "grad_norm": 0.5737040042877197,
+      "learning_rate": 8e-05,
+      "loss": 1.4741,
+      "num_input_tokens_seen": 791648756,
+      "step": 5610
+    },
+    {
+      "epoch": 0.34251845531490216,
+      "grad_norm": 0.5083377361297607,
+      "learning_rate": 8e-05,
+      "loss": 1.4006,
+      "num_input_tokens_seen": 793038600,
+      "step": 5620
+    },
+    {
+      "epoch": 0.34312791875852294,
+      "grad_norm": 0.48619720339775085,
+      "learning_rate": 8e-05,
+      "loss": 1.4241,
+      "num_input_tokens_seen": 794466268,
+      "step": 5630
+    },
+    {
+      "epoch": 0.3437373822021438,
+      "grad_norm": 0.5126612782478333,
+      "learning_rate": 8e-05,
+      "loss": 1.472,
+      "num_input_tokens_seen": 795840520,
+      "step": 5640
+    },
+    {
+      "epoch": 0.3443468456457646,
+      "grad_norm": 0.553537905216217,
+      "learning_rate": 8e-05,
+      "loss": 1.4802,
+      "num_input_tokens_seen": 797249844,
+      "step": 5650
+    },
+    {
+      "epoch": 0.34495630908938546,
+      "grad_norm": 0.576506495475769,
+      "learning_rate": 8e-05,
+      "loss": 1.4802,
+      "num_input_tokens_seen": 798662736,
+      "step": 5660
+    },
+    {
+      "epoch": 0.34556577253300624,
+      "grad_norm": 0.5432490706443787,
+      "learning_rate": 8e-05,
+      "loss": 1.5314,
+      "num_input_tokens_seen": 800085376,
+      "step": 5670
+    },
+    {
+      "epoch": 0.3461752359766271,
+      "grad_norm": 0.5892384648323059,
+      "learning_rate": 8e-05,
+      "loss": 1.4365,
+      "num_input_tokens_seen": 801519988,
+      "step": 5680
+    },
+    {
+      "epoch": 0.3467846994202479,
+      "grad_norm": 0.550477147102356,
+      "learning_rate": 8e-05,
+      "loss": 1.4772,
+      "num_input_tokens_seen": 802930932,
+      "step": 5690
+    },
+    {
+      "epoch": 0.3473941628638687,
+      "grad_norm": 0.46749433875083923,
+      "learning_rate": 8e-05,
+      "loss": 1.4563,
+      "num_input_tokens_seen": 804355416,
+      "step": 5700
+    },
+    {
+      "epoch": 0.34800362630748954,
+      "grad_norm": 0.6174665689468384,
+      "learning_rate": 8e-05,
+      "loss": 1.4796,
+      "num_input_tokens_seen": 805809472,
+      "step": 5710
+    },
+    {
+      "epoch": 0.3486130897511104,
+      "grad_norm": 0.5981230139732361,
+      "learning_rate": 8e-05,
+      "loss": 1.4873,
+      "num_input_tokens_seen": 807216568,
+      "step": 5720
+    },
+    {
+      "epoch": 0.34922255319473117,
+      "grad_norm": 0.8428714275360107,
+      "learning_rate": 8e-05,
+      "loss": 1.5775,
+      "num_input_tokens_seen": 808599216,
+      "step": 5730
+    },
+    {
+      "epoch": 0.349832016638352,
+      "grad_norm": 0.5814129710197449,
+      "learning_rate": 8e-05,
+      "loss": 1.4566,
+      "num_input_tokens_seen": 810003948,
+      "step": 5740
+    },
+    {
+      "epoch": 0.35044148008197284,
+      "grad_norm": 0.6042664647102356,
+      "learning_rate": 8e-05,
+      "loss": 1.5147,
+      "num_input_tokens_seen": 811385708,
+      "step": 5750
+    },
+    {
+      "epoch": 0.35105094352559363,
+      "grad_norm": 0.5238978862762451,
+      "learning_rate": 8e-05,
+      "loss": 1.477,
+      "num_input_tokens_seen": 812784196,
+      "step": 5760
+    },
+    {
+      "epoch": 0.35166040696921447,
+      "grad_norm": 0.5527967810630798,
+      "learning_rate": 8e-05,
+      "loss": 1.5258,
+      "num_input_tokens_seen": 814197040,
+      "step": 5770
+    },
+    {
+      "epoch": 0.3522698704128353,
+      "grad_norm": 0.539685070514679,
+      "learning_rate": 8e-05,
+      "loss": 1.505,
+      "num_input_tokens_seen": 815610940,
+      "step": 5780
+    },
+    {
+      "epoch": 0.35287933385645615,
+      "grad_norm": 0.5669323801994324,
+      "learning_rate": 8e-05,
+      "loss": 1.4015,
+      "num_input_tokens_seen": 817022692,
+      "step": 5790
+    },
+    {
+      "epoch": 0.35348879730007693,
+      "grad_norm": 0.6599160432815552,
+      "learning_rate": 8e-05,
+      "loss": 1.456,
+      "num_input_tokens_seen": 818400184,
+      "step": 5800
+    },
+    {
+      "epoch": 0.35409826074369777,
+      "grad_norm": 0.6508094668388367,
+      "learning_rate": 8e-05,
+      "loss": 1.432,
+      "num_input_tokens_seen": 819819924,
+      "step": 5810
+    },
+    {
+      "epoch": 0.3547077241873186,
+      "grad_norm": 0.45001381635665894,
+      "learning_rate": 8e-05,
+      "loss": 1.4672,
+      "num_input_tokens_seen": 821219128,
+      "step": 5820
+    },
+    {
+      "epoch": 0.3553171876309394,
+      "grad_norm": 0.5801172256469727,
+      "learning_rate": 8e-05,
+      "loss": 1.4242,
+      "num_input_tokens_seen": 822620580,
+      "step": 5830
+    },
+    {
+      "epoch": 0.35592665107456023,
+      "grad_norm": 0.588651716709137,
+      "learning_rate": 8e-05,
+      "loss": 1.5112,
+      "num_input_tokens_seen": 824046412,
+      "step": 5840
+    },
+    {
+      "epoch": 0.35653611451818107,
+      "grad_norm": 0.5115150213241577,
+      "learning_rate": 8e-05,
+      "loss": 1.4171,
+      "num_input_tokens_seen": 825453892,
+      "step": 5850
+    },
+    {
+      "epoch": 0.35714557796180185,
+      "grad_norm": 0.6320896744728088,
+      "learning_rate": 8e-05,
+      "loss": 1.5575,
+      "num_input_tokens_seen": 826866976,
+      "step": 5860
+    },
+    {
+      "epoch": 0.3577550414054227,
+      "grad_norm": 0.5521752834320068,
+      "learning_rate": 8e-05,
+      "loss": 1.4981,
+      "num_input_tokens_seen": 828276444,
+      "step": 5870
+    },
+    {
+      "epoch": 0.35836450484904353,
+      "grad_norm": 0.4956464469432831,
+      "learning_rate": 8e-05,
+      "loss": 1.4582,
+      "num_input_tokens_seen": 829717872,
+      "step": 5880
+    },
+    {
+      "epoch": 0.35897396829266437,
+      "grad_norm": 0.6160319447517395,
+      "learning_rate": 8e-05,
+      "loss": 1.4735,
+      "num_input_tokens_seen": 831124236,
+      "step": 5890
+    },
+    {
+      "epoch": 0.35958343173628515,
+      "grad_norm": 0.5546999573707581,
+      "learning_rate": 8e-05,
+      "loss": 1.5093,
+      "num_input_tokens_seen": 832544524,
+      "step": 5900
+    },
+    {
+      "epoch": 0.360192895179906,
+      "grad_norm": 0.5691624879837036,
+      "learning_rate": 8e-05,
+      "loss": 1.4528,
+      "num_input_tokens_seen": 833976044,
+      "step": 5910
+    },
+    {
+      "epoch": 0.36080235862352683,
+      "grad_norm": 0.5685278177261353,
+      "learning_rate": 8e-05,
+      "loss": 1.4905,
+      "num_input_tokens_seen": 835420540,
+      "step": 5920
+    },
+    {
+      "epoch": 0.3614118220671476,
+      "grad_norm": 0.5637968182563782,
+      "learning_rate": 8e-05,
+      "loss": 1.438,
+      "num_input_tokens_seen": 836844744,
+      "step": 5930
+    },
+    {
+      "epoch": 0.36202128551076845,
+      "grad_norm": 0.5134168267250061,
+      "learning_rate": 8e-05,
+      "loss": 1.4268,
+      "num_input_tokens_seen": 838204660,
+      "step": 5940
+    },
+    {
+      "epoch": 0.3626307489543893,
+      "grad_norm": 0.6179930567741394,
+      "learning_rate": 8e-05,
+      "loss": 1.4223,
+      "num_input_tokens_seen": 839639804,
+      "step": 5950
+    },
+    {
+      "epoch": 0.3632402123980101,
+      "grad_norm": 0.6156236529350281,
+      "learning_rate": 8e-05,
+      "loss": 1.4914,
+      "num_input_tokens_seen": 841052816,
+      "step": 5960
+    },
+    {
+      "epoch": 0.3638496758416309,
+      "grad_norm": 0.5875065922737122,
+      "learning_rate": 8e-05,
+      "loss": 1.4722,
+      "num_input_tokens_seen": 842451068,
+      "step": 5970
+    },
+    {
+      "epoch": 0.36445913928525175,
+      "grad_norm": 0.6806820631027222,
+      "learning_rate": 8e-05,
+      "loss": 1.4535,
+      "num_input_tokens_seen": 843881212,
+      "step": 5980
+    },
+    {
+      "epoch": 0.3650686027288726,
+      "grad_norm": 0.5545870065689087,
+      "learning_rate": 8e-05,
+      "loss": 1.4698,
+      "num_input_tokens_seen": 845263884,
+      "step": 5990
+    },
+    {
+      "epoch": 0.3656780661724934,
+      "grad_norm": 0.4307273328304291,
+      "learning_rate": 8e-05,
+      "loss": 1.5082,
+      "num_input_tokens_seen": 846667064,
+      "step": 6000
+    },
+    {
+      "epoch": 0.3662875296161142,
+      "grad_norm": 0.58298659324646,
+      "learning_rate": 8e-05,
+      "loss": 1.4679,
+      "num_input_tokens_seen": 848068848,
+      "step": 6010
+    },
+    {
+      "epoch": 0.36689699305973505,
+      "grad_norm": 0.45190927386283875,
+      "learning_rate": 8e-05,
+      "loss": 1.4729,
+      "num_input_tokens_seen": 849463044,
+      "step": 6020
+    },
+    {
+      "epoch": 0.36750645650335584,
+      "grad_norm": 0.5742015838623047,
+      "learning_rate": 8e-05,
+      "loss": 1.4787,
+      "num_input_tokens_seen": 850877448,
+      "step": 6030
+    },
+    {
+      "epoch": 0.3681159199469767,
+      "grad_norm": 0.5748038291931152,
+      "learning_rate": 8e-05,
+      "loss": 1.4364,
+      "num_input_tokens_seen": 852315284,
+      "step": 6040
+    },
+    {
+      "epoch": 0.3687253833905975,
+      "grad_norm": 0.5301962494850159,
+      "learning_rate": 8e-05,
+      "loss": 1.4812,
+      "num_input_tokens_seen": 853729908,
+      "step": 6050
+    },
+    {
+      "epoch": 0.3693348468342183,
+      "grad_norm": 0.589756965637207,
+      "learning_rate": 8e-05,
+      "loss": 1.4993,
+      "num_input_tokens_seen": 855128224,
+      "step": 6060
+    },
+    {
+      "epoch": 0.36994431027783914,
+      "grad_norm": 0.5072458982467651,
+      "learning_rate": 8e-05,
+      "loss": 1.5019,
+      "num_input_tokens_seen": 856560260,
+      "step": 6070
+    },
+    {
+      "epoch": 0.37055377372146,
+      "grad_norm": 0.532714307308197,
+      "learning_rate": 8e-05,
+      "loss": 1.4981,
+      "num_input_tokens_seen": 857947816,
+      "step": 6080
+    },
+    {
+      "epoch": 0.3711632371650808,
+      "grad_norm": 0.5182844400405884,
+      "learning_rate": 8e-05,
+      "loss": 1.4869,
+      "num_input_tokens_seen": 859354540,
+      "step": 6090
+    },
+    {
+      "epoch": 0.3717727006087016,
+      "grad_norm": 0.5242214798927307,
+      "learning_rate": 8e-05,
+      "loss": 1.498,
+      "num_input_tokens_seen": 860784760,
+      "step": 6100
+    },
+    {
+      "epoch": 0.37238216405232244,
+      "grad_norm": 0.5378220677375793,
+      "learning_rate": 8e-05,
+      "loss": 1.4231,
+      "num_input_tokens_seen": 862165624,
+      "step": 6110
+    },
+    {
+      "epoch": 0.3729916274959433,
+      "grad_norm": 0.5064565539360046,
+      "learning_rate": 8e-05,
+      "loss": 1.4333,
+      "num_input_tokens_seen": 863576044,
+      "step": 6120
+    },
+    {
+      "epoch": 0.37360109093956406,
+      "grad_norm": 0.6327691078186035,
+      "learning_rate": 8e-05,
+      "loss": 1.4521,
+      "num_input_tokens_seen": 864950492,
+      "step": 6130
+    },
+    {
+      "epoch": 0.3742105543831849,
+      "grad_norm": 0.6391866207122803,
+      "learning_rate": 8e-05,
+      "loss": 1.4526,
+      "num_input_tokens_seen": 866373744,
+      "step": 6140
+    },
+    {
+      "epoch": 0.37482001782680574,
+      "grad_norm": 0.5664558410644531,
+      "learning_rate": 8e-05,
+      "loss": 1.5089,
+      "num_input_tokens_seen": 867827248,
+      "step": 6150
+    },
+    {
+      "epoch": 0.3754294812704265,
+      "grad_norm": 0.6216340065002441,
+      "learning_rate": 8e-05,
+      "loss": 1.4829,
+      "num_input_tokens_seen": 869255428,
+      "step": 6160
+    },
+    {
+      "epoch": 0.37603894471404736,
+      "grad_norm": 0.612844705581665,
+      "learning_rate": 8e-05,
+      "loss": 1.4621,
+      "num_input_tokens_seen": 870649988,
+      "step": 6170
+    },
+    {
+      "epoch": 0.3766484081576682,
+      "grad_norm": 0.5435082912445068,
+      "learning_rate": 8e-05,
+      "loss": 1.3834,
+      "num_input_tokens_seen": 872079800,
+      "step": 6180
+    },
+    {
+      "epoch": 0.37725787160128904,
+      "grad_norm": 0.6071653366088867,
+      "learning_rate": 8e-05,
+      "loss": 1.4301,
+      "num_input_tokens_seen": 873466904,
+      "step": 6190
+    },
+    {
+      "epoch": 0.3778673350449098,
+      "grad_norm": 0.5115202069282532,
+      "learning_rate": 8e-05,
+      "loss": 1.4156,
+      "num_input_tokens_seen": 874891528,
+      "step": 6200
+    },
+    {
+      "epoch": 0.37847679848853066,
+      "grad_norm": 0.5494471788406372,
+      "learning_rate": 8e-05,
+      "loss": 1.4658,
+      "num_input_tokens_seen": 876284912,
+      "step": 6210
+    },
+    {
+      "epoch": 0.3790862619321515,
+      "grad_norm": 0.4871610701084137,
+      "learning_rate": 8e-05,
+      "loss": 1.4497,
+      "num_input_tokens_seen": 877703108,
+      "step": 6220
+    },
+    {
+      "epoch": 0.3796957253757723,
+      "grad_norm": 0.6249947547912598,
+      "learning_rate": 8e-05,
+      "loss": 1.4456,
+      "num_input_tokens_seen": 879123340,
+      "step": 6230
+    },
+    {
+      "epoch": 0.3803051888193931,
+      "grad_norm": 0.5746230483055115,
+      "learning_rate": 8e-05,
+      "loss": 1.5171,
+      "num_input_tokens_seen": 880553048,
+      "step": 6240
+    },
+    {
+      "epoch": 0.38091465226301396,
+      "grad_norm": 0.5843877792358398,
+      "learning_rate": 8e-05,
+      "loss": 1.5046,
+      "num_input_tokens_seen": 881984776,
+      "step": 6250
+    },
+    {
+      "epoch": 0.38152411570663475,
+      "grad_norm": 0.5240760445594788,
+      "learning_rate": 8e-05,
+      "loss": 1.4689,
+      "num_input_tokens_seen": 883371668,
+      "step": 6260
+    },
+    {
+      "epoch": 0.3821335791502556,
+      "grad_norm": 0.5815708041191101,
+      "learning_rate": 8e-05,
+      "loss": 1.5634,
+      "num_input_tokens_seen": 884792444,
+      "step": 6270
+    },
+    {
+      "epoch": 0.3827430425938764,
+      "grad_norm": 0.6558341383934021,
+      "learning_rate": 8e-05,
+      "loss": 1.4519,
+      "num_input_tokens_seen": 886181612,
+      "step": 6280
+    },
+    {
+      "epoch": 0.38335250603749726,
+      "grad_norm": 0.5229777097702026,
+      "learning_rate": 8e-05,
+      "loss": 1.4538,
+      "num_input_tokens_seen": 887601452,
+      "step": 6290
+    },
+    {
+      "epoch": 0.38396196948111805,
+      "grad_norm": 0.5209792852401733,
+      "learning_rate": 8e-05,
+      "loss": 1.4415,
+      "num_input_tokens_seen": 889034508,
+      "step": 6300
+    },
+    {
+      "epoch": 0.3845714329247389,
+      "grad_norm": 0.5684558749198914,
+      "learning_rate": 8e-05,
+      "loss": 1.4671,
+      "num_input_tokens_seen": 890425468,
+      "step": 6310
+    },
+    {
+      "epoch": 0.3851808963683597,
+      "grad_norm": 0.578091025352478,
+      "learning_rate": 8e-05,
+      "loss": 1.4735,
+      "num_input_tokens_seen": 891826040,
+      "step": 6320
+    },
+    {
+      "epoch": 0.3857903598119805,
+      "grad_norm": 0.5623794198036194,
+      "learning_rate": 8e-05,
+      "loss": 1.4926,
+      "num_input_tokens_seen": 893262592,
+      "step": 6330
+    },
+    {
+      "epoch": 0.38639982325560135,
+      "grad_norm": 0.5411602854728699,
+      "learning_rate": 8e-05,
+      "loss": 1.4318,
+      "num_input_tokens_seen": 894682080,
+      "step": 6340
+    },
+    {
+      "epoch": 0.3870092866992222,
+      "grad_norm": 0.5540904998779297,
+      "learning_rate": 8e-05,
+      "loss": 1.4068,
+      "num_input_tokens_seen": 896099164,
+      "step": 6350
+    },
+    {
+      "epoch": 0.38761875014284297,
+      "grad_norm": 0.5864216685295105,
+      "learning_rate": 8e-05,
+      "loss": 1.399,
+      "num_input_tokens_seen": 897476228,
+      "step": 6360
+    },
+    {
+      "epoch": 0.3882282135864638,
+      "grad_norm": 0.6339991092681885,
+      "learning_rate": 8e-05,
+      "loss": 1.467,
+      "num_input_tokens_seen": 898854080,
+      "step": 6370
+    },
+    {
+      "epoch": 0.38883767703008465,
+      "grad_norm": 0.555851399898529,
+      "learning_rate": 8e-05,
+      "loss": 1.4563,
+      "num_input_tokens_seen": 900256152,
+      "step": 6380
+    },
+    {
+      "epoch": 0.3894471404737055,
+      "grad_norm": 0.5713673233985901,
+      "learning_rate": 8e-05,
+      "loss": 1.5061,
+      "num_input_tokens_seen": 901627728,
+      "step": 6390
+    },
+    {
+      "epoch": 0.39005660391732627,
+      "grad_norm": 0.5585047006607056,
+      "learning_rate": 8e-05,
+      "loss": 1.4458,
+      "num_input_tokens_seen": 902992920,
+      "step": 6400
+    },
+    {
+      "epoch": 0.3906660673609471,
+      "grad_norm": 0.4973145127296448,
+      "learning_rate": 8e-05,
+      "loss": 1.4566,
+      "num_input_tokens_seen": 904318848,
+      "step": 6410
+    },
+    {
+      "epoch": 0.39127553080456795,
+      "grad_norm": 0.5424864888191223,
+      "learning_rate": 8e-05,
+      "loss": 1.4537,
+      "num_input_tokens_seen": 905700004,
+      "step": 6420
+    },
+    {
+      "epoch": 0.39188499424818873,
+      "grad_norm": 0.5117954611778259,
+      "learning_rate": 8e-05,
+      "loss": 1.4243,
+      "num_input_tokens_seen": 907103420,
+      "step": 6430
+    },
+    {
+      "epoch": 0.3924944576918096,
+      "grad_norm": 0.5525716543197632,
+      "learning_rate": 8e-05,
+      "loss": 1.4567,
+      "num_input_tokens_seen": 908554396,
+      "step": 6440
+    },
+    {
+      "epoch": 0.3931039211354304,
+      "grad_norm": 0.5715943574905396,
+      "learning_rate": 8e-05,
+      "loss": 1.44,
+      "num_input_tokens_seen": 909956772,
+      "step": 6450
+    },
+    {
+      "epoch": 0.3937133845790512,
+      "grad_norm": 0.5808925628662109,
+      "learning_rate": 8e-05,
+      "loss": 1.4815,
+      "num_input_tokens_seen": 911395988,
+      "step": 6460
+    },
+    {
+      "epoch": 0.39432284802267203,
+      "grad_norm": 0.6523997187614441,
+      "learning_rate": 8e-05,
+      "loss": 1.4218,
+      "num_input_tokens_seen": 912784372,
+      "step": 6470
+    },
+    {
+      "epoch": 0.3949323114662929,
+      "grad_norm": 0.6092135310173035,
+      "learning_rate": 8e-05,
+      "loss": 1.4576,
+      "num_input_tokens_seen": 914207364,
+      "step": 6480
+    },
+    {
+      "epoch": 0.39554177490991366,
+      "grad_norm": 0.5254921317100525,
+      "learning_rate": 8e-05,
+      "loss": 1.4644,
+      "num_input_tokens_seen": 915636420,
+      "step": 6490
+    },
+    {
+      "epoch": 0.3961512383535345,
+      "grad_norm": 0.6733999252319336,
+      "learning_rate": 8e-05,
+      "loss": 1.4234,
+      "num_input_tokens_seen": 917065908,
+      "step": 6500
+    },
+    {
+      "epoch": 0.39676070179715534,
+      "grad_norm": 0.5586018562316895,
+      "learning_rate": 8e-05,
+      "loss": 1.5169,
+      "num_input_tokens_seen": 918451632,
+      "step": 6510
+    },
+    {
+      "epoch": 0.3973701652407762,
+      "grad_norm": 0.5588156580924988,
+      "learning_rate": 8e-05,
+      "loss": 1.3798,
+      "num_input_tokens_seen": 919829072,
+      "step": 6520
+    },
+    {
+      "epoch": 0.39797962868439696,
+      "grad_norm": 0.5383242964744568,
+      "learning_rate": 8e-05,
+      "loss": 1.468,
+      "num_input_tokens_seen": 921242500,
+      "step": 6530
+    },
+    {
+      "epoch": 0.3985890921280178,
+      "grad_norm": 0.516828715801239,
+      "learning_rate": 8e-05,
+      "loss": 1.4953,
+      "num_input_tokens_seen": 922668428,
+      "step": 6540
+    },
+    {
+      "epoch": 0.39919855557163864,
+      "grad_norm": 0.5715090036392212,
+      "learning_rate": 8e-05,
+      "loss": 1.4606,
+      "num_input_tokens_seen": 924115712,
+      "step": 6550
+    },
+    {
+      "epoch": 0.3998080190152594,
+      "grad_norm": 0.5680440664291382,
+      "learning_rate": 8e-05,
+      "loss": 1.498,
+      "num_input_tokens_seen": 925525280,
+      "step": 6560
+    },
+    {
+      "epoch": 0.40041748245888026,
+      "grad_norm": 0.5580743551254272,
+      "learning_rate": 8e-05,
+      "loss": 1.4742,
+      "num_input_tokens_seen": 926948888,
+      "step": 6570
+    },
+    {
+      "epoch": 0.4010269459025011,
+      "grad_norm": 0.5276727676391602,
+      "learning_rate": 8e-05,
+      "loss": 1.4808,
+      "num_input_tokens_seen": 928345036,
+      "step": 6580
+    },
+    {
+      "epoch": 0.4016364093461219,
+      "grad_norm": 0.5557569265365601,
+      "learning_rate": 8e-05,
+      "loss": 1.5104,
+      "num_input_tokens_seen": 929788420,
+      "step": 6590
+    },
+    {
+      "epoch": 0.4022458727897427,
+      "grad_norm": 0.5946210026741028,
+      "learning_rate": 8e-05,
+      "loss": 1.4409,
+      "num_input_tokens_seen": 931238968,
+      "step": 6600
+    },
+    {
+      "epoch": 0.40285533623336356,
+      "grad_norm": 0.5625696778297424,
+      "learning_rate": 8e-05,
+      "loss": 1.4899,
+      "num_input_tokens_seen": 932685800,
+      "step": 6610
+    },
+    {
+      "epoch": 0.4034647996769844,
+      "grad_norm": 0.5148335695266724,
+      "learning_rate": 8e-05,
+      "loss": 1.4781,
+      "num_input_tokens_seen": 934066584,
+      "step": 6620
+    },
+    {
+      "epoch": 0.4040742631206052,
+      "grad_norm": 0.5163837671279907,
+      "learning_rate": 8e-05,
+      "loss": 1.3748,
+      "num_input_tokens_seen": 935478628,
+      "step": 6630
+    },
+    {
+      "epoch": 0.404683726564226,
+      "grad_norm": 0.5167310237884521,
+      "learning_rate": 8e-05,
+      "loss": 1.4414,
+      "num_input_tokens_seen": 936901648,
+      "step": 6640
+    },
+    {
+      "epoch": 0.40529319000784686,
+      "grad_norm": 0.5219048261642456,
+      "learning_rate": 8e-05,
+      "loss": 1.4501,
+      "num_input_tokens_seen": 938297212,
+      "step": 6650
+    },
+    {
+      "epoch": 0.40590265345146764,
+      "grad_norm": 0.4949333369731903,
+      "learning_rate": 8e-05,
+      "loss": 1.4314,
+      "num_input_tokens_seen": 939716088,
+      "step": 6660
+    },
+    {
+      "epoch": 0.4065121168950885,
+      "grad_norm": 0.6273387670516968,
+      "learning_rate": 8e-05,
+      "loss": 1.3889,
+      "num_input_tokens_seen": 941144088,
+      "step": 6670
+    },
+    {
+      "epoch": 0.4071215803387093,
+      "grad_norm": 0.7101430296897888,
+      "learning_rate": 8e-05,
+      "loss": 1.488,
+      "num_input_tokens_seen": 942563156,
+      "step": 6680
+    },
+    {
+      "epoch": 0.4077310437823301,
+      "grad_norm": 0.5433675646781921,
+      "learning_rate": 8e-05,
+      "loss": 1.419,
+      "num_input_tokens_seen": 943957008,
+      "step": 6690
+    },
+    {
+      "epoch": 0.40834050722595094,
+      "grad_norm": 0.5392646789550781,
+      "learning_rate": 8e-05,
+      "loss": 1.4103,
+      "num_input_tokens_seen": 945372644,
+      "step": 6700
+    },
+    {
+      "epoch": 0.4089499706695718,
+      "grad_norm": 0.537996768951416,
+      "learning_rate": 8e-05,
+      "loss": 1.4489,
+      "num_input_tokens_seen": 946820652,
+      "step": 6710
+    },
+    {
+      "epoch": 0.4095594341131926,
+      "grad_norm": 0.625306248664856,
+      "learning_rate": 8e-05,
+      "loss": 1.4446,
+      "num_input_tokens_seen": 948202592,
+      "step": 6720
+    },
+    {
+      "epoch": 0.4101688975568134,
+      "grad_norm": 0.5231274962425232,
+      "learning_rate": 8e-05,
+      "loss": 1.4562,
+      "num_input_tokens_seen": 949622864,
+      "step": 6730
+    },
+    {
+      "epoch": 0.41077836100043424,
+      "grad_norm": 0.5137037634849548,
+      "learning_rate": 8e-05,
+      "loss": 1.518,
+      "num_input_tokens_seen": 951016388,
+      "step": 6740
+    },
+    {
+      "epoch": 0.4113878244440551,
+      "grad_norm": 0.5600166320800781,
+      "learning_rate": 8e-05,
+      "loss": 1.3742,
+      "num_input_tokens_seen": 952467456,
+      "step": 6750
+    },
+    {
+      "epoch": 0.41199728788767587,
+      "grad_norm": 0.5758654475212097,
+      "learning_rate": 8e-05,
+      "loss": 1.4311,
+      "num_input_tokens_seen": 953914724,
+      "step": 6760
+    },
+    {
+      "epoch": 0.4126067513312967,
+      "grad_norm": 0.5547349452972412,
+      "learning_rate": 8e-05,
+      "loss": 1.4261,
+      "num_input_tokens_seen": 955295720,
+      "step": 6770
+    },
+    {
+      "epoch": 0.41321621477491755,
+      "grad_norm": 0.5143230557441711,
+      "learning_rate": 8e-05,
+      "loss": 1.4667,
+      "num_input_tokens_seen": 956681368,
+      "step": 6780
+    },
+    {
+      "epoch": 0.41382567821853833,
+      "grad_norm": 0.6212142109870911,
+      "learning_rate": 8e-05,
+      "loss": 1.5469,
+      "num_input_tokens_seen": 958080316,
+      "step": 6790
+    },
+    {
+      "epoch": 0.41443514166215917,
+      "grad_norm": 0.6437855362892151,
+      "learning_rate": 8e-05,
+      "loss": 1.4665,
+      "num_input_tokens_seen": 959483132,
+      "step": 6800
+    },
+    {
+      "epoch": 0.41504460510578,
+      "grad_norm": 0.6089012622833252,
+      "learning_rate": 8e-05,
+      "loss": 1.4368,
+      "num_input_tokens_seen": 960872924,
+      "step": 6810
+    },
+    {
+      "epoch": 0.41565406854940085,
+      "grad_norm": 0.5278199315071106,
+      "learning_rate": 8e-05,
+      "loss": 1.4487,
+      "num_input_tokens_seen": 962282748,
+      "step": 6820
+    },
+    {
+      "epoch": 0.41626353199302163,
+      "grad_norm": 0.5420494079589844,
+      "learning_rate": 8e-05,
+      "loss": 1.4431,
+      "num_input_tokens_seen": 963724892,
+      "step": 6830
+    },
+    {
+      "epoch": 0.41687299543664247,
+      "grad_norm": 0.5393320322036743,
+      "learning_rate": 8e-05,
+      "loss": 1.462,
+      "num_input_tokens_seen": 965116844,
+      "step": 6840
+    },
+    {
+      "epoch": 0.4174824588802633,
+      "grad_norm": 0.4785369634628296,
+      "learning_rate": 8e-05,
+      "loss": 1.4703,
+      "num_input_tokens_seen": 966494980,
+      "step": 6850
+    },
+    {
+      "epoch": 0.4180919223238841,
+      "grad_norm": 0.5344218611717224,
+      "learning_rate": 8e-05,
+      "loss": 1.4295,
+      "num_input_tokens_seen": 967908316,
+      "step": 6860
+    },
+    {
+      "epoch": 0.41870138576750493,
+      "grad_norm": 0.5722907781600952,
+      "learning_rate": 8e-05,
+      "loss": 1.5415,
+      "num_input_tokens_seen": 969300604,
+      "step": 6870
+    },
+    {
+      "epoch": 0.41931084921112577,
+      "grad_norm": 0.6001546382904053,
+      "learning_rate": 8e-05,
+      "loss": 1.4469,
+      "num_input_tokens_seen": 970686572,
+      "step": 6880
+    },
+    {
+      "epoch": 0.41992031265474655,
+      "grad_norm": 0.48998522758483887,
+      "learning_rate": 8e-05,
+      "loss": 1.497,
+      "num_input_tokens_seen": 972128596,
+      "step": 6890
+    },
+    {
+      "epoch": 0.4205297760983674,
+      "grad_norm": 0.5235675573348999,
+      "learning_rate": 8e-05,
+      "loss": 1.4051,
+      "num_input_tokens_seen": 973505548,
+      "step": 6900
+    },
+    {
+      "epoch": 0.42113923954198823,
+      "grad_norm": 0.45352238416671753,
+      "learning_rate": 8e-05,
+      "loss": 1.5063,
+      "num_input_tokens_seen": 974916112,
+      "step": 6910
+    },
+    {
+      "epoch": 0.42174870298560907,
+      "grad_norm": 0.6097639203071594,
+      "learning_rate": 8e-05,
+      "loss": 1.4945,
+      "num_input_tokens_seen": 976344436,
+      "step": 6920
+    },
+    {
+      "epoch": 0.42235816642922985,
+      "grad_norm": 0.6069915294647217,
+      "learning_rate": 8e-05,
+      "loss": 1.4796,
+      "num_input_tokens_seen": 977726860,
+      "step": 6930
+    },
+    {
+      "epoch": 0.4229676298728507,
+      "grad_norm": 0.555101752281189,
+      "learning_rate": 8e-05,
+      "loss": 1.465,
+      "num_input_tokens_seen": 979148684,
+      "step": 6940
+    },
+    {
+      "epoch": 0.42357709331647153,
+      "grad_norm": 0.5358740091323853,
+      "learning_rate": 8e-05,
+      "loss": 1.4698,
+      "num_input_tokens_seen": 980570088,
+      "step": 6950
+    },
+    {
+      "epoch": 0.4241865567600923,
+      "grad_norm": 0.5622055530548096,
+      "learning_rate": 8e-05,
+      "loss": 1.4742,
+      "num_input_tokens_seen": 982007352,
+      "step": 6960
+    },
+    {
+      "epoch": 0.42479602020371315,
+      "grad_norm": 0.548179566860199,
+      "learning_rate": 8e-05,
+      "loss": 1.3727,
+      "num_input_tokens_seen": 983386076,
+      "step": 6970
+    },
+    {
+      "epoch": 0.425405483647334,
+      "grad_norm": 0.49980705976486206,
+      "learning_rate": 8e-05,
+      "loss": 1.4423,
+      "num_input_tokens_seen": 984801120,
+      "step": 6980
+    },
+    {
+      "epoch": 0.4260149470909548,
+      "grad_norm": 0.5310669541358948,
+      "learning_rate": 8e-05,
+      "loss": 1.4295,
+      "num_input_tokens_seen": 986186972,
+      "step": 6990
+    },
+    {
+      "epoch": 0.4266244105345756,
+      "grad_norm": 0.5813121199607849,
+      "learning_rate": 8e-05,
+      "loss": 1.4334,
+      "num_input_tokens_seen": 987570284,
+      "step": 7000
+    },
+    {
+      "epoch": 0.42723387397819645,
+      "grad_norm": 0.5761700868606567,
+      "learning_rate": 8e-05,
+      "loss": 1.444,
+      "num_input_tokens_seen": 988955796,
+      "step": 7010
+    },
+    {
+      "epoch": 0.4278433374218173,
+      "grad_norm": 0.5148670673370361,
+      "learning_rate": 8e-05,
+      "loss": 1.4178,
+      "num_input_tokens_seen": 990350148,
+      "step": 7020
+    },
+    {
+      "epoch": 0.4284528008654381,
+      "grad_norm": 0.5637139678001404,
+      "learning_rate": 8e-05,
+      "loss": 1.452,
+      "num_input_tokens_seen": 991771360,
+      "step": 7030
+    },
+    {
+      "epoch": 0.4290622643090589,
+      "grad_norm": 0.5411230325698853,
+      "learning_rate": 8e-05,
+      "loss": 1.4548,
+      "num_input_tokens_seen": 993232832,
+      "step": 7040
+    },
+    {
+      "epoch": 0.42967172775267976,
+      "grad_norm": 0.5264154076576233,
+      "learning_rate": 8e-05,
+      "loss": 1.4627,
+      "num_input_tokens_seen": 994638464,
+      "step": 7050
+    },
+    {
+      "epoch": 0.43028119119630054,
+      "grad_norm": 0.6401540637016296,
+      "learning_rate": 8e-05,
+      "loss": 1.4346,
+      "num_input_tokens_seen": 996023220,
+      "step": 7060
+    },
+    {
+      "epoch": 0.4308906546399214,
+      "grad_norm": 0.5791023373603821,
+      "learning_rate": 8e-05,
+      "loss": 1.419,
+      "num_input_tokens_seen": 997433208,
+      "step": 7070
+    },
+    {
+      "epoch": 0.4315001180835422,
+      "grad_norm": 0.5640988945960999,
+      "learning_rate": 8e-05,
+      "loss": 1.4145,
+      "num_input_tokens_seen": 998853492,
+      "step": 7080
+    },
+    {
+      "epoch": 0.432109581527163,
+      "grad_norm": 0.6895711421966553,
+      "learning_rate": 8e-05,
+      "loss": 1.5127,
+      "num_input_tokens_seen": 1000269424,
+      "step": 7090
+    },
+    {
+      "epoch": 0.43271904497078384,
+      "grad_norm": 0.5173920392990112,
+      "learning_rate": 8e-05,
+      "loss": 1.4035,
+      "num_input_tokens_seen": 1001659736,
+      "step": 7100
+    },
+    {
+      "epoch": 0.4333285084144047,
+      "grad_norm": 0.518268346786499,
+      "learning_rate": 8e-05,
+      "loss": 1.4023,
+      "num_input_tokens_seen": 1003061084,
+      "step": 7110
+    },
+    {
+      "epoch": 0.4339379718580255,
+      "grad_norm": 0.4936695992946625,
+      "learning_rate": 8e-05,
+      "loss": 1.4437,
+      "num_input_tokens_seen": 1004496060,
+      "step": 7120
+    },
+    {
+      "epoch": 0.4345474353016463,
+      "grad_norm": 0.5039248466491699,
+      "learning_rate": 8e-05,
+      "loss": 1.4591,
+      "num_input_tokens_seen": 1005928784,
+      "step": 7130
+    },
+    {
+      "epoch": 0.43515689874526714,
+      "grad_norm": 0.5840840339660645,
+      "learning_rate": 8e-05,
+      "loss": 1.4275,
+      "num_input_tokens_seen": 1007361236,
+      "step": 7140
+    },
+    {
+      "epoch": 0.435766362188888,
+      "grad_norm": 0.5575258135795593,
+      "learning_rate": 8e-05,
+      "loss": 1.4881,
+      "num_input_tokens_seen": 1008766028,
+      "step": 7150
+    },
+    {
+      "epoch": 0.43637582563250876,
+      "grad_norm": 0.49963733553886414,
+      "learning_rate": 8e-05,
+      "loss": 1.4783,
+      "num_input_tokens_seen": 1010202448,
+      "step": 7160
+    },
+    {
+      "epoch": 0.4369852890761296,
+      "grad_norm": 0.4767080843448639,
+      "learning_rate": 8e-05,
+      "loss": 1.4741,
+      "num_input_tokens_seen": 1011606648,
+      "step": 7170
+    },
+    {
+      "epoch": 0.43759475251975044,
+      "grad_norm": 0.5568097829818726,
+      "learning_rate": 8e-05,
+      "loss": 1.5022,
+      "num_input_tokens_seen": 1013003276,
+      "step": 7180
+    },
+    {
+      "epoch": 0.4382042159633712,
+      "grad_norm": 0.5164647102355957,
+      "learning_rate": 8e-05,
+      "loss": 1.3946,
+      "num_input_tokens_seen": 1014431492,
+      "step": 7190
+    },
+    {
+      "epoch": 0.43881367940699206,
+      "grad_norm": 0.48684781789779663,
+      "learning_rate": 8e-05,
+      "loss": 1.4464,
+      "num_input_tokens_seen": 1015812592,
+      "step": 7200
+    },
+    {
+      "epoch": 0.4394231428506129,
+      "grad_norm": 0.4711934030056,
+      "learning_rate": 8e-05,
+      "loss": 1.4647,
+      "num_input_tokens_seen": 1017217112,
+      "step": 7210
+    },
+    {
+      "epoch": 0.44003260629423374,
+      "grad_norm": 0.5720543265342712,
+      "learning_rate": 8e-05,
+      "loss": 1.4723,
+      "num_input_tokens_seen": 1018655572,
+      "step": 7220
+    },
+    {
+      "epoch": 0.4406420697378545,
+      "grad_norm": 0.5549769997596741,
+      "learning_rate": 8e-05,
+      "loss": 1.4384,
+      "num_input_tokens_seen": 1020040320,
+      "step": 7230
+    },
+    {
+      "epoch": 0.44125153318147536,
+      "grad_norm": 0.5608406066894531,
+      "learning_rate": 8e-05,
+      "loss": 1.4143,
+      "num_input_tokens_seen": 1021483904,
+      "step": 7240
+    },
+    {
+      "epoch": 0.4418609966250962,
+      "grad_norm": 0.5251893401145935,
+      "learning_rate": 8e-05,
+      "loss": 1.3365,
+      "num_input_tokens_seen": 1022886696,
+      "step": 7250
+    },
+    {
+      "epoch": 0.442470460068717,
+      "grad_norm": 0.50985187292099,
+      "learning_rate": 8e-05,
+      "loss": 1.3904,
+      "num_input_tokens_seen": 1024290644,
+      "step": 7260
+    },
+    {
+      "epoch": 0.4430799235123378,
+      "grad_norm": 0.673911988735199,
+      "learning_rate": 8e-05,
+      "loss": 1.4286,
+      "num_input_tokens_seen": 1025736504,
+      "step": 7270
+    },
+    {
+      "epoch": 0.44368938695595866,
+      "grad_norm": 0.6181952357292175,
+      "learning_rate": 8e-05,
+      "loss": 1.4864,
+      "num_input_tokens_seen": 1027167200,
+      "step": 7280
+    },
+    {
+      "epoch": 0.44429885039957945,
+      "grad_norm": 0.6667662262916565,
+      "learning_rate": 8e-05,
+      "loss": 1.4002,
+      "num_input_tokens_seen": 1028578528,
+      "step": 7290
+    },
+    {
+      "epoch": 0.4449083138432003,
+      "grad_norm": 0.5670000910758972,
+      "learning_rate": 8e-05,
+      "loss": 1.4228,
+      "num_input_tokens_seen": 1029984048,
+      "step": 7300
+    },
+    {
+      "epoch": 0.4455177772868211,
+      "grad_norm": 0.527633786201477,
+      "learning_rate": 8e-05,
+      "loss": 1.423,
+      "num_input_tokens_seen": 1031348748,
+      "step": 7310
+    },
+    {
+      "epoch": 0.4461272407304419,
+      "grad_norm": 0.6152751445770264,
+      "learning_rate": 8e-05,
+      "loss": 1.4067,
+      "num_input_tokens_seen": 1032747992,
+      "step": 7320
+    },
+    {
+      "epoch": 0.44673670417406275,
+      "grad_norm": 0.6434319615364075,
+      "learning_rate": 8e-05,
+      "loss": 1.5126,
+      "num_input_tokens_seen": 1034148668,
+      "step": 7330
+    },
+    {
+      "epoch": 0.4473461676176836,
+      "grad_norm": 0.5347893238067627,
+      "learning_rate": 8e-05,
+      "loss": 1.5054,
+      "num_input_tokens_seen": 1035562136,
+      "step": 7340
+    },
+    {
+      "epoch": 0.4479556310613044,
+      "grad_norm": 0.6078560948371887,
+      "learning_rate": 8e-05,
+      "loss": 1.4428,
+      "num_input_tokens_seen": 1036939864,
+      "step": 7350
+    },
+    {
+      "epoch": 0.4485650945049252,
+      "grad_norm": 0.5783689618110657,
+      "learning_rate": 8e-05,
+      "loss": 1.4361,
+      "num_input_tokens_seen": 1038323952,
+      "step": 7360
+    },
+    {
+      "epoch": 0.44917455794854605,
+      "grad_norm": 0.5024809837341309,
+      "learning_rate": 8e-05,
+      "loss": 1.4971,
+      "num_input_tokens_seen": 1039722864,
+      "step": 7370
+    },
+    {
+      "epoch": 0.4497840213921669,
+      "grad_norm": 0.524659276008606,
+      "learning_rate": 8e-05,
+      "loss": 1.4948,
+      "num_input_tokens_seen": 1041139980,
+      "step": 7380
+    },
+    {
+      "epoch": 0.45039348483578767,
+      "grad_norm": 0.46882256865501404,
+      "learning_rate": 8e-05,
+      "loss": 1.4066,
+      "num_input_tokens_seen": 1042543716,
+      "step": 7390
+    },
+    {
+      "epoch": 0.4510029482794085,
+      "grad_norm": 1.1085089445114136,
+      "learning_rate": 8e-05,
+      "loss": 1.4335,
+      "num_input_tokens_seen": 1043957848,
+      "step": 7400
+    },
+    {
+      "epoch": 0.45161241172302935,
+      "grad_norm": 0.4720146059989929,
+      "learning_rate": 8e-05,
+      "loss": 1.5418,
+      "num_input_tokens_seen": 1045388008,
+      "step": 7410
+    },
+    {
+      "epoch": 0.45222187516665013,
+      "grad_norm": 0.47802358865737915,
+      "learning_rate": 8e-05,
+      "loss": 1.46,
+      "num_input_tokens_seen": 1046809520,
+      "step": 7420
+    },
+    {
+      "epoch": 0.452831338610271,
+      "grad_norm": 0.5443661212921143,
+      "learning_rate": 8e-05,
+      "loss": 1.4775,
+      "num_input_tokens_seen": 1048246316,
+      "step": 7430
+    },
+    {
+      "epoch": 0.4534408020538918,
+      "grad_norm": 0.5573719143867493,
+      "learning_rate": 8e-05,
+      "loss": 1.5155,
+      "num_input_tokens_seen": 1049699188,
+      "step": 7440
+    },
+    {
+      "epoch": 0.45405026549751265,
+      "grad_norm": 0.4769093692302704,
+      "learning_rate": 8e-05,
+      "loss": 1.3774,
+      "num_input_tokens_seen": 1051064820,
+      "step": 7450
+    },
+    {
+      "epoch": 0.45465972894113343,
+      "grad_norm": 0.5798795819282532,
+      "learning_rate": 8e-05,
+      "loss": 1.4688,
+      "num_input_tokens_seen": 1052484612,
+      "step": 7460
+    },
+    {
+      "epoch": 0.4552691923847543,
+      "grad_norm": 0.6087391376495361,
+      "learning_rate": 8e-05,
+      "loss": 1.416,
+      "num_input_tokens_seen": 1053910944,
+      "step": 7470
+    },
+    {
+      "epoch": 0.4558786558283751,
+      "grad_norm": 0.567535400390625,
+      "learning_rate": 8e-05,
+      "loss": 1.4634,
+      "num_input_tokens_seen": 1055332428,
+      "step": 7480
+    },
+    {
+      "epoch": 0.4564881192719959,
+      "grad_norm": 0.6195465922355652,
+      "learning_rate": 8e-05,
+      "loss": 1.3092,
+      "num_input_tokens_seen": 1056726892,
+      "step": 7490
+    },
+    {
+      "epoch": 0.45709758271561673,
+      "grad_norm": 0.5984314680099487,
+      "learning_rate": 8e-05,
+      "loss": 1.5001,
+      "num_input_tokens_seen": 1058136916,
+      "step": 7500
+    },
+    {
+      "epoch": 0.4577070461592376,
+      "grad_norm": 0.48998868465423584,
+      "learning_rate": 8e-05,
+      "loss": 1.4218,
+      "num_input_tokens_seen": 1059580108,
+      "step": 7510
+    },
+    {
+      "epoch": 0.45831650960285836,
+      "grad_norm": 0.5266983509063721,
+      "learning_rate": 8e-05,
+      "loss": 1.4329,
+      "num_input_tokens_seen": 1060958932,
+      "step": 7520
+    },
+    {
+      "epoch": 0.4589259730464792,
+      "grad_norm": 0.5505419969558716,
+      "learning_rate": 8e-05,
+      "loss": 1.4541,
+      "num_input_tokens_seen": 1062394804,
+      "step": 7530
+    },
+    {
+      "epoch": 0.45953543649010004,
+      "grad_norm": 0.5599749684333801,
+      "learning_rate": 8e-05,
+      "loss": 1.3939,
+      "num_input_tokens_seen": 1063805320,
+      "step": 7540
+    },
+    {
+      "epoch": 0.4601448999337209,
+      "grad_norm": 0.561083972454071,
+      "learning_rate": 8e-05,
+      "loss": 1.3821,
+      "num_input_tokens_seen": 1065164348,
+      "step": 7550
+    },
+    {
+      "epoch": 0.46075436337734166,
+      "grad_norm": 0.5383641719818115,
+      "learning_rate": 8e-05,
+      "loss": 1.5057,
+      "num_input_tokens_seen": 1066572824,
+      "step": 7560
+    },
+    {
+      "epoch": 0.4613638268209625,
+      "grad_norm": 0.4498058259487152,
+      "learning_rate": 8e-05,
+      "loss": 1.4079,
+      "num_input_tokens_seen": 1067999532,
+      "step": 7570
+    },
+    {
+      "epoch": 0.46197329026458334,
+      "grad_norm": 0.5523561239242554,
+      "learning_rate": 8e-05,
+      "loss": 1.4182,
+      "num_input_tokens_seen": 1069413536,
+      "step": 7580
+    },
+    {
+      "epoch": 0.4625827537082041,
+      "grad_norm": 0.5212712287902832,
+      "learning_rate": 8e-05,
+      "loss": 1.4475,
+      "num_input_tokens_seen": 1070846708,
+      "step": 7590
+    },
+    {
+      "epoch": 0.46319221715182496,
+      "grad_norm": 0.5242462158203125,
+      "learning_rate": 8e-05,
+      "loss": 1.427,
+      "num_input_tokens_seen": 1072226000,
+      "step": 7600
+    },
+    {
+      "epoch": 0.4638016805954458,
+      "grad_norm": 0.5377146601676941,
+      "learning_rate": 8e-05,
+      "loss": 1.4384,
+      "num_input_tokens_seen": 1073640188,
+      "step": 7610
+    },
+    {
+      "epoch": 0.4644111440390666,
+      "grad_norm": 0.5246424674987793,
+      "learning_rate": 8e-05,
+      "loss": 1.4643,
+      "num_input_tokens_seen": 1075053120,
+      "step": 7620
+    },
+    {
+      "epoch": 0.4650206074826874,
+      "grad_norm": 0.49351179599761963,
+      "learning_rate": 8e-05,
+      "loss": 1.4539,
+      "num_input_tokens_seen": 1076464372,
+      "step": 7630
+    },
+    {
+      "epoch": 0.46563007092630826,
+      "grad_norm": 0.4650513827800751,
+      "learning_rate": 8e-05,
+      "loss": 1.4067,
+      "num_input_tokens_seen": 1077887580,
+      "step": 7640
+    },
+    {
+      "epoch": 0.4662395343699291,
+      "grad_norm": 0.5255919694900513,
+      "learning_rate": 8e-05,
+      "loss": 1.4461,
+      "num_input_tokens_seen": 1079314396,
+      "step": 7650
+    },
+    {
+      "epoch": 0.4668489978135499,
+      "grad_norm": 0.45597824454307556,
+      "learning_rate": 8e-05,
+      "loss": 1.4076,
+      "num_input_tokens_seen": 1080722664,
+      "step": 7660
+    },
+    {
+      "epoch": 0.4674584612571707,
+      "grad_norm": 0.5342453718185425,
+      "learning_rate": 8e-05,
+      "loss": 1.3912,
+      "num_input_tokens_seen": 1082104756,
+      "step": 7670
+    },
+    {
+      "epoch": 0.46806792470079156,
+      "grad_norm": 0.5388374328613281,
+      "learning_rate": 8e-05,
+      "loss": 1.4392,
+      "num_input_tokens_seen": 1083515704,
+      "step": 7680
+    },
+    {
+      "epoch": 0.46867738814441234,
+      "grad_norm": 0.5510755777359009,
+      "learning_rate": 8e-05,
+      "loss": 1.4074,
+      "num_input_tokens_seen": 1084952776,
+      "step": 7690
+    },
+    {
+      "epoch": 0.4692868515880332,
+      "grad_norm": 0.5435276627540588,
+      "learning_rate": 8e-05,
+      "loss": 1.4402,
+      "num_input_tokens_seen": 1086371416,
+      "step": 7700
+    },
+    {
+      "epoch": 0.469896315031654,
+      "grad_norm": 0.4792904853820801,
+      "learning_rate": 8e-05,
+      "loss": 1.3989,
+      "num_input_tokens_seen": 1087779788,
+      "step": 7710
+    },
+    {
+      "epoch": 0.4705057784752748,
+      "grad_norm": 0.4804374575614929,
+      "learning_rate": 8e-05,
+      "loss": 1.3157,
+      "num_input_tokens_seen": 1089174612,
+      "step": 7720
+    },
+    {
+      "epoch": 0.47111524191889564,
+      "grad_norm": 0.47840744256973267,
+      "learning_rate": 8e-05,
+      "loss": 1.3208,
+      "num_input_tokens_seen": 1090588704,
+      "step": 7730
+    },
+    {
+      "epoch": 0.4717247053625165,
+      "grad_norm": 0.5292723178863525,
+      "learning_rate": 8e-05,
+      "loss": 1.4032,
+      "num_input_tokens_seen": 1091995860,
+      "step": 7740
+    },
+    {
+      "epoch": 0.4723341688061373,
+      "grad_norm": 0.5833985209465027,
+      "learning_rate": 8e-05,
+      "loss": 1.4286,
+      "num_input_tokens_seen": 1093395952,
+      "step": 7750
+    },
+    {
+      "epoch": 0.4729436322497581,
+      "grad_norm": 0.48498985171318054,
+      "learning_rate": 8e-05,
+      "loss": 1.4765,
+      "num_input_tokens_seen": 1094805108,
+      "step": 7760
+    },
+    {
+      "epoch": 0.47355309569337894,
+      "grad_norm": 0.5416792631149292,
+      "learning_rate": 8e-05,
+      "loss": 1.4559,
+      "num_input_tokens_seen": 1096191520,
+      "step": 7770
+    },
+    {
+      "epoch": 0.4741625591369998,
+      "grad_norm": 0.5118204951286316,
+      "learning_rate": 8e-05,
+      "loss": 1.4098,
+      "num_input_tokens_seen": 1097626836,
+      "step": 7780
+    },
+    {
+      "epoch": 0.47477202258062057,
+      "grad_norm": 0.5373117327690125,
+      "learning_rate": 8e-05,
+      "loss": 1.4222,
+      "num_input_tokens_seen": 1099008448,
+      "step": 7790
+    },
+    {
+      "epoch": 0.4753814860242414,
+      "grad_norm": 0.634472668170929,
+      "learning_rate": 8e-05,
+      "loss": 1.4626,
+      "num_input_tokens_seen": 1100430712,
+      "step": 7800
+    },
+    {
+      "epoch": 0.47599094946786225,
+      "grad_norm": 0.5606759190559387,
+      "learning_rate": 8e-05,
+      "loss": 1.4181,
+      "num_input_tokens_seen": 1101864384,
+      "step": 7810
+    },
+    {
+      "epoch": 0.47660041291148303,
+      "grad_norm": 0.6342650651931763,
+      "learning_rate": 8e-05,
+      "loss": 1.3675,
+      "num_input_tokens_seen": 1103281588,
+      "step": 7820
+    },
+    {
+      "epoch": 0.47720987635510387,
+      "grad_norm": 0.5315597057342529,
+      "learning_rate": 8e-05,
+      "loss": 1.4346,
+      "num_input_tokens_seen": 1104678480,
+      "step": 7830
+    },
+    {
+      "epoch": 0.4778193397987247,
+      "grad_norm": 0.6562740802764893,
+      "learning_rate": 8e-05,
+      "loss": 1.4669,
+      "num_input_tokens_seen": 1106087552,
+      "step": 7840
+    },
+    {
+      "epoch": 0.47842880324234555,
+      "grad_norm": 0.5265791416168213,
+      "learning_rate": 8e-05,
+      "loss": 1.4921,
+      "num_input_tokens_seen": 1107533152,
+      "step": 7850
+    },
+    {
+      "epoch": 0.47903826668596633,
+      "grad_norm": 0.4799894094467163,
+      "learning_rate": 8e-05,
+      "loss": 1.4559,
+      "num_input_tokens_seen": 1108933276,
+      "step": 7860
+    },
+    {
+      "epoch": 0.47964773012958717,
+      "grad_norm": 0.5308011174201965,
+      "learning_rate": 8e-05,
+      "loss": 1.3988,
+      "num_input_tokens_seen": 1110344220,
+      "step": 7870
+    },
+    {
+      "epoch": 0.480257193573208,
+      "grad_norm": 0.5046650171279907,
+      "learning_rate": 8e-05,
+      "loss": 1.4366,
+      "num_input_tokens_seen": 1111735944,
+      "step": 7880
+    },
+    {
+      "epoch": 0.4808666570168288,
+      "grad_norm": 0.5083670616149902,
+      "learning_rate": 8e-05,
+      "loss": 1.4219,
+      "num_input_tokens_seen": 1113172872,
+      "step": 7890
+    },
+    {
+      "epoch": 0.48147612046044963,
+      "grad_norm": 0.5572594404220581,
+      "learning_rate": 8e-05,
+      "loss": 1.4198,
+      "num_input_tokens_seen": 1114579892,
+      "step": 7900
+    },
+    {
+      "epoch": 0.48208558390407047,
+      "grad_norm": 0.5634285807609558,
+      "learning_rate": 8e-05,
+      "loss": 1.468,
+      "num_input_tokens_seen": 1115978076,
+      "step": 7910
+    },
+    {
+      "epoch": 0.48269504734769125,
+      "grad_norm": 0.4851561486721039,
+      "learning_rate": 8e-05,
+      "loss": 1.3592,
+      "num_input_tokens_seen": 1117361664,
+      "step": 7920
+    },
+    {
+      "epoch": 0.4833045107913121,
+      "grad_norm": 0.4718756079673767,
+      "learning_rate": 8e-05,
+      "loss": 1.439,
+      "num_input_tokens_seen": 1118816252,
+      "step": 7930
+    },
+    {
+      "epoch": 0.48391397423493293,
+      "grad_norm": 0.5959479808807373,
+      "learning_rate": 8e-05,
+      "loss": 1.4517,
+      "num_input_tokens_seen": 1120264108,
+      "step": 7940
+    },
+    {
+      "epoch": 0.48452343767855377,
+      "grad_norm": 0.5836747288703918,
+      "learning_rate": 8e-05,
+      "loss": 1.4244,
+      "num_input_tokens_seen": 1121650124,
+      "step": 7950
+    },
+    {
+      "epoch": 0.48513290112217455,
+      "grad_norm": 0.5627513527870178,
+      "learning_rate": 8e-05,
+      "loss": 1.4043,
+      "num_input_tokens_seen": 1123080060,
+      "step": 7960
+    },
+    {
+      "epoch": 0.4857423645657954,
+      "grad_norm": 0.5587437152862549,
+      "learning_rate": 8e-05,
+      "loss": 1.4019,
+      "num_input_tokens_seen": 1124511240,
+      "step": 7970
+    },
+    {
+      "epoch": 0.48635182800941623,
+      "grad_norm": 0.6421571969985962,
+      "learning_rate": 8e-05,
+      "loss": 1.4026,
+      "num_input_tokens_seen": 1125934080,
+      "step": 7980
+    },
+    {
+      "epoch": 0.486961291453037,
+      "grad_norm": 0.4620668888092041,
+      "learning_rate": 8e-05,
+      "loss": 1.4142,
+      "num_input_tokens_seen": 1127349476,
+      "step": 7990
+    },
+    {
+      "epoch": 0.48757075489665785,
+      "grad_norm": 0.5190436840057373,
+      "learning_rate": 8e-05,
+      "loss": 1.4625,
+      "num_input_tokens_seen": 1128787572,
+      "step": 8000
+    },
+    {
+      "epoch": 0.4881802183402787,
+      "grad_norm": 0.5595162510871887,
+      "learning_rate": 8e-05,
+      "loss": 1.4101,
+      "num_input_tokens_seen": 1130217996,
+      "step": 8010
+    },
+    {
+      "epoch": 0.4887896817838995,
+      "grad_norm": 0.4808787703514099,
+      "learning_rate": 8e-05,
+      "loss": 1.4088,
+      "num_input_tokens_seen": 1131625032,
+      "step": 8020
+    },
+    {
+      "epoch": 0.4893991452275203,
+      "grad_norm": 0.5082031488418579,
+      "learning_rate": 8e-05,
+      "loss": 1.4811,
+      "num_input_tokens_seen": 1133074808,
+      "step": 8030
+    },
+    {
+      "epoch": 0.49000860867114115,
+      "grad_norm": 0.7117087841033936,
+      "learning_rate": 8e-05,
+      "loss": 1.3978,
+      "num_input_tokens_seen": 1134471984,
+      "step": 8040
+    },
+    {
+      "epoch": 0.490618072114762,
+      "grad_norm": 0.4917582869529724,
+      "learning_rate": 8e-05,
+      "loss": 1.4326,
+      "num_input_tokens_seen": 1135871844,
+      "step": 8050
+    },
+    {
+      "epoch": 0.4912275355583828,
+      "grad_norm": 0.5196986794471741,
+      "learning_rate": 8e-05,
+      "loss": 1.4705,
+      "num_input_tokens_seen": 1137251528,
+      "step": 8060
+    },
+    {
+      "epoch": 0.4918369990020036,
+      "grad_norm": 0.45626333355903625,
+      "learning_rate": 8e-05,
+      "loss": 1.4173,
+      "num_input_tokens_seen": 1138668772,
+      "step": 8070
+    },
+    {
+      "epoch": 0.49244646244562446,
+      "grad_norm": 0.7028423547744751,
+      "learning_rate": 8e-05,
+      "loss": 1.4415,
+      "num_input_tokens_seen": 1140072992,
+      "step": 8080
+    },
+    {
+      "epoch": 0.49305592588924524,
+      "grad_norm": 0.4503718912601471,
+      "learning_rate": 8e-05,
+      "loss": 1.4979,
+      "num_input_tokens_seen": 1141473780,
+      "step": 8090
+    },
+    {
+      "epoch": 0.4936653893328661,
+      "grad_norm": 0.49622249603271484,
+      "learning_rate": 8e-05,
+      "loss": 1.4959,
+      "num_input_tokens_seen": 1142896620,
+      "step": 8100
+    },
+    {
+      "epoch": 0.4942748527764869,
+      "grad_norm": 0.45446139574050903,
+      "learning_rate": 8e-05,
+      "loss": 1.4125,
+      "num_input_tokens_seen": 1144295328,
+      "step": 8110
+    },
+    {
+      "epoch": 0.4948843162201077,
+      "grad_norm": 0.4794887602329254,
+      "learning_rate": 8e-05,
+      "loss": 1.4337,
+      "num_input_tokens_seen": 1145761996,
+      "step": 8120
+    },
+    {
+      "epoch": 0.49549377966372854,
+      "grad_norm": 0.535640299320221,
+      "learning_rate": 8e-05,
+      "loss": 1.4061,
+      "num_input_tokens_seen": 1147167508,
+      "step": 8130
+    },
+    {
+      "epoch": 0.4961032431073494,
+      "grad_norm": 0.5999556183815002,
+      "learning_rate": 8e-05,
+      "loss": 1.3923,
+      "num_input_tokens_seen": 1148590576,
+      "step": 8140
+    },
+    {
+      "epoch": 0.49671270655097016,
+      "grad_norm": 0.4626942574977875,
+      "learning_rate": 8e-05,
+      "loss": 1.3859,
+      "num_input_tokens_seen": 1149968016,
+      "step": 8150
+    },
+    {
+      "epoch": 0.497322169994591,
+      "grad_norm": 0.5697856545448303,
+      "learning_rate": 8e-05,
+      "loss": 1.4115,
+      "num_input_tokens_seen": 1151379092,
+      "step": 8160
+    },
+    {
+      "epoch": 0.49793163343821184,
+      "grad_norm": 0.5212153196334839,
+      "learning_rate": 8e-05,
+      "loss": 1.4006,
+      "num_input_tokens_seen": 1152789164,
+      "step": 8170
+    },
+    {
+      "epoch": 0.4985410968818327,
+      "grad_norm": 0.5307455062866211,
+      "learning_rate": 8e-05,
+      "loss": 1.4147,
+      "num_input_tokens_seen": 1154215056,
+      "step": 8180
+    },
+    {
+      "epoch": 0.49915056032545346,
+      "grad_norm": 0.49318286776542664,
+      "learning_rate": 8e-05,
+      "loss": 1.3641,
+      "num_input_tokens_seen": 1155560728,
+      "step": 8190
+    },
+    {
+      "epoch": 0.4997600237690743,
+      "grad_norm": 0.5201637148857117,
+      "learning_rate": 8e-05,
+      "loss": 1.4909,
+      "num_input_tokens_seen": 1156956988,
+      "step": 8200
+    },
+    {
+      "epoch": 0.5003694872126951,
+      "grad_norm": 0.5208278894424438,
+      "learning_rate": 8e-05,
+      "loss": 1.4412,
+      "num_input_tokens_seen": 1158393764,
+      "step": 8210
+    },
+    {
+      "epoch": 0.5009789506563159,
+      "grad_norm": 0.5729420185089111,
+      "learning_rate": 8e-05,
+      "loss": 1.3848,
+      "num_input_tokens_seen": 1159824056,
+      "step": 8220
+    },
+    {
+      "epoch": 0.5015884140999368,
+      "grad_norm": 0.4863448739051819,
+      "learning_rate": 8e-05,
+      "loss": 1.3707,
+      "num_input_tokens_seen": 1161241976,
+      "step": 8230
+    },
+    {
+      "epoch": 0.5021978775435576,
+      "grad_norm": 0.544262707233429,
+      "learning_rate": 8e-05,
+      "loss": 1.452,
+      "num_input_tokens_seen": 1162622704,
+      "step": 8240
+    },
+    {
+      "epoch": 0.5028073409871784,
+      "grad_norm": 0.5542010068893433,
+      "learning_rate": 8e-05,
+      "loss": 1.4808,
+      "num_input_tokens_seen": 1164019924,
+      "step": 8250
+    },
+    {
+      "epoch": 0.5034168044307993,
+      "grad_norm": 0.466775506734848,
+      "learning_rate": 8e-05,
+      "loss": 1.4303,
+      "num_input_tokens_seen": 1165465436,
+      "step": 8260
+    },
+    {
+      "epoch": 0.5040262678744201,
+      "grad_norm": 0.6469402313232422,
+      "learning_rate": 8e-05,
+      "loss": 1.4665,
+      "num_input_tokens_seen": 1166886136,
+      "step": 8270
+    },
+    {
+      "epoch": 0.5046357313180408,
+      "grad_norm": 0.6200391054153442,
+      "learning_rate": 8e-05,
+      "loss": 1.4426,
+      "num_input_tokens_seen": 1168299392,
+      "step": 8280
+    },
+    {
+      "epoch": 0.5052451947616617,
+      "grad_norm": 0.5519371628761292,
+      "learning_rate": 8e-05,
+      "loss": 1.4484,
+      "num_input_tokens_seen": 1169721572,
+      "step": 8290
+    },
+    {
+      "epoch": 0.5058546582052825,
+      "grad_norm": 0.5275024771690369,
+      "learning_rate": 8e-05,
+      "loss": 1.3321,
+      "num_input_tokens_seen": 1171103648,
+      "step": 8300
+    },
+    {
+      "epoch": 0.5064641216489033,
+      "grad_norm": 0.522892951965332,
+      "learning_rate": 8e-05,
+      "loss": 1.3864,
+      "num_input_tokens_seen": 1172502288,
+      "step": 8310
+    },
+    {
+      "epoch": 0.5070735850925242,
+      "grad_norm": 0.47808873653411865,
+      "learning_rate": 8e-05,
+      "loss": 1.3831,
+      "num_input_tokens_seen": 1173938752,
+      "step": 8320
+    },
+    {
+      "epoch": 0.507683048536145,
+      "grad_norm": 0.6184147596359253,
+      "learning_rate": 8e-05,
+      "loss": 1.4538,
+      "num_input_tokens_seen": 1175364784,
+      "step": 8330
+    },
+    {
+      "epoch": 0.5082925119797658,
+      "grad_norm": 0.5707573890686035,
+      "learning_rate": 8e-05,
+      "loss": 1.3267,
+      "num_input_tokens_seen": 1176787692,
+      "step": 8340
+    },
+    {
+      "epoch": 0.5089019754233867,
+      "grad_norm": 0.5967716574668884,
+      "learning_rate": 8e-05,
+      "loss": 1.4053,
+      "num_input_tokens_seen": 1178179564,
+      "step": 8350
+    },
+    {
+      "epoch": 0.5095114388670074,
+      "grad_norm": 0.5708630681037903,
+      "learning_rate": 8e-05,
+      "loss": 1.478,
+      "num_input_tokens_seen": 1179580276,
+      "step": 8360
+    },
+    {
+      "epoch": 0.5101209023106282,
+      "grad_norm": 0.5356885194778442,
+      "learning_rate": 8e-05,
+      "loss": 1.4038,
+      "num_input_tokens_seen": 1180980160,
+      "step": 8370
+    },
+    {
+      "epoch": 0.5107303657542491,
+      "grad_norm": 0.6001008749008179,
+      "learning_rate": 8e-05,
+      "loss": 1.4924,
+      "num_input_tokens_seen": 1182394556,
+      "step": 8380
+    },
+    {
+      "epoch": 0.5113398291978699,
+      "grad_norm": 0.4999195635318756,
+      "learning_rate": 8e-05,
+      "loss": 1.3733,
+      "num_input_tokens_seen": 1183836592,
+      "step": 8390
+    },
+    {
+      "epoch": 0.5119492926414907,
+      "grad_norm": 0.5139046311378479,
+      "learning_rate": 8e-05,
+      "loss": 1.399,
+      "num_input_tokens_seen": 1185238392,
+      "step": 8400
+    },
+    {
+      "epoch": 0.5125587560851116,
+      "grad_norm": 0.614023745059967,
+      "learning_rate": 8e-05,
+      "loss": 1.426,
+      "num_input_tokens_seen": 1186634124,
+      "step": 8410
+    },
+    {
+      "epoch": 0.5131682195287324,
+      "grad_norm": 0.5585746765136719,
+      "learning_rate": 8e-05,
+      "loss": 1.4441,
+      "num_input_tokens_seen": 1188016924,
+      "step": 8420
+    },
+    {
+      "epoch": 0.5137776829723533,
+      "grad_norm": 0.6003695726394653,
+      "learning_rate": 8e-05,
+      "loss": 1.4042,
+      "num_input_tokens_seen": 1189415668,
+      "step": 8430
+    },
+    {
+      "epoch": 0.514387146415974,
+      "grad_norm": 0.6207943558692932,
+      "learning_rate": 8e-05,
+      "loss": 1.3925,
+      "num_input_tokens_seen": 1190833848,
+      "step": 8440
+    },
+    {
+      "epoch": 0.5149966098595948,
+      "grad_norm": 0.5384374260902405,
+      "learning_rate": 8e-05,
+      "loss": 1.3952,
+      "num_input_tokens_seen": 1192256296,
+      "step": 8450
+    },
+    {
+      "epoch": 0.5156060733032157,
+      "grad_norm": 0.5247148275375366,
+      "learning_rate": 8e-05,
+      "loss": 1.471,
+      "num_input_tokens_seen": 1193697300,
+      "step": 8460
+    },
+    {
+      "epoch": 0.5162155367468365,
+      "grad_norm": 0.5606392621994019,
+      "learning_rate": 8e-05,
+      "loss": 1.5178,
+      "num_input_tokens_seen": 1195149596,
+      "step": 8470
+    },
+    {
+      "epoch": 0.5168250001904573,
+      "grad_norm": 0.5033918619155884,
+      "learning_rate": 8e-05,
+      "loss": 1.3954,
+      "num_input_tokens_seen": 1196543328,
+      "step": 8480
+    },
+    {
+      "epoch": 0.5174344636340782,
+      "grad_norm": 0.5382705330848694,
+      "learning_rate": 8e-05,
+      "loss": 1.3954,
+      "num_input_tokens_seen": 1197953020,
+      "step": 8490
+    },
+    {
+      "epoch": 0.518043927077699,
+      "grad_norm": 0.508390486240387,
+      "learning_rate": 8e-05,
+      "loss": 1.448,
+      "num_input_tokens_seen": 1199356544,
+      "step": 8500
+    },
+    {
+      "epoch": 0.5186533905213198,
+      "grad_norm": 0.49335700273513794,
+      "learning_rate": 8e-05,
+      "loss": 1.4365,
+      "num_input_tokens_seen": 1200772144,
+      "step": 8510
+    },
+    {
+      "epoch": 0.5192628539649407,
+      "grad_norm": 0.511746883392334,
+      "learning_rate": 8e-05,
+      "loss": 1.4413,
+      "num_input_tokens_seen": 1202172256,
+      "step": 8520
+    },
+    {
+      "epoch": 0.5198723174085614,
+      "grad_norm": 0.7011821269989014,
+      "learning_rate": 8e-05,
+      "loss": 1.4147,
+      "num_input_tokens_seen": 1203597072,
+      "step": 8530
+    },
+    {
+      "epoch": 0.5204817808521822,
+      "grad_norm": 0.612265408039093,
+      "learning_rate": 8e-05,
+      "loss": 1.4681,
+      "num_input_tokens_seen": 1204985968,
+      "step": 8540
+    },
+    {
+      "epoch": 0.5210912442958031,
+      "grad_norm": 0.510905385017395,
+      "learning_rate": 8e-05,
+      "loss": 1.4615,
+      "num_input_tokens_seen": 1206391220,
+      "step": 8550
+    },
+    {
+      "epoch": 0.5217007077394239,
+      "grad_norm": 0.6800599694252014,
+      "learning_rate": 8e-05,
+      "loss": 1.4496,
+      "num_input_tokens_seen": 1207786772,
+      "step": 8560
+    },
+    {
+      "epoch": 0.5223101711830447,
+      "grad_norm": 0.6727950572967529,
+      "learning_rate": 8e-05,
+      "loss": 1.4528,
+      "num_input_tokens_seen": 1209206112,
+      "step": 8570
+    },
+    {
+      "epoch": 0.5229196346266656,
+      "grad_norm": 0.6258131265640259,
+      "learning_rate": 8e-05,
+      "loss": 1.3601,
+      "num_input_tokens_seen": 1210620340,
+      "step": 8580
+    },
+    {
+      "epoch": 0.5235290980702864,
+      "grad_norm": 0.5519835948944092,
+      "learning_rate": 8e-05,
+      "loss": 1.3857,
+      "num_input_tokens_seen": 1212035396,
+      "step": 8590
+    },
+    {
+      "epoch": 0.5241385615139071,
+      "grad_norm": 0.49297013878822327,
+      "learning_rate": 8e-05,
+      "loss": 1.4223,
+      "num_input_tokens_seen": 1213445184,
+      "step": 8600
+    },
+    {
+      "epoch": 0.524748024957528,
+      "grad_norm": 0.47780099511146545,
+      "learning_rate": 8e-05,
+      "loss": 1.378,
+      "num_input_tokens_seen": 1214854316,
+      "step": 8610
+    },
+    {
+      "epoch": 0.5253574884011488,
+      "grad_norm": 0.5096369981765747,
+      "learning_rate": 8e-05,
+      "loss": 1.4467,
+      "num_input_tokens_seen": 1216250744,
+      "step": 8620
+    },
+    {
+      "epoch": 0.5259669518447697,
+      "grad_norm": 0.5387448072433472,
+      "learning_rate": 8e-05,
+      "loss": 1.4201,
+      "num_input_tokens_seen": 1217686440,
+      "step": 8630
+    },
+    {
+      "epoch": 0.5265764152883905,
+      "grad_norm": 0.5782696604728699,
+      "learning_rate": 8e-05,
+      "loss": 1.4183,
+      "num_input_tokens_seen": 1219090264,
+      "step": 8640
+    },
+    {
+      "epoch": 0.5271858787320113,
+      "grad_norm": 0.4248006343841553,
+      "learning_rate": 8e-05,
+      "loss": 1.5077,
+      "num_input_tokens_seen": 1220493872,
+      "step": 8650
+    },
+    {
+      "epoch": 0.5277953421756322,
+      "grad_norm": 0.5155441164970398,
+      "learning_rate": 8e-05,
+      "loss": 1.461,
+      "num_input_tokens_seen": 1221904396,
+      "step": 8660
+    },
+    {
+      "epoch": 0.528404805619253,
+      "grad_norm": 0.7794909477233887,
+      "learning_rate": 8e-05,
+      "loss": 1.359,
+      "num_input_tokens_seen": 1223272932,
+      "step": 8670
+    },
+    {
+      "epoch": 0.5290142690628737,
+      "grad_norm": 0.5529311299324036,
+      "learning_rate": 8e-05,
+      "loss": 1.4378,
+      "num_input_tokens_seen": 1224665304,
+      "step": 8680
+    },
+    {
+      "epoch": 0.5296237325064946,
+      "grad_norm": 0.5060685276985168,
+      "learning_rate": 8e-05,
+      "loss": 1.4173,
+      "num_input_tokens_seen": 1226062180,
+      "step": 8690
+    },
+    {
+      "epoch": 0.5302331959501154,
+      "grad_norm": 0.4951160252094269,
+      "learning_rate": 8e-05,
+      "loss": 1.4475,
+      "num_input_tokens_seen": 1227516508,
+      "step": 8700
+    },
+    {
+      "epoch": 0.5308426593937362,
+      "grad_norm": 0.48395630717277527,
+      "learning_rate": 8e-05,
+      "loss": 1.374,
+      "num_input_tokens_seen": 1228928164,
+      "step": 8710
+    },
+    {
+      "epoch": 0.5314521228373571,
+      "grad_norm": 0.5231735110282898,
+      "learning_rate": 8e-05,
+      "loss": 1.3518,
+      "num_input_tokens_seen": 1230332108,
+      "step": 8720
+    },
+    {
+      "epoch": 0.5320615862809779,
+      "grad_norm": 0.6426222324371338,
+      "learning_rate": 8e-05,
+      "loss": 1.3299,
+      "num_input_tokens_seen": 1231769688,
+      "step": 8730
+    },
+    {
+      "epoch": 0.5326710497245987,
+      "grad_norm": 0.496803343296051,
+      "learning_rate": 8e-05,
+      "loss": 1.3993,
+      "num_input_tokens_seen": 1233210764,
+      "step": 8740
+    },
+    {
+      "epoch": 0.5332805131682196,
+      "grad_norm": 0.4480886459350586,
+      "learning_rate": 8e-05,
+      "loss": 1.4256,
+      "num_input_tokens_seen": 1234671524,
+      "step": 8750
+    },
+    {
+      "epoch": 0.5338899766118403,
+      "grad_norm": 0.48204657435417175,
+      "learning_rate": 8e-05,
+      "loss": 1.3814,
+      "num_input_tokens_seen": 1236085908,
+      "step": 8760
+    },
+    {
+      "epoch": 0.5344994400554611,
+      "grad_norm": 0.5191913843154907,
+      "learning_rate": 8e-05,
+      "loss": 1.3768,
+      "num_input_tokens_seen": 1237467536,
+      "step": 8770
+    },
+    {
+      "epoch": 0.535108903499082,
+      "grad_norm": 0.5462549924850464,
+      "learning_rate": 8e-05,
+      "loss": 1.415,
+      "num_input_tokens_seen": 1238886584,
+      "step": 8780
+    },
+    {
+      "epoch": 0.5357183669427028,
+      "grad_norm": 0.574504554271698,
+      "learning_rate": 8e-05,
+      "loss": 1.4029,
+      "num_input_tokens_seen": 1240281716,
+      "step": 8790
+    },
+    {
+      "epoch": 0.5363278303863236,
+      "grad_norm": 0.5356009602546692,
+      "learning_rate": 8e-05,
+      "loss": 1.4261,
+      "num_input_tokens_seen": 1241675192,
+      "step": 8800
+    },
+    {
+      "epoch": 0.5369372938299445,
+      "grad_norm": 0.5325603485107422,
+      "learning_rate": 8e-05,
+      "loss": 1.3992,
+      "num_input_tokens_seen": 1243056036,
+      "step": 8810
+    },
+    {
+      "epoch": 0.5375467572735653,
+      "grad_norm": 0.47047409415245056,
+      "learning_rate": 8e-05,
+      "loss": 1.3463,
+      "num_input_tokens_seen": 1244426652,
+      "step": 8820
+    },
+    {
+      "epoch": 0.5381562207171862,
+      "grad_norm": 0.5718605518341064,
+      "learning_rate": 8e-05,
+      "loss": 1.4106,
+      "num_input_tokens_seen": 1245827804,
+      "step": 8830
+    },
+    {
+      "epoch": 0.538765684160807,
+      "grad_norm": 0.5218635201454163,
+      "learning_rate": 8e-05,
+      "loss": 1.4177,
+      "num_input_tokens_seen": 1247204320,
+      "step": 8840
+    },
+    {
+      "epoch": 0.5393751476044277,
+      "grad_norm": 0.6211861968040466,
+      "learning_rate": 8e-05,
+      "loss": 1.5218,
+      "num_input_tokens_seen": 1248618312,
+      "step": 8850
+    },
+    {
+      "epoch": 0.5399846110480486,
+      "grad_norm": 0.49085572361946106,
+      "learning_rate": 8e-05,
+      "loss": 1.4029,
+      "num_input_tokens_seen": 1250024496,
+      "step": 8860
+    },
+    {
+      "epoch": 0.5405940744916694,
+      "grad_norm": 0.5032560229301453,
+      "learning_rate": 8e-05,
+      "loss": 1.4333,
+      "num_input_tokens_seen": 1251433236,
+      "step": 8870
+    },
+    {
+      "epoch": 0.5412035379352902,
+      "grad_norm": 0.7042189240455627,
+      "learning_rate": 8e-05,
+      "loss": 1.3757,
+      "num_input_tokens_seen": 1252853548,
+      "step": 8880
+    },
+    {
+      "epoch": 0.5418130013789111,
+      "grad_norm": 0.5665264129638672,
+      "learning_rate": 8e-05,
+      "loss": 1.413,
+      "num_input_tokens_seen": 1254255272,
+      "step": 8890
+    },
+    {
+      "epoch": 0.5424224648225319,
+      "grad_norm": 0.5010056495666504,
+      "learning_rate": 8e-05,
+      "loss": 1.328,
+      "num_input_tokens_seen": 1255651628,
+      "step": 8900
+    },
+    {
+      "epoch": 0.5430319282661527,
+      "grad_norm": 0.5278030037879944,
+      "learning_rate": 8e-05,
+      "loss": 1.4238,
+      "num_input_tokens_seen": 1257075628,
+      "step": 8910
+    },
+    {
+      "epoch": 0.5436413917097735,
+      "grad_norm": 0.5137773156166077,
+      "learning_rate": 8e-05,
+      "loss": 1.4509,
+      "num_input_tokens_seen": 1258492128,
+      "step": 8920
+    },
+    {
+      "epoch": 0.5442508551533943,
+      "grad_norm": 0.49045154452323914,
+      "learning_rate": 8e-05,
+      "loss": 1.473,
+      "num_input_tokens_seen": 1259921876,
+      "step": 8930
+    },
+    {
+      "epoch": 0.5448603185970151,
+      "grad_norm": 0.5632349848747253,
+      "learning_rate": 8e-05,
+      "loss": 1.4155,
+      "num_input_tokens_seen": 1261282328,
+      "step": 8940
+    },
+    {
+      "epoch": 0.545469782040636,
+      "grad_norm": 0.4805395007133484,
+      "learning_rate": 8e-05,
+      "loss": 1.445,
+      "num_input_tokens_seen": 1262718360,
+      "step": 8950
+    },
+    {
+      "epoch": 0.5460792454842568,
+      "grad_norm": 0.5406620502471924,
+      "learning_rate": 8e-05,
+      "loss": 1.3724,
+      "num_input_tokens_seen": 1264130680,
+      "step": 8960
+    },
+    {
+      "epoch": 0.5466887089278776,
+      "grad_norm": 0.5531741976737976,
+      "learning_rate": 8e-05,
+      "loss": 1.4404,
+      "num_input_tokens_seen": 1265513744,
+      "step": 8970
+    },
+    {
+      "epoch": 0.5472981723714985,
+      "grad_norm": 0.5422140955924988,
+      "learning_rate": 8e-05,
+      "loss": 1.3957,
+      "num_input_tokens_seen": 1266903584,
+      "step": 8980
+    },
+    {
+      "epoch": 0.5479076358151193,
+      "grad_norm": 0.5059279203414917,
+      "learning_rate": 8e-05,
+      "loss": 1.36,
+      "num_input_tokens_seen": 1268300852,
+      "step": 8990
+    },
+    {
+      "epoch": 0.54851709925874,
+      "grad_norm": 0.4933377504348755,
+      "learning_rate": 8e-05,
+      "loss": 1.3714,
+      "num_input_tokens_seen": 1269717552,
+      "step": 9000
+    },
+    {
+      "epoch": 0.5491265627023609,
+      "grad_norm": 0.6613243818283081,
+      "learning_rate": 8e-05,
+      "loss": 1.4211,
+      "num_input_tokens_seen": 1271115644,
+      "step": 9010
+    },
+    {
+      "epoch": 0.5497360261459817,
+      "grad_norm": 0.49182480573654175,
+      "learning_rate": 8e-05,
+      "loss": 1.352,
+      "num_input_tokens_seen": 1272504180,
+      "step": 9020
+    },
+    {
+      "epoch": 0.5503454895896025,
+      "grad_norm": 0.5405098795890808,
+      "learning_rate": 8e-05,
+      "loss": 1.4031,
+      "num_input_tokens_seen": 1273932024,
+      "step": 9030
+    },
+    {
+      "epoch": 0.5509549530332234,
+      "grad_norm": 0.5463997721672058,
+      "learning_rate": 8e-05,
+      "loss": 1.5083,
+      "num_input_tokens_seen": 1275332648,
+      "step": 9040
+    },
+    {
+      "epoch": 0.5515644164768442,
+      "grad_norm": 0.5602612495422363,
+      "learning_rate": 8e-05,
+      "loss": 1.3726,
+      "num_input_tokens_seen": 1276735532,
+      "step": 9050
+    },
+    {
+      "epoch": 0.5521738799204651,
+      "grad_norm": 0.5101543068885803,
+      "learning_rate": 8e-05,
+      "loss": 1.3778,
+      "num_input_tokens_seen": 1278146092,
+      "step": 9060
+    },
+    {
+      "epoch": 0.5527833433640859,
+      "grad_norm": 0.5481278896331787,
+      "learning_rate": 8e-05,
+      "loss": 1.4419,
+      "num_input_tokens_seen": 1279583276,
+      "step": 9070
+    },
+    {
+      "epoch": 0.5533928068077066,
+      "grad_norm": 0.4749618172645569,
+      "learning_rate": 8e-05,
+      "loss": 1.4162,
+      "num_input_tokens_seen": 1281000036,
+      "step": 9080
+    },
+    {
+      "epoch": 0.5540022702513275,
+      "grad_norm": 0.5409462451934814,
+      "learning_rate": 8e-05,
+      "loss": 1.3868,
+      "num_input_tokens_seen": 1282445072,
+      "step": 9090
+    },
+    {
+      "epoch": 0.5546117336949483,
+      "grad_norm": 0.5022688508033752,
+      "learning_rate": 8e-05,
+      "loss": 1.4162,
+      "num_input_tokens_seen": 1283856564,
+      "step": 9100
+    },
+    {
+      "epoch": 0.5552211971385691,
+      "grad_norm": 0.4955657124519348,
+      "learning_rate": 8e-05,
+      "loss": 1.41,
+      "num_input_tokens_seen": 1285249116,
+      "step": 9110
+    },
+    {
+      "epoch": 0.55583066058219,
+      "grad_norm": 0.4351109564304352,
+      "learning_rate": 8e-05,
+      "loss": 1.4364,
+      "num_input_tokens_seen": 1286611828,
+      "step": 9120
+    },
+    {
+      "epoch": 0.5564401240258108,
+      "grad_norm": 0.47502756118774414,
+      "learning_rate": 8e-05,
+      "loss": 1.3688,
+      "num_input_tokens_seen": 1288059680,
+      "step": 9130
+    },
+    {
+      "epoch": 0.5570495874694316,
+      "grad_norm": 0.5219740867614746,
+      "learning_rate": 8e-05,
+      "loss": 1.3995,
+      "num_input_tokens_seen": 1289506464,
+      "step": 9140
+    },
+    {
+      "epoch": 0.5576590509130525,
+      "grad_norm": 0.5647391676902771,
+      "learning_rate": 8e-05,
+      "loss": 1.3867,
+      "num_input_tokens_seen": 1290890052,
+      "step": 9150
+    },
+    {
+      "epoch": 0.5582685143566732,
+      "grad_norm": 0.5420753359794617,
+      "learning_rate": 8e-05,
+      "loss": 1.4256,
+      "num_input_tokens_seen": 1292311544,
+      "step": 9160
+    },
+    {
+      "epoch": 0.558877977800294,
+      "grad_norm": 0.629084050655365,
+      "learning_rate": 8e-05,
+      "loss": 1.422,
+      "num_input_tokens_seen": 1293705252,
+      "step": 9170
+    },
+    {
+      "epoch": 0.5594874412439149,
+      "grad_norm": 0.5770242214202881,
+      "learning_rate": 8e-05,
+      "loss": 1.3623,
+      "num_input_tokens_seen": 1295093572,
+      "step": 9180
+    },
+    {
+      "epoch": 0.5600969046875357,
+      "grad_norm": 0.5409170985221863,
+      "learning_rate": 8e-05,
+      "loss": 1.4737,
+      "num_input_tokens_seen": 1296519364,
+      "step": 9190
+    },
+    {
+      "epoch": 0.5607063681311565,
+      "grad_norm": 0.5729120373725891,
+      "learning_rate": 8e-05,
+      "loss": 1.4459,
+      "num_input_tokens_seen": 1297897980,
+      "step": 9200
+    },
+    {
+      "epoch": 0.5613158315747774,
+      "grad_norm": 0.5346562266349792,
+      "learning_rate": 8e-05,
+      "loss": 1.4169,
+      "num_input_tokens_seen": 1299282408,
+      "step": 9210
+    },
+    {
+      "epoch": 0.5619252950183982,
+      "grad_norm": 0.6186701655387878,
+      "learning_rate": 8e-05,
+      "loss": 1.3818,
+      "num_input_tokens_seen": 1300736024,
+      "step": 9220
+    },
+    {
+      "epoch": 0.562534758462019,
+      "grad_norm": 0.5668373107910156,
+      "learning_rate": 8e-05,
+      "loss": 1.4676,
+      "num_input_tokens_seen": 1302156036,
+      "step": 9230
+    },
+    {
+      "epoch": 0.5631442219056398,
+      "grad_norm": 0.603315532207489,
+      "learning_rate": 8e-05,
+      "loss": 1.3285,
+      "num_input_tokens_seen": 1303545448,
+      "step": 9240
+    },
+    {
+      "epoch": 0.5637536853492606,
+      "grad_norm": 0.5525988936424255,
+      "learning_rate": 8e-05,
+      "loss": 1.4286,
+      "num_input_tokens_seen": 1304987840,
+      "step": 9250
+    },
+    {
+      "epoch": 0.5643631487928815,
+      "grad_norm": 0.5281280279159546,
+      "learning_rate": 8e-05,
+      "loss": 1.4153,
+      "num_input_tokens_seen": 1306349096,
+      "step": 9260
+    },
+    {
+      "epoch": 0.5649726122365023,
+      "grad_norm": 0.539382815361023,
+      "learning_rate": 8e-05,
+      "loss": 1.4315,
+      "num_input_tokens_seen": 1307741544,
+      "step": 9270
+    },
+    {
+      "epoch": 0.5655820756801231,
+      "grad_norm": 0.49413955211639404,
+      "learning_rate": 8e-05,
+      "loss": 1.422,
+      "num_input_tokens_seen": 1309121068,
+      "step": 9280
+    },
+    {
+      "epoch": 0.566191539123744,
+      "grad_norm": 0.7388852834701538,
+      "learning_rate": 8e-05,
+      "loss": 1.3662,
+      "num_input_tokens_seen": 1310517248,
+      "step": 9290
+    },
+    {
+      "epoch": 0.5668010025673648,
+      "grad_norm": 0.5051785707473755,
+      "learning_rate": 8e-05,
+      "loss": 1.3247,
+      "num_input_tokens_seen": 1311906244,
+      "step": 9300
+    },
+    {
+      "epoch": 0.5674104660109855,
+      "grad_norm": 0.5112394094467163,
+      "learning_rate": 8e-05,
+      "loss": 1.3472,
+      "num_input_tokens_seen": 1313330152,
+      "step": 9310
+    },
+    {
+      "epoch": 0.5680199294546064,
+      "grad_norm": 0.5090803503990173,
+      "learning_rate": 8e-05,
+      "loss": 1.3978,
+      "num_input_tokens_seen": 1314772320,
+      "step": 9320
+    },
+    {
+      "epoch": 0.5686293928982272,
+      "grad_norm": 0.4785972237586975,
+      "learning_rate": 8e-05,
+      "loss": 1.3579,
+      "num_input_tokens_seen": 1316168788,
+      "step": 9330
+    },
+    {
+      "epoch": 0.569238856341848,
+      "grad_norm": 0.565299928188324,
+      "learning_rate": 8e-05,
+      "loss": 1.386,
+      "num_input_tokens_seen": 1317603684,
+      "step": 9340
+    },
+    {
+      "epoch": 0.5698483197854689,
+      "grad_norm": 0.5050956010818481,
+      "learning_rate": 8e-05,
+      "loss": 1.4233,
+      "num_input_tokens_seen": 1319033440,
+      "step": 9350
+    },
+    {
+      "epoch": 0.5704577832290897,
+      "grad_norm": 0.5199946761131287,
+      "learning_rate": 8e-05,
+      "loss": 1.4047,
+      "num_input_tokens_seen": 1320462816,
+      "step": 9360
+    },
+    {
+      "epoch": 0.5710672466727105,
+      "grad_norm": 0.47691354155540466,
+      "learning_rate": 8e-05,
+      "loss": 1.3643,
+      "num_input_tokens_seen": 1321889000,
+      "step": 9370
+    },
+    {
+      "epoch": 0.5716767101163314,
+      "grad_norm": 0.5591989159584045,
+      "learning_rate": 8e-05,
+      "loss": 1.4623,
+      "num_input_tokens_seen": 1323308556,
+      "step": 9380
+    },
+    {
+      "epoch": 0.5722861735599521,
+      "grad_norm": 0.516518771648407,
+      "learning_rate": 8e-05,
+      "loss": 1.4165,
+      "num_input_tokens_seen": 1324702200,
+      "step": 9390
+    },
+    {
+      "epoch": 0.5728956370035729,
+      "grad_norm": 0.5647040009498596,
+      "learning_rate": 8e-05,
+      "loss": 1.3805,
+      "num_input_tokens_seen": 1326076652,
+      "step": 9400
+    },
+    {
+      "epoch": 0.5735051004471938,
+      "grad_norm": 0.5209704637527466,
+      "learning_rate": 8e-05,
+      "loss": 1.3816,
+      "num_input_tokens_seen": 1327492928,
+      "step": 9410
+    },
+    {
+      "epoch": 0.5741145638908146,
+      "grad_norm": 0.6285092234611511,
+      "learning_rate": 8e-05,
+      "loss": 1.3143,
+      "num_input_tokens_seen": 1328873088,
+      "step": 9420
+    },
+    {
+      "epoch": 0.5747240273344354,
+      "grad_norm": 0.42731142044067383,
+      "learning_rate": 8e-05,
+      "loss": 1.3784,
+      "num_input_tokens_seen": 1330316768,
+      "step": 9430
+    },
+    {
+      "epoch": 0.5753334907780563,
+      "grad_norm": 0.500942051410675,
+      "learning_rate": 8e-05,
+      "loss": 1.3928,
+      "num_input_tokens_seen": 1331718868,
+      "step": 9440
+    },
+    {
+      "epoch": 0.5759429542216771,
+      "grad_norm": 0.5587007403373718,
+      "learning_rate": 8e-05,
+      "loss": 1.423,
+      "num_input_tokens_seen": 1333111484,
+      "step": 9450
+    },
+    {
+      "epoch": 0.576552417665298,
+      "grad_norm": 0.48135367035865784,
+      "learning_rate": 8e-05,
+      "loss": 1.4287,
+      "num_input_tokens_seen": 1334496056,
+      "step": 9460
+    },
+    {
+      "epoch": 0.5771618811089188,
+      "grad_norm": 0.5771949887275696,
+      "learning_rate": 8e-05,
+      "loss": 1.4181,
+      "num_input_tokens_seen": 1335922768,
+      "step": 9470
+    },
+    {
+      "epoch": 0.5777713445525395,
+      "grad_norm": 0.5315853953361511,
+      "learning_rate": 8e-05,
+      "loss": 1.4274,
+      "num_input_tokens_seen": 1337302964,
+      "step": 9480
+    },
+    {
+      "epoch": 0.5783808079961604,
+      "grad_norm": 0.5134252309799194,
+      "learning_rate": 8e-05,
+      "loss": 1.4307,
+      "num_input_tokens_seen": 1338719692,
+      "step": 9490
+    },
+    {
+      "epoch": 0.5789902714397812,
+      "grad_norm": 0.5051198601722717,
+      "learning_rate": 8e-05,
+      "loss": 1.496,
+      "num_input_tokens_seen": 1340134348,
+      "step": 9500
+    },
+    {
+      "epoch": 0.579599734883402,
+      "grad_norm": 0.549379289150238,
+      "learning_rate": 8e-05,
+      "loss": 1.4031,
+      "num_input_tokens_seen": 1341522456,
+      "step": 9510
+    },
+    {
+      "epoch": 0.5802091983270229,
+      "grad_norm": 0.56511390209198,
+      "learning_rate": 8e-05,
+      "loss": 1.3683,
+      "num_input_tokens_seen": 1342933524,
+      "step": 9520
+    },
+    {
+      "epoch": 0.5808186617706437,
+      "grad_norm": 0.55156409740448,
+      "learning_rate": 8e-05,
+      "loss": 1.4012,
+      "num_input_tokens_seen": 1344352168,
+      "step": 9530
+    },
+    {
+      "epoch": 0.5814281252142645,
+      "grad_norm": 0.4940817654132843,
+      "learning_rate": 8e-05,
+      "loss": 1.4177,
+      "num_input_tokens_seen": 1345777572,
+      "step": 9540
+    },
+    {
+      "epoch": 0.5820375886578854,
+      "grad_norm": 0.4556514620780945,
+      "learning_rate": 8e-05,
+      "loss": 1.3985,
+      "num_input_tokens_seen": 1347204108,
+      "step": 9550
+    },
+    {
+      "epoch": 0.5826470521015061,
+      "grad_norm": 0.5312452912330627,
+      "learning_rate": 8e-05,
+      "loss": 1.3973,
+      "num_input_tokens_seen": 1348650808,
+      "step": 9560
+    },
+    {
+      "epoch": 0.5832565155451269,
+      "grad_norm": 0.5042407512664795,
+      "learning_rate": 8e-05,
+      "loss": 1.4426,
+      "num_input_tokens_seen": 1350015484,
+      "step": 9570
+    },
+    {
+      "epoch": 0.5838659789887478,
+      "grad_norm": 0.549752950668335,
+      "learning_rate": 8e-05,
+      "loss": 1.4271,
+      "num_input_tokens_seen": 1351431476,
+      "step": 9580
+    },
+    {
+      "epoch": 0.5844754424323686,
+      "grad_norm": 0.5132181644439697,
+      "learning_rate": 8e-05,
+      "loss": 1.3289,
+      "num_input_tokens_seen": 1352869096,
+      "step": 9590
+    },
+    {
+      "epoch": 0.5850849058759894,
+      "grad_norm": 0.6704673767089844,
+      "learning_rate": 8e-05,
+      "loss": 1.3606,
+      "num_input_tokens_seen": 1354273284,
+      "step": 9600
+    },
+    {
+      "epoch": 0.5856943693196103,
+      "grad_norm": 0.5381009578704834,
+      "learning_rate": 8e-05,
+      "loss": 1.3736,
+      "num_input_tokens_seen": 1355701548,
+      "step": 9610
+    },
+    {
+      "epoch": 0.5863038327632311,
+      "grad_norm": 0.4848478138446808,
+      "learning_rate": 8e-05,
+      "loss": 1.4624,
+      "num_input_tokens_seen": 1357098184,
+      "step": 9620
+    },
+    {
+      "epoch": 0.5869132962068518,
+      "grad_norm": 0.5218216180801392,
+      "learning_rate": 8e-05,
+      "loss": 1.3911,
+      "num_input_tokens_seen": 1358496652,
+      "step": 9630
+    },
+    {
+      "epoch": 0.5875227596504727,
+      "grad_norm": 0.5365182757377625,
+      "learning_rate": 8e-05,
+      "loss": 1.4491,
+      "num_input_tokens_seen": 1359933264,
+      "step": 9640
+    },
+    {
+      "epoch": 0.5881322230940935,
+      "grad_norm": 0.5306704044342041,
+      "learning_rate": 8e-05,
+      "loss": 1.4075,
+      "num_input_tokens_seen": 1361319892,
+      "step": 9650
+    },
+    {
+      "epoch": 0.5887416865377144,
+      "grad_norm": 0.5075612664222717,
+      "learning_rate": 8e-05,
+      "loss": 1.3719,
+      "num_input_tokens_seen": 1362739532,
+      "step": 9660
+    },
+    {
+      "epoch": 0.5893511499813352,
+      "grad_norm": 0.5192306041717529,
+      "learning_rate": 8e-05,
+      "loss": 1.3277,
+      "num_input_tokens_seen": 1364184264,
+      "step": 9670
+    },
+    {
+      "epoch": 0.589960613424956,
+      "grad_norm": 0.5479230880737305,
+      "learning_rate": 8e-05,
+      "loss": 1.4084,
+      "num_input_tokens_seen": 1365614432,
+      "step": 9680
+    },
+    {
+      "epoch": 0.5905700768685769,
+      "grad_norm": 0.6008402109146118,
+      "learning_rate": 8e-05,
+      "loss": 1.4572,
+      "num_input_tokens_seen": 1367052460,
+      "step": 9690
+    },
+    {
+      "epoch": 0.5911795403121977,
+      "grad_norm": 0.42700648307800293,
+      "learning_rate": 8e-05,
+      "loss": 1.3478,
+      "num_input_tokens_seen": 1368451128,
+      "step": 9700
+    },
+    {
+      "epoch": 0.5917890037558184,
+      "grad_norm": 0.5453005433082581,
+      "learning_rate": 8e-05,
+      "loss": 1.4283,
+      "num_input_tokens_seen": 1369849936,
+      "step": 9710
+    },
+    {
+      "epoch": 0.5923984671994393,
+      "grad_norm": 0.5360944271087646,
+      "learning_rate": 8e-05,
+      "loss": 1.4105,
+      "num_input_tokens_seen": 1371263976,
+      "step": 9720
+    },
+    {
+      "epoch": 0.5930079306430601,
+      "grad_norm": 0.5020470023155212,
+      "learning_rate": 8e-05,
+      "loss": 1.39,
+      "num_input_tokens_seen": 1372649640,
+      "step": 9730
+    },
+    {
+      "epoch": 0.5936173940866809,
+      "grad_norm": 0.6608033180236816,
+      "learning_rate": 8e-05,
+      "loss": 1.3897,
+      "num_input_tokens_seen": 1374020584,
+      "step": 9740
+    },
+    {
+      "epoch": 0.5942268575303018,
+      "grad_norm": 0.4657529294490814,
+      "learning_rate": 8e-05,
+      "loss": 1.3444,
+      "num_input_tokens_seen": 1375393228,
+      "step": 9750
+    },
+    {
+      "epoch": 0.5948363209739226,
+      "grad_norm": 0.6202511191368103,
+      "learning_rate": 8e-05,
+      "loss": 1.4421,
+      "num_input_tokens_seen": 1376785448,
+      "step": 9760
+    },
+    {
+      "epoch": 0.5954457844175434,
+      "grad_norm": 0.48330217599868774,
+      "learning_rate": 8e-05,
+      "loss": 1.3878,
+      "num_input_tokens_seen": 1378221156,
+      "step": 9770
+    },
+    {
+      "epoch": 0.5960552478611643,
+      "grad_norm": 0.5171040296554565,
+      "learning_rate": 8e-05,
+      "loss": 1.406,
+      "num_input_tokens_seen": 1379642012,
+      "step": 9780
+    },
+    {
+      "epoch": 0.596664711304785,
+      "grad_norm": 0.5539959669113159,
+      "learning_rate": 8e-05,
+      "loss": 1.3386,
+      "num_input_tokens_seen": 1381034376,
+      "step": 9790
+    },
+    {
+      "epoch": 0.5972741747484058,
+      "grad_norm": 0.5682281851768494,
+      "learning_rate": 8e-05,
+      "loss": 1.3227,
+      "num_input_tokens_seen": 1382464908,
+      "step": 9800
+    },
+    {
+      "epoch": 0.5978836381920267,
+      "grad_norm": 0.5348504781723022,
+      "learning_rate": 8e-05,
+      "loss": 1.3212,
+      "num_input_tokens_seen": 1383873968,
+      "step": 9810
+    },
+    {
+      "epoch": 0.5984931016356475,
+      "grad_norm": 0.5483260750770569,
+      "learning_rate": 8e-05,
+      "loss": 1.4158,
+      "num_input_tokens_seen": 1385281380,
+      "step": 9820
+    },
+    {
+      "epoch": 0.5991025650792683,
+      "grad_norm": 0.4727989137172699,
+      "learning_rate": 8e-05,
+      "loss": 1.3502,
+      "num_input_tokens_seen": 1386679784,
+      "step": 9830
+    },
+    {
+      "epoch": 0.5997120285228892,
+      "grad_norm": 0.5440919399261475,
+      "learning_rate": 8e-05,
+      "loss": 1.5048,
+      "num_input_tokens_seen": 1388088844,
+      "step": 9840
+    },
+    {
+      "epoch": 0.60032149196651,
+      "grad_norm": 0.5022852420806885,
+      "learning_rate": 8e-05,
+      "loss": 1.3446,
+      "num_input_tokens_seen": 1389562772,
+      "step": 9850
+    },
+    {
+      "epoch": 0.6009309554101308,
+      "grad_norm": 0.5559846758842468,
+      "learning_rate": 8e-05,
+      "loss": 1.4179,
+      "num_input_tokens_seen": 1390956260,
+      "step": 9860
+    },
+    {
+      "epoch": 0.6015404188537516,
+      "grad_norm": 0.5346333384513855,
+      "learning_rate": 8e-05,
+      "loss": 1.3865,
+      "num_input_tokens_seen": 1392385600,
+      "step": 9870
+    },
+    {
+      "epoch": 0.6021498822973724,
+      "grad_norm": 0.5209230780601501,
+      "learning_rate": 8e-05,
+      "loss": 1.3285,
+      "num_input_tokens_seen": 1393782632,
+      "step": 9880
+    },
+    {
+      "epoch": 0.6027593457409933,
+      "grad_norm": 0.5146684646606445,
+      "learning_rate": 8e-05,
+      "loss": 1.419,
+      "num_input_tokens_seen": 1395203648,
+      "step": 9890
+    },
+    {
+      "epoch": 0.6033688091846141,
+      "grad_norm": 0.49187934398651123,
+      "learning_rate": 8e-05,
+      "loss": 1.3834,
+      "num_input_tokens_seen": 1396636232,
+      "step": 9900
+    },
+    {
+      "epoch": 0.6039782726282349,
+      "grad_norm": 0.5156924724578857,
+      "learning_rate": 8e-05,
+      "loss": 1.3366,
+      "num_input_tokens_seen": 1398070672,
+      "step": 9910
+    },
+    {
+      "epoch": 0.6045877360718558,
+      "grad_norm": 0.5331023931503296,
+      "learning_rate": 8e-05,
+      "loss": 1.3713,
+      "num_input_tokens_seen": 1399473452,
+      "step": 9920
+    },
+    {
+      "epoch": 0.6051971995154766,
+      "grad_norm": 0.6120520234107971,
+      "learning_rate": 8e-05,
+      "loss": 1.4259,
+      "num_input_tokens_seen": 1400826964,
+      "step": 9930
+    },
+    {
+      "epoch": 0.6058066629590974,
+      "grad_norm": 0.4598010182380676,
+      "learning_rate": 8e-05,
+      "loss": 1.3581,
+      "num_input_tokens_seen": 1402235776,
+      "step": 9940
+    },
+    {
+      "epoch": 0.6064161264027182,
+      "grad_norm": 0.4815860688686371,
+      "learning_rate": 8e-05,
+      "loss": 1.3952,
+      "num_input_tokens_seen": 1403663484,
+      "step": 9950
+    },
+    {
+      "epoch": 0.607025589846339,
+      "grad_norm": 0.5356654524803162,
+      "learning_rate": 8e-05,
+      "loss": 1.4294,
+      "num_input_tokens_seen": 1405093388,
+      "step": 9960
+    },
+    {
+      "epoch": 0.6076350532899598,
+      "grad_norm": 0.4952469766139984,
+      "learning_rate": 8e-05,
+      "loss": 1.4177,
+      "num_input_tokens_seen": 1406509716,
+      "step": 9970
+    },
+    {
+      "epoch": 0.6082445167335807,
+      "grad_norm": 0.497787207365036,
+      "learning_rate": 8e-05,
+      "loss": 1.39,
+      "num_input_tokens_seen": 1407883080,
+      "step": 9980
+    },
+    {
+      "epoch": 0.6088539801772015,
+      "grad_norm": 0.41468948125839233,
+      "learning_rate": 8e-05,
+      "loss": 1.3927,
+      "num_input_tokens_seen": 1409283056,
+      "step": 9990
+    },
+    {
+      "epoch": 0.6094634436208223,
+      "grad_norm": 0.45570316910743713,
+      "learning_rate": 8e-05,
+      "loss": 1.289,
+      "num_input_tokens_seen": 1410699192,
+      "step": 10000
+    },
+    {
+      "epoch": 0.6100729070644432,
+      "grad_norm": 0.5052545666694641,
+      "learning_rate": 8e-05,
+      "loss": 1.3546,
+      "num_input_tokens_seen": 1412106116,
+      "step": 10010
+    },
+    {
+      "epoch": 0.610682370508064,
+      "grad_norm": 0.5244643688201904,
+      "learning_rate": 8e-05,
+      "loss": 1.3356,
+      "num_input_tokens_seen": 1413547016,
+      "step": 10020
+    },
+    {
+      "epoch": 0.6112918339516847,
+      "grad_norm": 0.5760179758071899,
+      "learning_rate": 8e-05,
+      "loss": 1.389,
+      "num_input_tokens_seen": 1414973076,
+      "step": 10030
+    },
+    {
+      "epoch": 0.6119012973953056,
+      "grad_norm": 0.5389821529388428,
+      "learning_rate": 8e-05,
+      "loss": 1.3859,
+      "num_input_tokens_seen": 1416396436,
+      "step": 10040
+    },
+    {
+      "epoch": 0.6125107608389264,
+      "grad_norm": 0.5720279812812805,
+      "learning_rate": 8e-05,
+      "loss": 1.349,
+      "num_input_tokens_seen": 1417828128,
+      "step": 10050
+    },
+    {
+      "epoch": 0.6131202242825472,
+      "grad_norm": 0.5402315258979797,
+      "learning_rate": 8e-05,
+      "loss": 1.4391,
+      "num_input_tokens_seen": 1419232568,
+      "step": 10060
+    },
+    {
+      "epoch": 0.6137296877261681,
+      "grad_norm": 0.5044508576393127,
+      "learning_rate": 8e-05,
+      "loss": 1.3909,
+      "num_input_tokens_seen": 1420627952,
+      "step": 10070
+    },
+    {
+      "epoch": 0.6143391511697889,
+      "grad_norm": 0.48339608311653137,
+      "learning_rate": 8e-05,
+      "loss": 1.3168,
+      "num_input_tokens_seen": 1422007340,
+      "step": 10080
+    },
+    {
+      "epoch": 0.6149486146134098,
+      "grad_norm": 0.4976171553134918,
+      "learning_rate": 8e-05,
+      "loss": 1.3754,
+      "num_input_tokens_seen": 1423445552,
+      "step": 10090
+    },
+    {
+      "epoch": 0.6155580780570306,
+      "grad_norm": 0.49609145522117615,
+      "learning_rate": 8e-05,
+      "loss": 1.3285,
+      "num_input_tokens_seen": 1424851720,
+      "step": 10100
+    },
+    {
+      "epoch": 0.6161675415006513,
+      "grad_norm": 0.5570063591003418,
+      "learning_rate": 8e-05,
+      "loss": 1.4115,
+      "num_input_tokens_seen": 1426280464,
+      "step": 10110
+    },
+    {
+      "epoch": 0.6167770049442722,
+      "grad_norm": 0.511796236038208,
+      "learning_rate": 8e-05,
+      "loss": 1.391,
+      "num_input_tokens_seen": 1427688620,
+      "step": 10120
+    },
+    {
+      "epoch": 0.617386468387893,
+      "grad_norm": 0.5466093420982361,
+      "learning_rate": 8e-05,
+      "loss": 1.3736,
+      "num_input_tokens_seen": 1429118608,
+      "step": 10130
+    },
+    {
+      "epoch": 0.6179959318315138,
+      "grad_norm": 0.5146467685699463,
+      "learning_rate": 8e-05,
+      "loss": 1.4094,
+      "num_input_tokens_seen": 1430487804,
+      "step": 10140
+    },
+    {
+      "epoch": 0.6186053952751347,
+      "grad_norm": 0.5151812434196472,
+      "learning_rate": 8e-05,
+      "loss": 1.3828,
+      "num_input_tokens_seen": 1431876676,
+      "step": 10150
+    },
+    {
+      "epoch": 0.6192148587187555,
+      "grad_norm": 0.5586668848991394,
+      "learning_rate": 8e-05,
+      "loss": 1.3261,
+      "num_input_tokens_seen": 1433285472,
+      "step": 10160
+    },
+    {
+      "epoch": 0.6198243221623763,
+      "grad_norm": 0.5817645788192749,
+      "learning_rate": 8e-05,
+      "loss": 1.4316,
+      "num_input_tokens_seen": 1434693624,
+      "step": 10170
+    },
+    {
+      "epoch": 0.6204337856059972,
+      "grad_norm": 0.6884422302246094,
+      "learning_rate": 8e-05,
+      "loss": 1.4434,
+      "num_input_tokens_seen": 1436113752,
+      "step": 10180
+    },
+    {
+      "epoch": 0.6210432490496179,
+      "grad_norm": 0.5640652179718018,
+      "learning_rate": 8e-05,
+      "loss": 1.3756,
+      "num_input_tokens_seen": 1437519516,
+      "step": 10190
+    },
+    {
+      "epoch": 0.6216527124932387,
+      "grad_norm": 0.49332916736602783,
+      "learning_rate": 8e-05,
+      "loss": 1.3701,
+      "num_input_tokens_seen": 1438919604,
+      "step": 10200
+    },
+    {
+      "epoch": 0.6222621759368596,
+      "grad_norm": 0.5096072554588318,
+      "learning_rate": 8e-05,
+      "loss": 1.3413,
+      "num_input_tokens_seen": 1440347500,
+      "step": 10210
+    },
+    {
+      "epoch": 0.6228716393804804,
+      "grad_norm": 0.44264650344848633,
+      "learning_rate": 8e-05,
+      "loss": 1.4,
+      "num_input_tokens_seen": 1441772520,
+      "step": 10220
+    },
+    {
+      "epoch": 0.6234811028241012,
+      "grad_norm": 0.4569202661514282,
+      "learning_rate": 8e-05,
+      "loss": 1.3895,
+      "num_input_tokens_seen": 1443207380,
+      "step": 10230
+    },
+    {
+      "epoch": 0.6240905662677221,
+      "grad_norm": 0.5005388855934143,
+      "learning_rate": 8e-05,
+      "loss": 1.3637,
+      "num_input_tokens_seen": 1444575216,
+      "step": 10240
+    },
+    {
+      "epoch": 0.6247000297113429,
+      "grad_norm": 0.5004163384437561,
+      "learning_rate": 8e-05,
+      "loss": 1.4176,
+      "num_input_tokens_seen": 1445983064,
+      "step": 10250
+    },
+    {
+      "epoch": 0.6253094931549636,
+      "grad_norm": 0.5014557242393494,
+      "learning_rate": 8e-05,
+      "loss": 1.3762,
+      "num_input_tokens_seen": 1447412308,
+      "step": 10260
+    },
+    {
+      "epoch": 0.6259189565985845,
+      "grad_norm": 0.5607008337974548,
+      "learning_rate": 8e-05,
+      "loss": 1.3564,
+      "num_input_tokens_seen": 1448812476,
+      "step": 10270
+    },
+    {
+      "epoch": 0.6265284200422053,
+      "grad_norm": 0.5385088324546814,
+      "learning_rate": 8e-05,
+      "loss": 1.324,
+      "num_input_tokens_seen": 1450237216,
+      "step": 10280
+    },
+    {
+      "epoch": 0.6271378834858262,
+      "grad_norm": 0.5262163281440735,
+      "learning_rate": 8e-05,
+      "loss": 1.4133,
+      "num_input_tokens_seen": 1451658832,
+      "step": 10290
+    },
+    {
+      "epoch": 0.627747346929447,
+      "grad_norm": 0.7675474286079407,
+      "learning_rate": 8e-05,
+      "loss": 1.3457,
+      "num_input_tokens_seen": 1453044480,
+      "step": 10300
+    },
+    {
+      "epoch": 0.6283568103730678,
+      "grad_norm": 0.48304829001426697,
+      "learning_rate": 8e-05,
+      "loss": 1.3758,
+      "num_input_tokens_seen": 1454441984,
+      "step": 10310
+    },
+    {
+      "epoch": 0.6289662738166887,
+      "grad_norm": 0.49265825748443604,
+      "learning_rate": 8e-05,
+      "loss": 1.3955,
+      "num_input_tokens_seen": 1455862852,
+      "step": 10320
+    },
+    {
+      "epoch": 0.6295757372603095,
+      "grad_norm": 0.4803888499736786,
+      "learning_rate": 8e-05,
+      "loss": 1.3637,
+      "num_input_tokens_seen": 1457252176,
+      "step": 10330
+    },
+    {
+      "epoch": 0.6301852007039302,
+      "grad_norm": 0.4618512690067291,
+      "learning_rate": 8e-05,
+      "loss": 1.3878,
+      "num_input_tokens_seen": 1458690980,
+      "step": 10340
+    },
+    {
+      "epoch": 0.6307946641475511,
+      "grad_norm": 0.486068457365036,
+      "learning_rate": 8e-05,
+      "loss": 1.3597,
+      "num_input_tokens_seen": 1460126928,
+      "step": 10350
+    },
+    {
+      "epoch": 0.6314041275911719,
+      "grad_norm": 0.482714980840683,
+      "learning_rate": 8e-05,
+      "loss": 1.3995,
+      "num_input_tokens_seen": 1461522124,
+      "step": 10360
+    },
+    {
+      "epoch": 0.6320135910347927,
+      "grad_norm": 0.5192288756370544,
+      "learning_rate": 8e-05,
+      "loss": 1.4576,
+      "num_input_tokens_seen": 1462943012,
+      "step": 10370
+    },
+    {
+      "epoch": 0.6326230544784136,
+      "grad_norm": 0.4682476818561554,
+      "learning_rate": 8e-05,
+      "loss": 1.3796,
+      "num_input_tokens_seen": 1464333524,
+      "step": 10380
+    },
+    {
+      "epoch": 0.6332325179220344,
+      "grad_norm": 0.5397130250930786,
+      "learning_rate": 8e-05,
+      "loss": 1.3822,
+      "num_input_tokens_seen": 1465743076,
+      "step": 10390
+    },
+    {
+      "epoch": 0.6338419813656552,
+      "grad_norm": 0.5656686425209045,
+      "learning_rate": 8e-05,
+      "loss": 1.3642,
+      "num_input_tokens_seen": 1467142196,
+      "step": 10400
+    },
+    {
+      "epoch": 0.6344514448092761,
+      "grad_norm": 0.6183952689170837,
+      "learning_rate": 8e-05,
+      "loss": 1.3855,
+      "num_input_tokens_seen": 1468562740,
+      "step": 10410
+    },
+    {
+      "epoch": 0.6350609082528968,
+      "grad_norm": 0.4923710227012634,
+      "learning_rate": 8e-05,
+      "loss": 1.3049,
+      "num_input_tokens_seen": 1469960856,
+      "step": 10420
+    },
+    {
+      "epoch": 0.6356703716965176,
+      "grad_norm": 0.6067305207252502,
+      "learning_rate": 8e-05,
+      "loss": 1.3892,
+      "num_input_tokens_seen": 1471400408,
+      "step": 10430
+    },
+    {
+      "epoch": 0.6362798351401385,
+      "grad_norm": 0.5619065761566162,
+      "learning_rate": 8e-05,
+      "loss": 1.4251,
+      "num_input_tokens_seen": 1472802868,
+      "step": 10440
+    },
+    {
+      "epoch": 0.6368892985837593,
+      "grad_norm": 0.5561144351959229,
+      "learning_rate": 8e-05,
+      "loss": 1.4278,
+      "num_input_tokens_seen": 1474193064,
+      "step": 10450
+    },
+    {
+      "epoch": 0.6374987620273801,
+      "grad_norm": 0.5016257762908936,
+      "learning_rate": 8e-05,
+      "loss": 1.3927,
+      "num_input_tokens_seen": 1475612660,
+      "step": 10460
+    },
+    {
+      "epoch": 0.638108225471001,
+      "grad_norm": 0.4828028380870819,
+      "learning_rate": 8e-05,
+      "loss": 1.3506,
+      "num_input_tokens_seen": 1477059616,
+      "step": 10470
+    },
+    {
+      "epoch": 0.6387176889146218,
+      "grad_norm": 0.7187060713768005,
+      "learning_rate": 8e-05,
+      "loss": 1.3665,
+      "num_input_tokens_seen": 1478402568,
+      "step": 10480
+    },
+    {
+      "epoch": 0.6393271523582427,
+      "grad_norm": 0.6509134769439697,
+      "learning_rate": 8e-05,
+      "loss": 1.4104,
+      "num_input_tokens_seen": 1479782212,
+      "step": 10490
+    },
+    {
+      "epoch": 0.6399366158018635,
+      "grad_norm": 0.5177718997001648,
+      "learning_rate": 8e-05,
+      "loss": 1.2858,
+      "num_input_tokens_seen": 1481194272,
+      "step": 10500
+    },
+    {
+      "epoch": 0.6405460792454842,
+      "grad_norm": 0.542962908744812,
+      "learning_rate": 8e-05,
+      "loss": 1.4324,
+      "num_input_tokens_seen": 1482567368,
+      "step": 10510
+    },
+    {
+      "epoch": 0.6411555426891051,
+      "grad_norm": 0.5583025217056274,
+      "learning_rate": 8e-05,
+      "loss": 1.3911,
+      "num_input_tokens_seen": 1484014248,
+      "step": 10520
+    },
+    {
+      "epoch": 0.6417650061327259,
+      "grad_norm": 0.4803900718688965,
+      "learning_rate": 8e-05,
+      "loss": 1.4556,
+      "num_input_tokens_seen": 1485396060,
+      "step": 10530
+    },
+    {
+      "epoch": 0.6423744695763467,
+      "grad_norm": 0.4976584017276764,
+      "learning_rate": 8e-05,
+      "loss": 1.3407,
+      "num_input_tokens_seen": 1486803768,
+      "step": 10540
+    },
+    {
+      "epoch": 0.6429839330199676,
+      "grad_norm": 0.6051338911056519,
+      "learning_rate": 8e-05,
+      "loss": 1.4611,
+      "num_input_tokens_seen": 1488205756,
+      "step": 10550
+    },
+    {
+      "epoch": 0.6435933964635884,
+      "grad_norm": 0.5262444615364075,
+      "learning_rate": 8e-05,
+      "loss": 1.3961,
+      "num_input_tokens_seen": 1489615612,
+      "step": 10560
+    },
+    {
+      "epoch": 0.6442028599072092,
+      "grad_norm": 0.484791100025177,
+      "learning_rate": 8e-05,
+      "loss": 1.4121,
+      "num_input_tokens_seen": 1491011084,
+      "step": 10570
+    },
+    {
+      "epoch": 0.64481232335083,
+      "grad_norm": 0.5317909717559814,
+      "learning_rate": 8e-05,
+      "loss": 1.3567,
+      "num_input_tokens_seen": 1492429136,
+      "step": 10580
+    },
+    {
+      "epoch": 0.6454217867944508,
+      "grad_norm": 0.5404983162879944,
+      "learning_rate": 8e-05,
+      "loss": 1.3642,
+      "num_input_tokens_seen": 1493834740,
+      "step": 10590
+    },
+    {
+      "epoch": 0.6460312502380716,
+      "grad_norm": 0.538868248462677,
+      "learning_rate": 8e-05,
+      "loss": 1.3155,
+      "num_input_tokens_seen": 1495264904,
+      "step": 10600
+    },
+    {
+      "epoch": 0.6466407136816925,
+      "grad_norm": 0.5658362507820129,
+      "learning_rate": 8e-05,
+      "loss": 1.3228,
+      "num_input_tokens_seen": 1496659280,
+      "step": 10610
+    },
+    {
+      "epoch": 0.6472501771253133,
+      "grad_norm": 0.5738557577133179,
+      "learning_rate": 8e-05,
+      "loss": 1.3978,
+      "num_input_tokens_seen": 1498064828,
+      "step": 10620
+    },
+    {
+      "epoch": 0.6478596405689341,
+      "grad_norm": 0.5166726112365723,
+      "learning_rate": 8e-05,
+      "loss": 1.4355,
+      "num_input_tokens_seen": 1499487812,
+      "step": 10630
+    },
+    {
+      "epoch": 0.648469104012555,
+      "grad_norm": 0.4428934156894684,
+      "learning_rate": 8e-05,
+      "loss": 1.3573,
+      "num_input_tokens_seen": 1500892436,
+      "step": 10640
+    },
+    {
+      "epoch": 0.6490785674561758,
+      "grad_norm": 0.559181809425354,
+      "learning_rate": 8e-05,
+      "loss": 1.4064,
+      "num_input_tokens_seen": 1502337536,
+      "step": 10650
+    },
+    {
+      "epoch": 0.6496880308997965,
+      "grad_norm": 0.46578314900398254,
+      "learning_rate": 8e-05,
+      "loss": 1.3296,
+      "num_input_tokens_seen": 1503735980,
+      "step": 10660
+    },
+    {
+      "epoch": 0.6502974943434174,
+      "grad_norm": 0.5593947768211365,
+      "learning_rate": 8e-05,
+      "loss": 1.444,
+      "num_input_tokens_seen": 1505164384,
+      "step": 10670
+    },
+    {
+      "epoch": 0.6509069577870382,
+      "grad_norm": 2.141988754272461,
+      "learning_rate": 8e-05,
+      "loss": 1.3533,
+      "num_input_tokens_seen": 1506537700,
+      "step": 10680
+    },
+    {
+      "epoch": 0.651516421230659,
+      "grad_norm": 0.6614646315574646,
+      "learning_rate": 8e-05,
+      "loss": 1.4131,
+      "num_input_tokens_seen": 1507983352,
+      "step": 10690
+    },
+    {
+      "epoch": 0.6521258846742799,
+      "grad_norm": 0.49016857147216797,
+      "learning_rate": 8e-05,
+      "loss": 1.389,
+      "num_input_tokens_seen": 1509361728,
+      "step": 10700
+    },
+    {
+      "epoch": 0.6527353481179007,
+      "grad_norm": 0.5343895554542542,
+      "learning_rate": 8e-05,
+      "loss": 1.4211,
+      "num_input_tokens_seen": 1510712804,
+      "step": 10710
+    },
+    {
+      "epoch": 0.6533448115615216,
+      "grad_norm": 0.44679155945777893,
+      "learning_rate": 8e-05,
+      "loss": 1.3696,
+      "num_input_tokens_seen": 1512120068,
+      "step": 10720
+    },
+    {
+      "epoch": 0.6539542750051424,
+      "grad_norm": 0.5805819630622864,
+      "learning_rate": 8e-05,
+      "loss": 1.3976,
+      "num_input_tokens_seen": 1513546256,
+      "step": 10730
+    },
+    {
+      "epoch": 0.6545637384487631,
+      "grad_norm": 0.5583277940750122,
+      "learning_rate": 8e-05,
+      "loss": 1.3012,
+      "num_input_tokens_seen": 1514925448,
+      "step": 10740
+    },
+    {
+      "epoch": 0.655173201892384,
+      "grad_norm": 0.49840471148490906,
+      "learning_rate": 8e-05,
+      "loss": 1.462,
+      "num_input_tokens_seen": 1516367584,
+      "step": 10750
+    },
+    {
+      "epoch": 0.6557826653360048,
+      "grad_norm": 0.4973022937774658,
+      "learning_rate": 8e-05,
+      "loss": 1.3537,
+      "num_input_tokens_seen": 1517757988,
+      "step": 10760
+    },
+    {
+      "epoch": 0.6563921287796256,
+      "grad_norm": 0.5231379866600037,
+      "learning_rate": 8e-05,
+      "loss": 1.3669,
+      "num_input_tokens_seen": 1519164136,
+      "step": 10770
+    },
+    {
+      "epoch": 0.6570015922232465,
+      "grad_norm": 0.5220736861228943,
+      "learning_rate": 8e-05,
+      "loss": 1.3633,
+      "num_input_tokens_seen": 1520588564,
+      "step": 10780
+    },
+    {
+      "epoch": 0.6576110556668673,
+      "grad_norm": 0.5354141592979431,
+      "learning_rate": 8e-05,
+      "loss": 1.4131,
+      "num_input_tokens_seen": 1522010348,
+      "step": 10790
+    },
+    {
+      "epoch": 0.6582205191104881,
+      "grad_norm": 0.5864176154136658,
+      "learning_rate": 8e-05,
+      "loss": 1.3681,
+      "num_input_tokens_seen": 1523415004,
+      "step": 10800
+    },
+    {
+      "epoch": 0.658829982554109,
+      "grad_norm": 0.5721861720085144,
+      "learning_rate": 8e-05,
+      "loss": 1.308,
+      "num_input_tokens_seen": 1524815832,
+      "step": 10810
+    },
+    {
+      "epoch": 0.6594394459977297,
+      "grad_norm": 0.6083350777626038,
+      "learning_rate": 8e-05,
+      "loss": 1.3567,
+      "num_input_tokens_seen": 1526230472,
+      "step": 10820
+    },
+    {
+      "epoch": 0.6600489094413505,
+      "grad_norm": 0.46110281348228455,
+      "learning_rate": 8e-05,
+      "loss": 1.4186,
+      "num_input_tokens_seen": 1527644032,
+      "step": 10830
+    },
+    {
+      "epoch": 0.6606583728849714,
+      "grad_norm": 0.5084540843963623,
+      "learning_rate": 8e-05,
+      "loss": 1.32,
+      "num_input_tokens_seen": 1529057312,
+      "step": 10840
+    },
+    {
+      "epoch": 0.6612678363285922,
+      "grad_norm": 0.5809466242790222,
+      "learning_rate": 8e-05,
+      "loss": 1.3788,
+      "num_input_tokens_seen": 1530445328,
+      "step": 10850
+    },
+    {
+      "epoch": 0.661877299772213,
+      "grad_norm": 0.46439307928085327,
+      "learning_rate": 8e-05,
+      "loss": 1.3833,
+      "num_input_tokens_seen": 1531850316,
+      "step": 10860
+    },
+    {
+      "epoch": 0.6624867632158339,
+      "grad_norm": 0.5628945231437683,
+      "learning_rate": 8e-05,
+      "loss": 1.3472,
+      "num_input_tokens_seen": 1533252868,
+      "step": 10870
+    },
+    {
+      "epoch": 0.6630962266594547,
+      "grad_norm": 0.6179889440536499,
+      "learning_rate": 8e-05,
+      "loss": 1.4127,
+      "num_input_tokens_seen": 1534660880,
+      "step": 10880
+    },
+    {
+      "epoch": 0.6637056901030755,
+      "grad_norm": 0.5281222462654114,
+      "learning_rate": 8e-05,
+      "loss": 1.4029,
+      "num_input_tokens_seen": 1536068992,
+      "step": 10890
+    },
+    {
+      "epoch": 0.6643151535466963,
+      "grad_norm": 0.5171144008636475,
+      "learning_rate": 8e-05,
+      "loss": 1.4088,
+      "num_input_tokens_seen": 1537464796,
+      "step": 10900
+    },
+    {
+      "epoch": 0.6649246169903171,
+      "grad_norm": 0.529052197933197,
+      "learning_rate": 8e-05,
+      "loss": 1.3975,
+      "num_input_tokens_seen": 1538875160,
+      "step": 10910
+    },
+    {
+      "epoch": 0.665534080433938,
+      "grad_norm": 0.5157914757728577,
+      "learning_rate": 8e-05,
+      "loss": 1.3578,
+      "num_input_tokens_seen": 1540279084,
+      "step": 10920
+    },
+    {
+      "epoch": 0.6661435438775588,
+      "grad_norm": 0.5008856058120728,
+      "learning_rate": 8e-05,
+      "loss": 1.3453,
+      "num_input_tokens_seen": 1541714696,
+      "step": 10930
+    },
+    {
+      "epoch": 0.6667530073211796,
+      "grad_norm": 0.45337000489234924,
+      "learning_rate": 8e-05,
+      "loss": 1.4766,
+      "num_input_tokens_seen": 1543170740,
+      "step": 10940
+    },
+    {
+      "epoch": 0.6673624707648005,
+      "grad_norm": 0.5083340406417847,
+      "learning_rate": 8e-05,
+      "loss": 1.4436,
+      "num_input_tokens_seen": 1544592796,
+      "step": 10950
+    },
+    {
+      "epoch": 0.6679719342084213,
+      "grad_norm": 0.48253244161605835,
+      "learning_rate": 8e-05,
+      "loss": 1.3303,
+      "num_input_tokens_seen": 1545997384,
+      "step": 10960
+    },
+    {
+      "epoch": 0.668581397652042,
+      "grad_norm": 0.4723127484321594,
+      "learning_rate": 8e-05,
+      "loss": 1.3656,
+      "num_input_tokens_seen": 1547390872,
+      "step": 10970
+    },
+    {
+      "epoch": 0.669190861095663,
+      "grad_norm": 0.539252758026123,
+      "learning_rate": 8e-05,
+      "loss": 1.3333,
+      "num_input_tokens_seen": 1548803244,
+      "step": 10980
+    },
+    {
+      "epoch": 0.6698003245392837,
+      "grad_norm": 0.5402015447616577,
+      "learning_rate": 8e-05,
+      "loss": 1.434,
+      "num_input_tokens_seen": 1550175492,
+      "step": 10990
+    },
+    {
+      "epoch": 0.6704097879829045,
+      "grad_norm": 0.6111288070678711,
+      "learning_rate": 8e-05,
+      "loss": 1.3002,
+      "num_input_tokens_seen": 1551566972,
+      "step": 11000
+    },
+    {
+      "epoch": 0.6710192514265254,
+      "grad_norm": 0.5599400997161865,
+      "learning_rate": 8e-05,
+      "loss": 1.3756,
+      "num_input_tokens_seen": 1552990300,
+      "step": 11010
+    },
+    {
+      "epoch": 0.6716287148701462,
+      "grad_norm": 0.4863987863063812,
+      "learning_rate": 8e-05,
+      "loss": 1.3602,
+      "num_input_tokens_seen": 1554395040,
+      "step": 11020
+    },
+    {
+      "epoch": 0.672238178313767,
+      "grad_norm": 0.5732718706130981,
+      "learning_rate": 8e-05,
+      "loss": 1.3012,
+      "num_input_tokens_seen": 1555814488,
+      "step": 11030
+    },
+    {
+      "epoch": 0.6728476417573879,
+      "grad_norm": 0.48249438405036926,
+      "learning_rate": 8e-05,
+      "loss": 1.3947,
+      "num_input_tokens_seen": 1557230564,
+      "step": 11040
+    },
+    {
+      "epoch": 0.6734571052010087,
+      "grad_norm": 0.5319753289222717,
+      "learning_rate": 8e-05,
+      "loss": 1.3969,
+      "num_input_tokens_seen": 1558697588,
+      "step": 11050
+    },
+    {
+      "epoch": 0.6740665686446294,
+      "grad_norm": 0.4940206706523895,
+      "learning_rate": 8e-05,
+      "loss": 1.3153,
+      "num_input_tokens_seen": 1560096556,
+      "step": 11060
+    },
+    {
+      "epoch": 0.6746760320882503,
+      "grad_norm": 0.5515936017036438,
+      "learning_rate": 8e-05,
+      "loss": 1.3674,
+      "num_input_tokens_seen": 1561517816,
+      "step": 11070
+    },
+    {
+      "epoch": 0.6752854955318711,
+      "grad_norm": 0.4390547573566437,
+      "learning_rate": 8e-05,
+      "loss": 1.3617,
+      "num_input_tokens_seen": 1562934520,
+      "step": 11080
+    },
+    {
+      "epoch": 0.6758949589754919,
+      "grad_norm": 0.5946500897407532,
+      "learning_rate": 8e-05,
+      "loss": 1.4425,
+      "num_input_tokens_seen": 1564382372,
+      "step": 11090
+    },
+    {
+      "epoch": 0.6765044224191128,
+      "grad_norm": 0.46804702281951904,
+      "learning_rate": 8e-05,
+      "loss": 1.295,
+      "num_input_tokens_seen": 1565750404,
+      "step": 11100
+    },
+    {
+      "epoch": 0.6771138858627336,
+      "grad_norm": 0.5131279826164246,
+      "learning_rate": 8e-05,
+      "loss": 1.4145,
+      "num_input_tokens_seen": 1567168144,
+      "step": 11110
+    },
+    {
+      "epoch": 0.6777233493063545,
+      "grad_norm": 0.5205957293510437,
+      "learning_rate": 8e-05,
+      "loss": 1.3669,
+      "num_input_tokens_seen": 1568518060,
+      "step": 11120
+    },
+    {
+      "epoch": 0.6783328127499753,
+      "grad_norm": 0.4875277876853943,
+      "learning_rate": 8e-05,
+      "loss": 1.4563,
+      "num_input_tokens_seen": 1569937112,
+      "step": 11130
+    },
+    {
+      "epoch": 0.678942276193596,
+      "grad_norm": 0.5259339809417725,
+      "learning_rate": 8e-05,
+      "loss": 1.3498,
+      "num_input_tokens_seen": 1571385304,
+      "step": 11140
+    },
+    {
+      "epoch": 0.6795517396372169,
+      "grad_norm": 0.48126786947250366,
+      "learning_rate": 8e-05,
+      "loss": 1.3787,
+      "num_input_tokens_seen": 1572780332,
+      "step": 11150
+    },
+    {
+      "epoch": 0.6801612030808377,
+      "grad_norm": 0.45843496918678284,
+      "learning_rate": 8e-05,
+      "loss": 1.3826,
+      "num_input_tokens_seen": 1574176036,
+      "step": 11160
+    },
+    {
+      "epoch": 0.6807706665244585,
+      "grad_norm": 0.5293328762054443,
+      "learning_rate": 8e-05,
+      "loss": 1.3625,
+      "num_input_tokens_seen": 1575584344,
+      "step": 11170
+    },
+    {
+      "epoch": 0.6813801299680794,
+      "grad_norm": 0.4788746237754822,
+      "learning_rate": 8e-05,
+      "loss": 1.401,
+      "num_input_tokens_seen": 1577029088,
+      "step": 11180
+    },
+    {
+      "epoch": 0.6819895934117002,
+      "grad_norm": 0.5486621260643005,
+      "learning_rate": 8e-05,
+      "loss": 1.4309,
+      "num_input_tokens_seen": 1578437036,
+      "step": 11190
+    },
+    {
+      "epoch": 0.682599056855321,
+      "grad_norm": 0.5115844011306763,
+      "learning_rate": 8e-05,
+      "loss": 1.417,
+      "num_input_tokens_seen": 1579839792,
+      "step": 11200
+    },
+    {
+      "epoch": 0.6832085202989419,
+      "grad_norm": 0.46927279233932495,
+      "learning_rate": 8e-05,
+      "loss": 1.3891,
+      "num_input_tokens_seen": 1581258564,
+      "step": 11210
+    },
+    {
+      "epoch": 0.6838179837425626,
+      "grad_norm": 0.4486519694328308,
+      "learning_rate": 8e-05,
+      "loss": 1.3672,
+      "num_input_tokens_seen": 1582668792,
+      "step": 11220
+    },
+    {
+      "epoch": 0.6844274471861834,
+      "grad_norm": 0.5005569458007812,
+      "learning_rate": 8e-05,
+      "loss": 1.421,
+      "num_input_tokens_seen": 1584050876,
+      "step": 11230
+    },
+    {
+      "epoch": 0.6850369106298043,
+      "grad_norm": 0.605232834815979,
+      "learning_rate": 8e-05,
+      "loss": 1.3457,
+      "num_input_tokens_seen": 1585466308,
+      "step": 11240
+    },
+    {
+      "epoch": 0.6856463740734251,
+      "grad_norm": 0.5172209739685059,
+      "learning_rate": 8e-05,
+      "loss": 1.3576,
+      "num_input_tokens_seen": 1586888240,
+      "step": 11250
+    },
+    {
+      "epoch": 0.6862558375170459,
+      "grad_norm": 0.47500714659690857,
+      "learning_rate": 8e-05,
+      "loss": 1.2991,
+      "num_input_tokens_seen": 1588319356,
+      "step": 11260
+    },
+    {
+      "epoch": 0.6868653009606668,
+      "grad_norm": 0.48329707980155945,
+      "learning_rate": 8e-05,
+      "loss": 1.3285,
+      "num_input_tokens_seen": 1589717748,
+      "step": 11270
+    },
+    {
+      "epoch": 0.6874747644042876,
+      "grad_norm": 0.5334200859069824,
+      "learning_rate": 8e-05,
+      "loss": 1.3861,
+      "num_input_tokens_seen": 1591160404,
+      "step": 11280
+    },
+    {
+      "epoch": 0.6880842278479083,
+      "grad_norm": 0.5618347525596619,
+      "learning_rate": 8e-05,
+      "loss": 1.3787,
+      "num_input_tokens_seen": 1592539424,
+      "step": 11290
+    },
+    {
+      "epoch": 0.6886936912915292,
+      "grad_norm": 0.47227922081947327,
+      "learning_rate": 8e-05,
+      "loss": 1.3481,
+      "num_input_tokens_seen": 1593941668,
+      "step": 11300
+    },
+    {
+      "epoch": 0.68930315473515,
+      "grad_norm": 0.4863712191581726,
+      "learning_rate": 8e-05,
+      "loss": 1.3482,
+      "num_input_tokens_seen": 1595349516,
+      "step": 11310
+    },
+    {
+      "epoch": 0.6899126181787709,
+      "grad_norm": 0.4943729341030121,
+      "learning_rate": 8e-05,
+      "loss": 1.3168,
+      "num_input_tokens_seen": 1596768640,
+      "step": 11320
+    },
+    {
+      "epoch": 0.6905220816223917,
+      "grad_norm": 0.5097691416740417,
+      "learning_rate": 8e-05,
+      "loss": 1.3259,
+      "num_input_tokens_seen": 1598213372,
+      "step": 11330
+    },
+    {
+      "epoch": 0.6911315450660125,
+      "grad_norm": 0.5228952169418335,
+      "learning_rate": 8e-05,
+      "loss": 1.4481,
+      "num_input_tokens_seen": 1599627692,
+      "step": 11340
+    },
+    {
+      "epoch": 0.6917410085096334,
+      "grad_norm": 0.4985556900501251,
+      "learning_rate": 8e-05,
+      "loss": 1.3566,
+      "num_input_tokens_seen": 1601010164,
+      "step": 11350
+    },
+    {
+      "epoch": 0.6923504719532542,
+      "grad_norm": 0.4366033673286438,
+      "learning_rate": 8e-05,
+      "loss": 1.27,
+      "num_input_tokens_seen": 1602418184,
+      "step": 11360
+    },
+    {
+      "epoch": 0.692959935396875,
+      "grad_norm": 0.43515148758888245,
+      "learning_rate": 8e-05,
+      "loss": 1.3676,
+      "num_input_tokens_seen": 1603824332,
+      "step": 11370
+    },
+    {
+      "epoch": 0.6935693988404958,
+      "grad_norm": 0.5262019634246826,
+      "learning_rate": 8e-05,
+      "loss": 1.3643,
+      "num_input_tokens_seen": 1605236940,
+      "step": 11380
+    },
+    {
+      "epoch": 0.6941788622841166,
+      "grad_norm": 0.5392087697982788,
+      "learning_rate": 8e-05,
+      "loss": 1.4025,
+      "num_input_tokens_seen": 1606675864,
+      "step": 11390
+    },
+    {
+      "epoch": 0.6947883257277374,
+      "grad_norm": 0.6022461652755737,
+      "learning_rate": 8e-05,
+      "loss": 1.408,
+      "num_input_tokens_seen": 1608094576,
+      "step": 11400
+    },
+    {
+      "epoch": 0.6953977891713583,
+      "grad_norm": 0.6072622537612915,
+      "learning_rate": 8e-05,
+      "loss": 1.3024,
+      "num_input_tokens_seen": 1609523344,
+      "step": 11410
+    },
+    {
+      "epoch": 0.6960072526149791,
+      "grad_norm": 0.5051203370094299,
+      "learning_rate": 8e-05,
+      "loss": 1.3671,
+      "num_input_tokens_seen": 1610934128,
+      "step": 11420
+    },
+    {
+      "epoch": 0.6966167160585999,
+      "grad_norm": 0.5395517349243164,
+      "learning_rate": 8e-05,
+      "loss": 1.3802,
+      "num_input_tokens_seen": 1612402656,
+      "step": 11430
+    },
+    {
+      "epoch": 0.6972261795022208,
+      "grad_norm": 0.5100081562995911,
+      "learning_rate": 8e-05,
+      "loss": 1.4357,
+      "num_input_tokens_seen": 1613828080,
+      "step": 11440
+    },
+    {
+      "epoch": 0.6978356429458415,
+      "grad_norm": 0.5012816190719604,
+      "learning_rate": 8e-05,
+      "loss": 1.3875,
+      "num_input_tokens_seen": 1615257472,
+      "step": 11450
+    },
+    {
+      "epoch": 0.6984451063894623,
+      "grad_norm": 0.5558944344520569,
+      "learning_rate": 8e-05,
+      "loss": 1.3962,
+      "num_input_tokens_seen": 1616629904,
+      "step": 11460
+    },
+    {
+      "epoch": 0.6990545698330832,
+      "grad_norm": 0.5226157307624817,
+      "learning_rate": 8e-05,
+      "loss": 1.3542,
+      "num_input_tokens_seen": 1618030980,
+      "step": 11470
+    },
+    {
+      "epoch": 0.699664033276704,
+      "grad_norm": 0.5259307622909546,
+      "learning_rate": 8e-05,
+      "loss": 1.3689,
+      "num_input_tokens_seen": 1619450276,
+      "step": 11480
+    },
+    {
+      "epoch": 0.7002734967203248,
+      "grad_norm": 0.5086653232574463,
+      "learning_rate": 8e-05,
+      "loss": 1.3484,
+      "num_input_tokens_seen": 1620874576,
+      "step": 11490
+    },
+    {
+      "epoch": 0.7008829601639457,
+      "grad_norm": 0.5540332198143005,
+      "learning_rate": 8e-05,
+      "loss": 1.3293,
+      "num_input_tokens_seen": 1622283328,
+      "step": 11500
+    },
+    {
+      "epoch": 0.7014924236075665,
+      "grad_norm": 0.5071319341659546,
+      "learning_rate": 8e-05,
+      "loss": 1.3199,
+      "num_input_tokens_seen": 1623641908,
+      "step": 11510
+    },
+    {
+      "epoch": 0.7021018870511873,
+      "grad_norm": 0.5358554124832153,
+      "learning_rate": 8e-05,
+      "loss": 1.4036,
+      "num_input_tokens_seen": 1625041264,
+      "step": 11520
+    },
+    {
+      "epoch": 0.7027113504948082,
+      "grad_norm": 0.5506737232208252,
+      "learning_rate": 8e-05,
+      "loss": 1.34,
+      "num_input_tokens_seen": 1626404844,
+      "step": 11530
+    },
+    {
+      "epoch": 0.7033208139384289,
+      "grad_norm": 0.44404277205467224,
+      "learning_rate": 8e-05,
+      "loss": 1.34,
+      "num_input_tokens_seen": 1627863172,
+      "step": 11540
+    },
+    {
+      "epoch": 0.7039302773820498,
+      "grad_norm": 0.46990063786506653,
+      "learning_rate": 8e-05,
+      "loss": 1.3722,
+      "num_input_tokens_seen": 1629274832,
+      "step": 11550
+    },
+    {
+      "epoch": 0.7045397408256706,
+      "grad_norm": 0.4636783301830292,
+      "learning_rate": 8e-05,
+      "loss": 1.3307,
+      "num_input_tokens_seen": 1630697752,
+      "step": 11560
+    },
+    {
+      "epoch": 0.7051492042692914,
+      "grad_norm": 0.43071427941322327,
+      "learning_rate": 8e-05,
+      "loss": 1.3574,
+      "num_input_tokens_seen": 1632095920,
+      "step": 11570
+    },
+    {
+      "epoch": 0.7057586677129123,
+      "grad_norm": 0.5207687616348267,
+      "learning_rate": 8e-05,
+      "loss": 1.3084,
+      "num_input_tokens_seen": 1633509780,
+      "step": 11580
+    },
+    {
+      "epoch": 0.7063681311565331,
+      "grad_norm": 0.4385841190814972,
+      "learning_rate": 8e-05,
+      "loss": 1.3712,
+      "num_input_tokens_seen": 1634958136,
+      "step": 11590
+    },
+    {
+      "epoch": 0.7069775946001539,
+      "grad_norm": 0.5395679473876953,
+      "learning_rate": 8e-05,
+      "loss": 1.3802,
+      "num_input_tokens_seen": 1636385408,
+      "step": 11600
+    },
+    {
+      "epoch": 0.7075870580437748,
+      "grad_norm": 0.605144202709198,
+      "learning_rate": 8e-05,
+      "loss": 1.3814,
+      "num_input_tokens_seen": 1637775280,
+      "step": 11610
+    },
+    {
+      "epoch": 0.7081965214873955,
+      "grad_norm": 0.46157678961753845,
+      "learning_rate": 8e-05,
+      "loss": 1.378,
+      "num_input_tokens_seen": 1639212620,
+      "step": 11620
+    },
+    {
+      "epoch": 0.7088059849310163,
+      "grad_norm": 0.4572867751121521,
+      "learning_rate": 8e-05,
+      "loss": 1.3809,
+      "num_input_tokens_seen": 1640607368,
+      "step": 11630
+    },
+    {
+      "epoch": 0.7094154483746372,
+      "grad_norm": 0.5079594254493713,
+      "learning_rate": 8e-05,
+      "loss": 1.2521,
+      "num_input_tokens_seen": 1642027784,
+      "step": 11640
+    },
+    {
+      "epoch": 0.710024911818258,
+      "grad_norm": 0.7235115766525269,
+      "learning_rate": 8e-05,
+      "loss": 1.3339,
+      "num_input_tokens_seen": 1643408672,
+      "step": 11650
+    },
+    {
+      "epoch": 0.7106343752618788,
+      "grad_norm": 0.459494411945343,
+      "learning_rate": 8e-05,
+      "loss": 1.3244,
+      "num_input_tokens_seen": 1644816008,
+      "step": 11660
+    },
+    {
+      "epoch": 0.7112438387054997,
+      "grad_norm": 0.4589853584766388,
+      "learning_rate": 8e-05,
+      "loss": 1.3231,
+      "num_input_tokens_seen": 1646220640,
+      "step": 11670
+    },
+    {
+      "epoch": 0.7118533021491205,
+      "grad_norm": 0.496324360370636,
+      "learning_rate": 8e-05,
+      "loss": 1.3495,
+      "num_input_tokens_seen": 1647628668,
+      "step": 11680
+    },
+    {
+      "epoch": 0.7124627655927412,
+      "grad_norm": 0.5813272595405579,
+      "learning_rate": 8e-05,
+      "loss": 1.3994,
+      "num_input_tokens_seen": 1649031944,
+      "step": 11690
+    },
+    {
+      "epoch": 0.7130722290363621,
+      "grad_norm": 0.528102695941925,
+      "learning_rate": 8e-05,
+      "loss": 1.4662,
+      "num_input_tokens_seen": 1650445412,
+      "step": 11700
+    },
+    {
+      "epoch": 0.7136816924799829,
+      "grad_norm": 0.5562940835952759,
+      "learning_rate": 8e-05,
+      "loss": 1.3603,
+      "num_input_tokens_seen": 1651821844,
+      "step": 11710
+    },
+    {
+      "epoch": 0.7142911559236037,
+      "grad_norm": 0.4890764057636261,
+      "learning_rate": 8e-05,
+      "loss": 1.3721,
+      "num_input_tokens_seen": 1653199044,
+      "step": 11720
+    },
+    {
+      "epoch": 0.7149006193672246,
+      "grad_norm": 0.5230799913406372,
+      "learning_rate": 8e-05,
+      "loss": 1.3574,
+      "num_input_tokens_seen": 1654565212,
+      "step": 11730
+    },
+    {
+      "epoch": 0.7155100828108454,
+      "grad_norm": 0.5011894106864929,
+      "learning_rate": 8e-05,
+      "loss": 1.3551,
+      "num_input_tokens_seen": 1655952188,
+      "step": 11740
+    },
+    {
+      "epoch": 0.7161195462544663,
+      "grad_norm": 0.48514172434806824,
+      "learning_rate": 8e-05,
+      "loss": 1.3334,
+      "num_input_tokens_seen": 1657357812,
+      "step": 11750
+    },
+    {
+      "epoch": 0.7167290096980871,
+      "grad_norm": 0.5416278839111328,
+      "learning_rate": 8e-05,
+      "loss": 1.3142,
+      "num_input_tokens_seen": 1658753864,
+      "step": 11760
+    },
+    {
+      "epoch": 0.7173384731417078,
+      "grad_norm": 0.5038822889328003,
+      "learning_rate": 8e-05,
+      "loss": 1.3044,
+      "num_input_tokens_seen": 1660171056,
+      "step": 11770
+    },
+    {
+      "epoch": 0.7179479365853287,
+      "grad_norm": 0.6000845432281494,
+      "learning_rate": 8e-05,
+      "loss": 1.3487,
+      "num_input_tokens_seen": 1661578792,
+      "step": 11780
+    },
+    {
+      "epoch": 0.7185574000289495,
+      "grad_norm": 0.4599187672138214,
+      "learning_rate": 8e-05,
+      "loss": 1.3169,
+      "num_input_tokens_seen": 1662961256,
+      "step": 11790
+    },
+    {
+      "epoch": 0.7191668634725703,
+      "grad_norm": 0.5955362915992737,
+      "learning_rate": 8e-05,
+      "loss": 1.4204,
+      "num_input_tokens_seen": 1664363192,
+      "step": 11800
+    },
+    {
+      "epoch": 0.7197763269161912,
+      "grad_norm": 0.5447018146514893,
+      "learning_rate": 8e-05,
+      "loss": 1.3821,
+      "num_input_tokens_seen": 1665776148,
+      "step": 11810
+    },
+    {
+      "epoch": 0.720385790359812,
+      "grad_norm": 0.49061569571495056,
+      "learning_rate": 8e-05,
+      "loss": 1.3661,
+      "num_input_tokens_seen": 1667199220,
+      "step": 11820
+    },
+    {
+      "epoch": 0.7209952538034328,
+      "grad_norm": 0.5659445524215698,
+      "learning_rate": 8e-05,
+      "loss": 1.3536,
+      "num_input_tokens_seen": 1668633580,
+      "step": 11830
+    },
+    {
+      "epoch": 0.7216047172470537,
+      "grad_norm": 0.44457441568374634,
+      "learning_rate": 8e-05,
+      "loss": 1.3176,
+      "num_input_tokens_seen": 1670048000,
+      "step": 11840
+    },
+    {
+      "epoch": 0.7222141806906744,
+      "grad_norm": 0.5029119849205017,
+      "learning_rate": 8e-05,
+      "loss": 1.4107,
+      "num_input_tokens_seen": 1671464340,
+      "step": 11850
+    },
+    {
+      "epoch": 0.7228236441342952,
+      "grad_norm": 0.4228610694408417,
+      "learning_rate": 8e-05,
+      "loss": 1.4419,
+      "num_input_tokens_seen": 1672855328,
+      "step": 11860
+    },
+    {
+      "epoch": 0.7234331075779161,
+      "grad_norm": 0.5061122179031372,
+      "learning_rate": 8e-05,
+      "loss": 1.3356,
+      "num_input_tokens_seen": 1674291744,
+      "step": 11870
+    },
+    {
+      "epoch": 0.7240425710215369,
+      "grad_norm": 0.4675373136997223,
+      "learning_rate": 8e-05,
+      "loss": 1.4002,
+      "num_input_tokens_seen": 1675717044,
+      "step": 11880
+    },
+    {
+      "epoch": 0.7246520344651577,
+      "grad_norm": 0.40757638216018677,
+      "learning_rate": 8e-05,
+      "loss": 1.3313,
+      "num_input_tokens_seen": 1677101540,
+      "step": 11890
+    },
+    {
+      "epoch": 0.7252614979087786,
+      "grad_norm": 0.5157292485237122,
+      "learning_rate": 8e-05,
+      "loss": 1.4145,
+      "num_input_tokens_seen": 1678515036,
+      "step": 11900
+    },
+    {
+      "epoch": 0.7258709613523994,
+      "grad_norm": 0.6204097270965576,
+      "learning_rate": 8e-05,
+      "loss": 1.3039,
+      "num_input_tokens_seen": 1679921248,
+      "step": 11910
+    },
+    {
+      "epoch": 0.7264804247960202,
+      "grad_norm": 0.44126880168914795,
+      "learning_rate": 8e-05,
+      "loss": 1.3353,
+      "num_input_tokens_seen": 1681306344,
+      "step": 11920
+    },
+    {
+      "epoch": 0.727089888239641,
+      "grad_norm": 0.6292856931686401,
+      "learning_rate": 8e-05,
+      "loss": 1.3355,
+      "num_input_tokens_seen": 1682719392,
+      "step": 11930
+    },
+    {
+      "epoch": 0.7276993516832618,
+      "grad_norm": 0.4729043245315552,
+      "learning_rate": 8e-05,
+      "loss": 1.2812,
+      "num_input_tokens_seen": 1684071768,
+      "step": 11940
+    },
+    {
+      "epoch": 0.7283088151268827,
+      "grad_norm": 0.484418660402298,
+      "learning_rate": 8e-05,
+      "loss": 1.3574,
+      "num_input_tokens_seen": 1685486372,
+      "step": 11950
+    },
+    {
+      "epoch": 0.7289182785705035,
+      "grad_norm": 0.48124608397483826,
+      "learning_rate": 8e-05,
+      "loss": 1.3325,
+      "num_input_tokens_seen": 1686917640,
+      "step": 11960
+    },
+    {
+      "epoch": 0.7295277420141243,
+      "grad_norm": 0.5151872634887695,
+      "learning_rate": 8e-05,
+      "loss": 1.3391,
+      "num_input_tokens_seen": 1688341640,
+      "step": 11970
+    },
+    {
+      "epoch": 0.7301372054577452,
+      "grad_norm": 0.5516366958618164,
+      "learning_rate": 8e-05,
+      "loss": 1.3006,
+      "num_input_tokens_seen": 1689765020,
+      "step": 11980
+    },
+    {
+      "epoch": 0.730746668901366,
+      "grad_norm": 0.5208513140678406,
+      "learning_rate": 8e-05,
+      "loss": 1.3503,
+      "num_input_tokens_seen": 1691179912,
+      "step": 11990
+    },
+    {
+      "epoch": 0.7313561323449868,
+      "grad_norm": 0.5036225914955139,
+      "learning_rate": 8e-05,
+      "loss": 1.3422,
+      "num_input_tokens_seen": 1692541944,
+      "step": 12000
+    },
+    {
+      "epoch": 0.7319655957886076,
+      "grad_norm": 0.44400182366371155,
+      "learning_rate": 8e-05,
+      "loss": 1.3675,
+      "num_input_tokens_seen": 1693937504,
+      "step": 12010
+    },
+    {
+      "epoch": 0.7325750592322284,
+      "grad_norm": 0.4907507598400116,
+      "learning_rate": 8e-05,
+      "loss": 1.3248,
+      "num_input_tokens_seen": 1695314680,
+      "step": 12020
+    },
+    {
+      "epoch": 0.7331845226758492,
+      "grad_norm": 0.5090814828872681,
+      "learning_rate": 8e-05,
+      "loss": 1.3548,
+      "num_input_tokens_seen": 1696725392,
+      "step": 12030
+    },
+    {
+      "epoch": 0.7337939861194701,
+      "grad_norm": 0.5077126622200012,
+      "learning_rate": 8e-05,
+      "loss": 1.3663,
+      "num_input_tokens_seen": 1698124780,
+      "step": 12040
+    },
+    {
+      "epoch": 0.7344034495630909,
+      "grad_norm": 0.5142135620117188,
+      "learning_rate": 8e-05,
+      "loss": 1.3235,
+      "num_input_tokens_seen": 1699533240,
+      "step": 12050
+    },
+    {
+      "epoch": 0.7350129130067117,
+      "grad_norm": 0.4585704207420349,
+      "learning_rate": 8e-05,
+      "loss": 1.3834,
+      "num_input_tokens_seen": 1700921096,
+      "step": 12060
+    },
+    {
+      "epoch": 0.7356223764503326,
+      "grad_norm": 0.6136194467544556,
+      "learning_rate": 8e-05,
+      "loss": 1.3458,
+      "num_input_tokens_seen": 1702319564,
+      "step": 12070
+    },
+    {
+      "epoch": 0.7362318398939534,
+      "grad_norm": 0.5828477740287781,
+      "learning_rate": 8e-05,
+      "loss": 1.3506,
+      "num_input_tokens_seen": 1703697712,
+      "step": 12080
+    },
+    {
+      "epoch": 0.7368413033375741,
+      "grad_norm": 0.5217992663383484,
+      "learning_rate": 8e-05,
+      "loss": 1.3363,
+      "num_input_tokens_seen": 1705110788,
+      "step": 12090
+    },
+    {
+      "epoch": 0.737450766781195,
+      "grad_norm": 0.5703504681587219,
+      "learning_rate": 8e-05,
+      "loss": 1.3916,
+      "num_input_tokens_seen": 1706529936,
+      "step": 12100
+    },
+    {
+      "epoch": 0.7380602302248158,
+      "grad_norm": 0.49754780530929565,
+      "learning_rate": 8e-05,
+      "loss": 1.3259,
+      "num_input_tokens_seen": 1707949384,
+      "step": 12110
+    },
+    {
+      "epoch": 0.7386696936684366,
+      "grad_norm": 0.46623051166534424,
+      "learning_rate": 8e-05,
+      "loss": 1.3169,
+      "num_input_tokens_seen": 1709343816,
+      "step": 12120
+    },
+    {
+      "epoch": 0.7392791571120575,
+      "grad_norm": 0.4529217481613159,
+      "learning_rate": 8e-05,
+      "loss": 1.3445,
+      "num_input_tokens_seen": 1710727904,
+      "step": 12130
+    },
+    {
+      "epoch": 0.7398886205556783,
+      "grad_norm": 0.5647047758102417,
+      "learning_rate": 8e-05,
+      "loss": 1.359,
+      "num_input_tokens_seen": 1712104092,
+      "step": 12140
+    },
+    {
+      "epoch": 0.7404980839992991,
+      "grad_norm": 0.4240126609802246,
+      "learning_rate": 8e-05,
+      "loss": 1.3276,
+      "num_input_tokens_seen": 1713494984,
+      "step": 12150
+    },
+    {
+      "epoch": 0.74110754744292,
+      "grad_norm": 0.4749116003513336,
+      "learning_rate": 8e-05,
+      "loss": 1.3626,
+      "num_input_tokens_seen": 1714923848,
+      "step": 12160
+    },
+    {
+      "epoch": 0.7417170108865407,
+      "grad_norm": 0.4912160038948059,
+      "learning_rate": 8e-05,
+      "loss": 1.3342,
+      "num_input_tokens_seen": 1716289192,
+      "step": 12170
+    },
+    {
+      "epoch": 0.7423264743301616,
+      "grad_norm": 0.47284796833992004,
+      "learning_rate": 8e-05,
+      "loss": 1.4014,
+      "num_input_tokens_seen": 1717724568,
+      "step": 12180
+    },
+    {
+      "epoch": 0.7429359377737824,
+      "grad_norm": 0.520675539970398,
+      "learning_rate": 8e-05,
+      "loss": 1.3207,
+      "num_input_tokens_seen": 1719167576,
+      "step": 12190
+    },
+    {
+      "epoch": 0.7435454012174032,
+      "grad_norm": 0.48047661781311035,
+      "learning_rate": 8e-05,
+      "loss": 1.288,
+      "num_input_tokens_seen": 1720576612,
+      "step": 12200
+    },
+    {
+      "epoch": 0.7441548646610241,
+      "grad_norm": 0.49878841638565063,
+      "learning_rate": 8e-05,
+      "loss": 1.3228,
+      "num_input_tokens_seen": 1721972364,
+      "step": 12210
+    },
+    {
+      "epoch": 0.7447643281046449,
+      "grad_norm": 0.4979226589202881,
+      "learning_rate": 8e-05,
+      "loss": 1.2814,
+      "num_input_tokens_seen": 1723431204,
+      "step": 12220
+    },
+    {
+      "epoch": 0.7453737915482657,
+      "grad_norm": 0.5070583820343018,
+      "learning_rate": 8e-05,
+      "loss": 1.3283,
+      "num_input_tokens_seen": 1724849792,
+      "step": 12230
+    },
+    {
+      "epoch": 0.7459832549918866,
+      "grad_norm": 0.4756496250629425,
+      "learning_rate": 8e-05,
+      "loss": 1.2717,
+      "num_input_tokens_seen": 1726305880,
+      "step": 12240
+    },
+    {
+      "epoch": 0.7465927184355073,
+      "grad_norm": 0.6402299404144287,
+      "learning_rate": 8e-05,
+      "loss": 1.3142,
+      "num_input_tokens_seen": 1727716648,
+      "step": 12250
+    },
+    {
+      "epoch": 0.7472021818791281,
+      "grad_norm": 0.5206665396690369,
+      "learning_rate": 8e-05,
+      "loss": 1.2944,
+      "num_input_tokens_seen": 1729132756,
+      "step": 12260
+    },
+    {
+      "epoch": 0.747811645322749,
+      "grad_norm": 0.5346994996070862,
+      "learning_rate": 8e-05,
+      "loss": 1.3813,
+      "num_input_tokens_seen": 1730565924,
+      "step": 12270
+    },
+    {
+      "epoch": 0.7484211087663698,
+      "grad_norm": 0.6076343059539795,
+      "learning_rate": 8e-05,
+      "loss": 1.3387,
+      "num_input_tokens_seen": 1731972796,
+      "step": 12280
+    },
+    {
+      "epoch": 0.7490305722099906,
+      "grad_norm": 0.5422996878623962,
+      "learning_rate": 8e-05,
+      "loss": 1.3192,
+      "num_input_tokens_seen": 1733365796,
+      "step": 12290
+    },
+    {
+      "epoch": 0.7496400356536115,
+      "grad_norm": 0.5899409055709839,
+      "learning_rate": 8e-05,
+      "loss": 1.3258,
+      "num_input_tokens_seen": 1734719368,
+      "step": 12300
+    },
+    {
+      "epoch": 0.7502494990972323,
+      "grad_norm": 0.4870734214782715,
+      "learning_rate": 8e-05,
+      "loss": 1.3172,
+      "num_input_tokens_seen": 1736156068,
+      "step": 12310
+    },
+    {
+      "epoch": 0.750858962540853,
+      "grad_norm": 0.44740596413612366,
+      "learning_rate": 8e-05,
+      "loss": 1.3336,
+      "num_input_tokens_seen": 1737561880,
+      "step": 12320
+    },
+    {
+      "epoch": 0.7514684259844739,
+      "grad_norm": 0.5884395241737366,
+      "learning_rate": 8e-05,
+      "loss": 1.3935,
+      "num_input_tokens_seen": 1738988796,
+      "step": 12330
+    },
+    {
+      "epoch": 0.7520778894280947,
+      "grad_norm": 0.5457208156585693,
+      "learning_rate": 8e-05,
+      "loss": 1.338,
+      "num_input_tokens_seen": 1740413288,
+      "step": 12340
+    },
+    {
+      "epoch": 0.7526873528717155,
+      "grad_norm": 0.4471927881240845,
+      "learning_rate": 8e-05,
+      "loss": 1.3132,
+      "num_input_tokens_seen": 1741852120,
+      "step": 12350
+    },
+    {
+      "epoch": 0.7532968163153364,
+      "grad_norm": 0.5655962228775024,
+      "learning_rate": 8e-05,
+      "loss": 1.265,
+      "num_input_tokens_seen": 1743221408,
+      "step": 12360
+    },
+    {
+      "epoch": 0.7539062797589572,
+      "grad_norm": 0.5733407735824585,
+      "learning_rate": 8e-05,
+      "loss": 1.3406,
+      "num_input_tokens_seen": 1744609012,
+      "step": 12370
+    },
+    {
+      "epoch": 0.7545157432025781,
+      "grad_norm": 0.5545304417610168,
+      "learning_rate": 8e-05,
+      "loss": 1.411,
+      "num_input_tokens_seen": 1746007428,
+      "step": 12380
+    },
+    {
+      "epoch": 0.7551252066461989,
+      "grad_norm": 0.5492582321166992,
+      "learning_rate": 8e-05,
+      "loss": 1.3644,
+      "num_input_tokens_seen": 1747422652,
+      "step": 12390
+    },
+    {
+      "epoch": 0.7557346700898196,
+      "grad_norm": 0.5435560345649719,
+      "learning_rate": 8e-05,
+      "loss": 1.37,
+      "num_input_tokens_seen": 1748808948,
+      "step": 12400
+    },
+    {
+      "epoch": 0.7563441335334405,
+      "grad_norm": 0.5100287795066833,
+      "learning_rate": 8e-05,
+      "loss": 1.3401,
+      "num_input_tokens_seen": 1750211396,
+      "step": 12410
+    },
+    {
+      "epoch": 0.7569535969770613,
+      "grad_norm": 0.5885925889015198,
+      "learning_rate": 8e-05,
+      "loss": 1.2552,
+      "num_input_tokens_seen": 1751641492,
+      "step": 12420
+    },
+    {
+      "epoch": 0.7575630604206821,
+      "grad_norm": 0.5710425972938538,
+      "learning_rate": 8e-05,
+      "loss": 1.3211,
+      "num_input_tokens_seen": 1753033724,
+      "step": 12430
+    },
+    {
+      "epoch": 0.758172523864303,
+      "grad_norm": 0.4631437659263611,
+      "learning_rate": 8e-05,
+      "loss": 1.3819,
+      "num_input_tokens_seen": 1754426200,
+      "step": 12440
+    },
+    {
+      "epoch": 0.7587819873079238,
+      "grad_norm": 0.5151511430740356,
+      "learning_rate": 8e-05,
+      "loss": 1.2802,
+      "num_input_tokens_seen": 1755811048,
+      "step": 12450
+    },
+    {
+      "epoch": 0.7593914507515446,
+      "grad_norm": 0.5333474278450012,
+      "learning_rate": 8e-05,
+      "loss": 1.3127,
+      "num_input_tokens_seen": 1757194620,
+      "step": 12460
+    },
+    {
+      "epoch": 0.7600009141951655,
+      "grad_norm": 0.49914056062698364,
+      "learning_rate": 8e-05,
+      "loss": 1.3185,
+      "num_input_tokens_seen": 1758622592,
+      "step": 12470
+    },
+    {
+      "epoch": 0.7606103776387863,
+      "grad_norm": 0.41091620922088623,
+      "learning_rate": 8e-05,
+      "loss": 1.2696,
+      "num_input_tokens_seen": 1759999208,
+      "step": 12480
+    },
+    {
+      "epoch": 0.761219841082407,
+      "grad_norm": 0.5769081115722656,
+      "learning_rate": 8e-05,
+      "loss": 1.3431,
+      "num_input_tokens_seen": 1761407552,
+      "step": 12490
+    },
+    {
+      "epoch": 0.7618293045260279,
+      "grad_norm": 0.5206863880157471,
+      "learning_rate": 8e-05,
+      "loss": 1.3317,
+      "num_input_tokens_seen": 1762800588,
+      "step": 12500
+    },
+    {
+      "epoch": 0.7624387679696487,
+      "grad_norm": 0.5158757567405701,
+      "learning_rate": 8e-05,
+      "loss": 1.3526,
+      "num_input_tokens_seen": 1764187880,
+      "step": 12510
+    },
+    {
+      "epoch": 0.7630482314132695,
+      "grad_norm": 0.5738706588745117,
+      "learning_rate": 8e-05,
+      "loss": 1.3989,
+      "num_input_tokens_seen": 1765586032,
+      "step": 12520
+    },
+    {
+      "epoch": 0.7636576948568904,
+      "grad_norm": 0.6022568941116333,
+      "learning_rate": 8e-05,
+      "loss": 1.2905,
+      "num_input_tokens_seen": 1767057448,
+      "step": 12530
+    },
+    {
+      "epoch": 0.7642671583005112,
+      "grad_norm": 0.5038744211196899,
+      "learning_rate": 8e-05,
+      "loss": 1.3278,
+      "num_input_tokens_seen": 1768443564,
+      "step": 12540
+    },
+    {
+      "epoch": 0.764876621744132,
+      "grad_norm": 0.5244400501251221,
+      "learning_rate": 8e-05,
+      "loss": 1.3228,
+      "num_input_tokens_seen": 1769849304,
+      "step": 12550
+    },
+    {
+      "epoch": 0.7654860851877529,
+      "grad_norm": 0.5055208802223206,
+      "learning_rate": 8e-05,
+      "loss": 1.3335,
+      "num_input_tokens_seen": 1771248796,
+      "step": 12560
+    },
+    {
+      "epoch": 0.7660955486313736,
+      "grad_norm": 0.48623979091644287,
+      "learning_rate": 8e-05,
+      "loss": 1.3765,
+      "num_input_tokens_seen": 1772625404,
+      "step": 12570
+    },
+    {
+      "epoch": 0.7667050120749945,
+      "grad_norm": 0.5019470453262329,
+      "learning_rate": 8e-05,
+      "loss": 1.3249,
+      "num_input_tokens_seen": 1774037708,
+      "step": 12580
+    },
+    {
+      "epoch": 0.7673144755186153,
+      "grad_norm": 0.5067726373672485,
+      "learning_rate": 8e-05,
+      "loss": 1.3194,
+      "num_input_tokens_seen": 1775423152,
+      "step": 12590
+    },
+    {
+      "epoch": 0.7679239389622361,
+      "grad_norm": 0.4977276027202606,
+      "learning_rate": 8e-05,
+      "loss": 1.3447,
+      "num_input_tokens_seen": 1776840448,
+      "step": 12600
+    },
+    {
+      "epoch": 0.768533402405857,
+      "grad_norm": 0.5764220356941223,
+      "learning_rate": 8e-05,
+      "loss": 1.3149,
+      "num_input_tokens_seen": 1778260160,
+      "step": 12610
+    },
+    {
+      "epoch": 0.7691428658494778,
+      "grad_norm": 0.5561099648475647,
+      "learning_rate": 8e-05,
+      "loss": 1.416,
+      "num_input_tokens_seen": 1779627344,
+      "step": 12620
+    },
+    {
+      "epoch": 0.7697523292930986,
+      "grad_norm": 0.4993090331554413,
+      "learning_rate": 8e-05,
+      "loss": 1.4345,
+      "num_input_tokens_seen": 1781020624,
+      "step": 12630
+    },
+    {
+      "epoch": 0.7703617927367195,
+      "grad_norm": 0.537917971611023,
+      "learning_rate": 8e-05,
+      "loss": 1.2816,
+      "num_input_tokens_seen": 1782457444,
+      "step": 12640
+    },
+    {
+      "epoch": 0.7709712561803402,
+      "grad_norm": 0.5599529147148132,
+      "learning_rate": 8e-05,
+      "loss": 1.3272,
+      "num_input_tokens_seen": 1783910788,
+      "step": 12650
+    },
+    {
+      "epoch": 0.771580719623961,
+      "grad_norm": 0.523169755935669,
+      "learning_rate": 8e-05,
+      "loss": 1.3633,
+      "num_input_tokens_seen": 1785320612,
+      "step": 12660
+    },
+    {
+      "epoch": 0.7721901830675819,
+      "grad_norm": 0.5435786247253418,
+      "learning_rate": 8e-05,
+      "loss": 1.2906,
+      "num_input_tokens_seen": 1786698656,
+      "step": 12670
+    },
+    {
+      "epoch": 0.7727996465112027,
+      "grad_norm": 0.489886999130249,
+      "learning_rate": 8e-05,
+      "loss": 1.365,
+      "num_input_tokens_seen": 1788067732,
+      "step": 12680
+    },
+    {
+      "epoch": 0.7734091099548235,
+      "grad_norm": 0.43583112955093384,
+      "learning_rate": 8e-05,
+      "loss": 1.2961,
+      "num_input_tokens_seen": 1789457292,
+      "step": 12690
+    },
+    {
+      "epoch": 0.7740185733984444,
+      "grad_norm": 0.4963041841983795,
+      "learning_rate": 8e-05,
+      "loss": 1.3571,
+      "num_input_tokens_seen": 1790862480,
+      "step": 12700
+    },
+    {
+      "epoch": 0.7746280368420652,
+      "grad_norm": 0.5106602907180786,
+      "learning_rate": 8e-05,
+      "loss": 1.4132,
+      "num_input_tokens_seen": 1792265412,
+      "step": 12710
+    },
+    {
+      "epoch": 0.7752375002856859,
+      "grad_norm": 0.47308048605918884,
+      "learning_rate": 8e-05,
+      "loss": 1.2768,
+      "num_input_tokens_seen": 1793681304,
+      "step": 12720
+    },
+    {
+      "epoch": 0.7758469637293068,
+      "grad_norm": 0.5513538718223572,
+      "learning_rate": 8e-05,
+      "loss": 1.4489,
+      "num_input_tokens_seen": 1795090636,
+      "step": 12730
+    },
+    {
+      "epoch": 0.7764564271729276,
+      "grad_norm": 0.5152673125267029,
+      "learning_rate": 8e-05,
+      "loss": 1.3806,
+      "num_input_tokens_seen": 1796505112,
+      "step": 12740
+    },
+    {
+      "epoch": 0.7770658906165484,
+      "grad_norm": 0.5017542243003845,
+      "learning_rate": 8e-05,
+      "loss": 1.2905,
+      "num_input_tokens_seen": 1797913208,
+      "step": 12750
+    },
+    {
+      "epoch": 0.7776753540601693,
+      "grad_norm": 0.513664186000824,
+      "learning_rate": 8e-05,
+      "loss": 1.4124,
+      "num_input_tokens_seen": 1799342844,
+      "step": 12760
+    },
+    {
+      "epoch": 0.7782848175037901,
+      "grad_norm": 0.47624465823173523,
+      "learning_rate": 8e-05,
+      "loss": 1.3233,
+      "num_input_tokens_seen": 1800772036,
+      "step": 12770
+    },
+    {
+      "epoch": 0.778894280947411,
+      "grad_norm": 0.5275976061820984,
+      "learning_rate": 8e-05,
+      "loss": 1.3211,
+      "num_input_tokens_seen": 1802188180,
+      "step": 12780
+    },
+    {
+      "epoch": 0.7795037443910318,
+      "grad_norm": 0.5230554938316345,
+      "learning_rate": 8e-05,
+      "loss": 1.3797,
+      "num_input_tokens_seen": 1803600956,
+      "step": 12790
+    },
+    {
+      "epoch": 0.7801132078346525,
+      "grad_norm": 0.6262668371200562,
+      "learning_rate": 8e-05,
+      "loss": 1.285,
+      "num_input_tokens_seen": 1805012032,
+      "step": 12800
+    },
+    {
+      "epoch": 0.7807226712782734,
+      "grad_norm": 0.5023617148399353,
+      "learning_rate": 8e-05,
+      "loss": 1.3513,
+      "num_input_tokens_seen": 1806435284,
+      "step": 12810
+    },
+    {
+      "epoch": 0.7813321347218942,
+      "grad_norm": 0.7828114032745361,
+      "learning_rate": 8e-05,
+      "loss": 1.3308,
+      "num_input_tokens_seen": 1807822968,
+      "step": 12820
+    },
+    {
+      "epoch": 0.781941598165515,
+      "grad_norm": 0.4700995981693268,
+      "learning_rate": 8e-05,
+      "loss": 1.4061,
+      "num_input_tokens_seen": 1809275984,
+      "step": 12830
+    },
+    {
+      "epoch": 0.7825510616091359,
+      "grad_norm": 0.48993438482284546,
+      "learning_rate": 8e-05,
+      "loss": 1.2856,
+      "num_input_tokens_seen": 1810719656,
+      "step": 12840
+    },
+    {
+      "epoch": 0.7831605250527567,
+      "grad_norm": 0.5454090237617493,
+      "learning_rate": 8e-05,
+      "loss": 1.3035,
+      "num_input_tokens_seen": 1812148332,
+      "step": 12850
+    },
+    {
+      "epoch": 0.7837699884963775,
+      "grad_norm": 0.4457705318927765,
+      "learning_rate": 8e-05,
+      "loss": 1.3257,
+      "num_input_tokens_seen": 1813587512,
+      "step": 12860
+    },
+    {
+      "epoch": 0.7843794519399984,
+      "grad_norm": 0.5165936350822449,
+      "learning_rate": 8e-05,
+      "loss": 1.3082,
+      "num_input_tokens_seen": 1814998812,
+      "step": 12870
+    },
+    {
+      "epoch": 0.7849889153836191,
+      "grad_norm": 0.5486757755279541,
+      "learning_rate": 8e-05,
+      "loss": 1.3164,
+      "num_input_tokens_seen": 1816431448,
+      "step": 12880
+    },
+    {
+      "epoch": 0.7855983788272399,
+      "grad_norm": 0.5352789163589478,
+      "learning_rate": 8e-05,
+      "loss": 1.333,
+      "num_input_tokens_seen": 1817905740,
+      "step": 12890
+    },
+    {
+      "epoch": 0.7862078422708608,
+      "grad_norm": 0.5381478071212769,
+      "learning_rate": 8e-05,
+      "loss": 1.3422,
+      "num_input_tokens_seen": 1819312544,
+      "step": 12900
+    },
+    {
+      "epoch": 0.7868173057144816,
+      "grad_norm": 0.4825122058391571,
+      "learning_rate": 8e-05,
+      "loss": 1.3764,
+      "num_input_tokens_seen": 1820735580,
+      "step": 12910
+    },
+    {
+      "epoch": 0.7874267691581024,
+      "grad_norm": 0.49656012654304504,
+      "learning_rate": 8e-05,
+      "loss": 1.3513,
+      "num_input_tokens_seen": 1822160700,
+      "step": 12920
+    },
+    {
+      "epoch": 0.7880362326017233,
+      "grad_norm": 0.4445992112159729,
+      "learning_rate": 8e-05,
+      "loss": 1.2595,
+      "num_input_tokens_seen": 1823572968,
+      "step": 12930
+    },
+    {
+      "epoch": 0.7886456960453441,
+      "grad_norm": 0.5262532234191895,
+      "learning_rate": 8e-05,
+      "loss": 1.3655,
+      "num_input_tokens_seen": 1824957320,
+      "step": 12940
+    },
+    {
+      "epoch": 0.7892551594889649,
+      "grad_norm": 0.5006152987480164,
+      "learning_rate": 8e-05,
+      "loss": 1.3366,
+      "num_input_tokens_seen": 1826389520,
+      "step": 12950
+    },
+    {
+      "epoch": 0.7898646229325857,
+      "grad_norm": 0.5019990801811218,
+      "learning_rate": 8e-05,
+      "loss": 1.3276,
+      "num_input_tokens_seen": 1827851336,
+      "step": 12960
+    },
+    {
+      "epoch": 0.7904740863762065,
+      "grad_norm": 0.5868325233459473,
+      "learning_rate": 8e-05,
+      "loss": 1.2845,
+      "num_input_tokens_seen": 1829273684,
+      "step": 12970
+    },
+    {
+      "epoch": 0.7910835498198273,
+      "grad_norm": 0.5358790159225464,
+      "learning_rate": 8e-05,
+      "loss": 1.2664,
+      "num_input_tokens_seen": 1830681708,
+      "step": 12980
+    },
+    {
+      "epoch": 0.7916930132634482,
+      "grad_norm": 0.5080293416976929,
+      "learning_rate": 8e-05,
+      "loss": 1.3323,
+      "num_input_tokens_seen": 1832051744,
+      "step": 12990
+    },
+    {
+      "epoch": 0.792302476707069,
+      "grad_norm": 0.48632749915122986,
+      "learning_rate": 8e-05,
+      "loss": 1.3402,
+      "num_input_tokens_seen": 1833451224,
+      "step": 13000
+    },
+    {
+      "epoch": 0.7929119401506899,
+      "grad_norm": 0.5741437673568726,
+      "learning_rate": 8e-05,
+      "loss": 1.3162,
+      "num_input_tokens_seen": 1834868992,
+      "step": 13010
+    },
+    {
+      "epoch": 0.7935214035943107,
+      "grad_norm": 0.5048441886901855,
+      "learning_rate": 8e-05,
+      "loss": 1.4296,
+      "num_input_tokens_seen": 1836321756,
+      "step": 13020
+    },
+    {
+      "epoch": 0.7941308670379315,
+      "grad_norm": 0.5186116695404053,
+      "learning_rate": 8e-05,
+      "loss": 1.3066,
+      "num_input_tokens_seen": 1837704368,
+      "step": 13030
+    },
+    {
+      "epoch": 0.7947403304815523,
+      "grad_norm": 0.5407278537750244,
+      "learning_rate": 8e-05,
+      "loss": 1.3587,
+      "num_input_tokens_seen": 1839125744,
+      "step": 13040
+    },
+    {
+      "epoch": 0.7953497939251731,
+      "grad_norm": 0.5114259719848633,
+      "learning_rate": 8e-05,
+      "loss": 1.2733,
+      "num_input_tokens_seen": 1840565796,
+      "step": 13050
+    },
+    {
+      "epoch": 0.7959592573687939,
+      "grad_norm": 0.5528411269187927,
+      "learning_rate": 8e-05,
+      "loss": 1.2579,
+      "num_input_tokens_seen": 1841969076,
+      "step": 13060
+    },
+    {
+      "epoch": 0.7965687208124148,
+      "grad_norm": 0.4976850748062134,
+      "learning_rate": 8e-05,
+      "loss": 1.3238,
+      "num_input_tokens_seen": 1843405600,
+      "step": 13070
+    },
+    {
+      "epoch": 0.7971781842560356,
+      "grad_norm": 0.48887088894844055,
+      "learning_rate": 8e-05,
+      "loss": 1.322,
+      "num_input_tokens_seen": 1844856608,
+      "step": 13080
+    },
+    {
+      "epoch": 0.7977876476996564,
+      "grad_norm": 0.44426125288009644,
+      "learning_rate": 8e-05,
+      "loss": 1.2833,
+      "num_input_tokens_seen": 1846275428,
+      "step": 13090
+    },
+    {
+      "epoch": 0.7983971111432773,
+      "grad_norm": 0.5178174376487732,
+      "learning_rate": 8e-05,
+      "loss": 1.3407,
+      "num_input_tokens_seen": 1847671452,
+      "step": 13100
+    },
+    {
+      "epoch": 0.799006574586898,
+      "grad_norm": 0.445492148399353,
+      "learning_rate": 8e-05,
+      "loss": 1.3597,
+      "num_input_tokens_seen": 1849090208,
+      "step": 13110
+    },
+    {
+      "epoch": 0.7996160380305188,
+      "grad_norm": 0.547702968120575,
+      "learning_rate": 8e-05,
+      "loss": 1.3509,
+      "num_input_tokens_seen": 1850510744,
+      "step": 13120
+    },
+    {
+      "epoch": 0.8002255014741397,
+      "grad_norm": 0.531181275844574,
+      "learning_rate": 8e-05,
+      "loss": 1.2971,
+      "num_input_tokens_seen": 1851915908,
+      "step": 13130
+    },
+    {
+      "epoch": 0.8008349649177605,
+      "grad_norm": 0.5193353891372681,
+      "learning_rate": 8e-05,
+      "loss": 1.2707,
+      "num_input_tokens_seen": 1853305048,
+      "step": 13140
+    },
+    {
+      "epoch": 0.8014444283613813,
+      "grad_norm": 0.5197256207466125,
+      "learning_rate": 8e-05,
+      "loss": 1.3619,
+      "num_input_tokens_seen": 1854735100,
+      "step": 13150
+    },
+    {
+      "epoch": 0.8020538918050022,
+      "grad_norm": 0.5376043319702148,
+      "learning_rate": 8e-05,
+      "loss": 1.2955,
+      "num_input_tokens_seen": 1856147500,
+      "step": 13160
+    },
+    {
+      "epoch": 0.802663355248623,
+      "grad_norm": 0.45468467473983765,
+      "learning_rate": 8e-05,
+      "loss": 1.375,
+      "num_input_tokens_seen": 1857560596,
+      "step": 13170
+    },
+    {
+      "epoch": 0.8032728186922438,
+      "grad_norm": 0.49229443073272705,
+      "learning_rate": 8e-05,
+      "loss": 1.3398,
+      "num_input_tokens_seen": 1858960492,
+      "step": 13180
+    },
+    {
+      "epoch": 0.8038822821358647,
+      "grad_norm": 0.49927592277526855,
+      "learning_rate": 8e-05,
+      "loss": 1.2409,
+      "num_input_tokens_seen": 1860333604,
+      "step": 13190
+    },
+    {
+      "epoch": 0.8044917455794854,
+      "grad_norm": 0.533245861530304,
+      "learning_rate": 8e-05,
+      "loss": 1.4103,
+      "num_input_tokens_seen": 1861704208,
+      "step": 13200
+    },
+    {
+      "epoch": 0.8051012090231063,
+      "grad_norm": 0.4365921914577484,
+      "learning_rate": 8e-05,
+      "loss": 1.3687,
+      "num_input_tokens_seen": 1863099408,
+      "step": 13210
+    },
+    {
+      "epoch": 0.8057106724667271,
+      "grad_norm": 0.48126909136772156,
+      "learning_rate": 8e-05,
+      "loss": 1.4127,
+      "num_input_tokens_seen": 1864469736,
+      "step": 13220
+    },
+    {
+      "epoch": 0.8063201359103479,
+      "grad_norm": 0.5094852447509766,
+      "learning_rate": 8e-05,
+      "loss": 1.3602,
+      "num_input_tokens_seen": 1865862852,
+      "step": 13230
+    },
+    {
+      "epoch": 0.8069295993539688,
+      "grad_norm": 0.45473966002464294,
+      "learning_rate": 8e-05,
+      "loss": 1.3174,
+      "num_input_tokens_seen": 1867276496,
+      "step": 13240
+    },
+    {
+      "epoch": 0.8075390627975896,
+      "grad_norm": 0.465964674949646,
+      "learning_rate": 8e-05,
+      "loss": 1.3023,
+      "num_input_tokens_seen": 1868676544,
+      "step": 13250
+    },
+    {
+      "epoch": 0.8081485262412104,
+      "grad_norm": 0.4686562716960907,
+      "learning_rate": 8e-05,
+      "loss": 1.3148,
+      "num_input_tokens_seen": 1870089728,
+      "step": 13260
+    },
+    {
+      "epoch": 0.8087579896848313,
+      "grad_norm": 0.510066568851471,
+      "learning_rate": 8e-05,
+      "loss": 1.3571,
+      "num_input_tokens_seen": 1871505476,
+      "step": 13270
+    },
+    {
+      "epoch": 0.809367453128452,
+      "grad_norm": 0.5353869199752808,
+      "learning_rate": 8e-05,
+      "loss": 1.3631,
+      "num_input_tokens_seen": 1872945652,
+      "step": 13280
+    },
+    {
+      "epoch": 0.8099769165720728,
+      "grad_norm": 0.46690231561660767,
+      "learning_rate": 8e-05,
+      "loss": 1.4069,
+      "num_input_tokens_seen": 1874334148,
+      "step": 13290
+    },
+    {
+      "epoch": 0.8105863800156937,
+      "grad_norm": 0.46225330233573914,
+      "learning_rate": 8e-05,
+      "loss": 1.2693,
+      "num_input_tokens_seen": 1875742380,
+      "step": 13300
+    },
+    {
+      "epoch": 0.8111958434593145,
+      "grad_norm": 0.39511632919311523,
+      "learning_rate": 8e-05,
+      "loss": 1.4481,
+      "num_input_tokens_seen": 1877145116,
+      "step": 13310
+    },
+    {
+      "epoch": 0.8118053069029353,
+      "grad_norm": 0.53965163230896,
+      "learning_rate": 8e-05,
+      "loss": 1.3699,
+      "num_input_tokens_seen": 1878564336,
+      "step": 13320
+    },
+    {
+      "epoch": 0.8124147703465562,
+      "grad_norm": 0.47026219964027405,
+      "learning_rate": 8e-05,
+      "loss": 1.2953,
+      "num_input_tokens_seen": 1879957336,
+      "step": 13330
+    },
+    {
+      "epoch": 0.813024233790177,
+      "grad_norm": 0.45782703161239624,
+      "learning_rate": 8e-05,
+      "loss": 1.3148,
+      "num_input_tokens_seen": 1881375308,
+      "step": 13340
+    },
+    {
+      "epoch": 0.8136336972337977,
+      "grad_norm": 0.5887371301651001,
+      "learning_rate": 8e-05,
+      "loss": 1.3117,
+      "num_input_tokens_seen": 1882757276,
+      "step": 13350
+    },
+    {
+      "epoch": 0.8142431606774186,
+      "grad_norm": 0.46690821647644043,
+      "learning_rate": 8e-05,
+      "loss": 1.3411,
+      "num_input_tokens_seen": 1884181132,
+      "step": 13360
+    },
+    {
+      "epoch": 0.8148526241210394,
+      "grad_norm": 0.44570302963256836,
+      "learning_rate": 8e-05,
+      "loss": 1.3642,
+      "num_input_tokens_seen": 1885627092,
+      "step": 13370
+    },
+    {
+      "epoch": 0.8154620875646602,
+      "grad_norm": 0.46674537658691406,
+      "learning_rate": 8e-05,
+      "loss": 1.3329,
+      "num_input_tokens_seen": 1887042028,
+      "step": 13380
+    },
+    {
+      "epoch": 0.8160715510082811,
+      "grad_norm": 0.4996930956840515,
+      "learning_rate": 8e-05,
+      "loss": 1.3414,
+      "num_input_tokens_seen": 1888444844,
+      "step": 13390
+    },
+    {
+      "epoch": 0.8166810144519019,
+      "grad_norm": 0.5288018584251404,
+      "learning_rate": 8e-05,
+      "loss": 1.3824,
+      "num_input_tokens_seen": 1889802940,
+      "step": 13400
+    },
+    {
+      "epoch": 0.8172904778955228,
+      "grad_norm": 0.5384576320648193,
+      "learning_rate": 8e-05,
+      "loss": 1.2916,
+      "num_input_tokens_seen": 1891232692,
+      "step": 13410
+    },
+    {
+      "epoch": 0.8178999413391436,
+      "grad_norm": 0.5390682816505432,
+      "learning_rate": 8e-05,
+      "loss": 1.334,
+      "num_input_tokens_seen": 1892682584,
+      "step": 13420
+    },
+    {
+      "epoch": 0.8185094047827643,
+      "grad_norm": 0.4713698923587799,
+      "learning_rate": 8e-05,
+      "loss": 1.3697,
+      "num_input_tokens_seen": 1894098956,
+      "step": 13430
+    },
+    {
+      "epoch": 0.8191188682263852,
+      "grad_norm": 0.4790286421775818,
+      "learning_rate": 8e-05,
+      "loss": 1.3651,
+      "num_input_tokens_seen": 1895517364,
+      "step": 13440
+    },
+    {
+      "epoch": 0.819728331670006,
+      "grad_norm": 0.5080155730247498,
+      "learning_rate": 8e-05,
+      "loss": 1.3155,
+      "num_input_tokens_seen": 1896935040,
+      "step": 13450
+    },
+    {
+      "epoch": 0.8203377951136268,
+      "grad_norm": 0.4799495339393616,
+      "learning_rate": 8e-05,
+      "loss": 1.3478,
+      "num_input_tokens_seen": 1898327544,
+      "step": 13460
+    },
+    {
+      "epoch": 0.8209472585572477,
+      "grad_norm": 0.5344340205192566,
+      "learning_rate": 8e-05,
+      "loss": 1.3948,
+      "num_input_tokens_seen": 1899741048,
+      "step": 13470
+    },
+    {
+      "epoch": 0.8215567220008685,
+      "grad_norm": 0.5373334288597107,
+      "learning_rate": 8e-05,
+      "loss": 1.376,
+      "num_input_tokens_seen": 1901150512,
+      "step": 13480
+    },
+    {
+      "epoch": 0.8221661854444893,
+      "grad_norm": 0.511161208152771,
+      "learning_rate": 8e-05,
+      "loss": 1.333,
+      "num_input_tokens_seen": 1902523144,
+      "step": 13490
+    },
+    {
+      "epoch": 0.8227756488881102,
+      "grad_norm": 0.4798104763031006,
+      "learning_rate": 8e-05,
+      "loss": 1.4043,
+      "num_input_tokens_seen": 1903981052,
+      "step": 13500
+    },
+    {
+      "epoch": 0.823385112331731,
+      "grad_norm": 0.5256845355033875,
+      "learning_rate": 8e-05,
+      "loss": 1.2564,
+      "num_input_tokens_seen": 1905352460,
+      "step": 13510
+    },
+    {
+      "epoch": 0.8239945757753517,
+      "grad_norm": 0.43500033020973206,
+      "learning_rate": 8e-05,
+      "loss": 1.3265,
+      "num_input_tokens_seen": 1906776468,
+      "step": 13520
+    },
+    {
+      "epoch": 0.8246040392189726,
+      "grad_norm": 0.47271063923835754,
+      "learning_rate": 8e-05,
+      "loss": 1.3805,
+      "num_input_tokens_seen": 1908155040,
+      "step": 13530
+    },
+    {
+      "epoch": 0.8252135026625934,
+      "grad_norm": 0.5768705606460571,
+      "learning_rate": 8e-05,
+      "loss": 1.3352,
+      "num_input_tokens_seen": 1909572824,
+      "step": 13540
+    },
+    {
+      "epoch": 0.8258229661062142,
+      "grad_norm": 0.496417373418808,
+      "learning_rate": 8e-05,
+      "loss": 1.33,
+      "num_input_tokens_seen": 1911018176,
+      "step": 13550
+    },
+    {
+      "epoch": 0.8264324295498351,
+      "grad_norm": 0.4653494954109192,
+      "learning_rate": 8e-05,
+      "loss": 1.3169,
+      "num_input_tokens_seen": 1912427256,
+      "step": 13560
+    },
+    {
+      "epoch": 0.8270418929934559,
+      "grad_norm": 0.4730968475341797,
+      "learning_rate": 8e-05,
+      "loss": 1.3272,
+      "num_input_tokens_seen": 1913827684,
+      "step": 13570
+    },
+    {
+      "epoch": 0.8276513564370767,
+      "grad_norm": 0.4840553104877472,
+      "learning_rate": 8e-05,
+      "loss": 1.3281,
+      "num_input_tokens_seen": 1915245556,
+      "step": 13580
+    },
+    {
+      "epoch": 0.8282608198806976,
+      "grad_norm": 0.4903997480869293,
+      "learning_rate": 8e-05,
+      "loss": 1.297,
+      "num_input_tokens_seen": 1916646964,
+      "step": 13590
+    },
+    {
+      "epoch": 0.8288702833243183,
+      "grad_norm": 0.539023756980896,
+      "learning_rate": 8e-05,
+      "loss": 1.3459,
+      "num_input_tokens_seen": 1918123372,
+      "step": 13600
+    },
+    {
+      "epoch": 0.8294797467679392,
+      "grad_norm": 0.5554112195968628,
+      "learning_rate": 8e-05,
+      "loss": 1.296,
+      "num_input_tokens_seen": 1919534892,
+      "step": 13610
+    },
+    {
+      "epoch": 0.83008921021156,
+      "grad_norm": 0.4873456656932831,
+      "learning_rate": 8e-05,
+      "loss": 1.3225,
+      "num_input_tokens_seen": 1920943776,
+      "step": 13620
+    },
+    {
+      "epoch": 0.8306986736551808,
+      "grad_norm": 0.5194123387336731,
+      "learning_rate": 8e-05,
+      "loss": 1.3441,
+      "num_input_tokens_seen": 1922312644,
+      "step": 13630
+    },
+    {
+      "epoch": 0.8313081370988017,
+      "grad_norm": 0.476192444562912,
+      "learning_rate": 8e-05,
+      "loss": 1.3575,
+      "num_input_tokens_seen": 1923732804,
+      "step": 13640
+    },
+    {
+      "epoch": 0.8319176005424225,
+      "grad_norm": 0.540345311164856,
+      "learning_rate": 8e-05,
+      "loss": 1.3457,
+      "num_input_tokens_seen": 1925149948,
+      "step": 13650
+    },
+    {
+      "epoch": 0.8325270639860433,
+      "grad_norm": 0.5137238502502441,
+      "learning_rate": 8e-05,
+      "loss": 1.3691,
+      "num_input_tokens_seen": 1926573700,
+      "step": 13660
+    },
+    {
+      "epoch": 0.8331365274296642,
+      "grad_norm": 0.5790938138961792,
+      "learning_rate": 8e-05,
+      "loss": 1.3639,
+      "num_input_tokens_seen": 1927971992,
+      "step": 13670
+    },
+    {
+      "epoch": 0.8337459908732849,
+      "grad_norm": 0.509369432926178,
+      "learning_rate": 8e-05,
+      "loss": 1.2713,
+      "num_input_tokens_seen": 1929379708,
+      "step": 13680
+    },
+    {
+      "epoch": 0.8343554543169057,
+      "grad_norm": 0.523574948310852,
+      "learning_rate": 8e-05,
+      "loss": 1.3607,
+      "num_input_tokens_seen": 1930737992,
+      "step": 13690
+    },
+    {
+      "epoch": 0.8349649177605266,
+      "grad_norm": 0.5223262310028076,
+      "learning_rate": 8e-05,
+      "loss": 1.34,
+      "num_input_tokens_seen": 1932126924,
+      "step": 13700
+    },
+    {
+      "epoch": 0.8355743812041474,
+      "grad_norm": 0.5009987354278564,
+      "learning_rate": 8e-05,
+      "loss": 1.3189,
+      "num_input_tokens_seen": 1933541832,
+      "step": 13710
+    },
+    {
+      "epoch": 0.8361838446477682,
+      "grad_norm": 0.4302278459072113,
+      "learning_rate": 8e-05,
+      "loss": 1.2678,
+      "num_input_tokens_seen": 1934971524,
+      "step": 13720
+    },
+    {
+      "epoch": 0.8367933080913891,
+      "grad_norm": 0.5610336661338806,
+      "learning_rate": 8e-05,
+      "loss": 1.3011,
+      "num_input_tokens_seen": 1936364524,
+      "step": 13730
+    },
+    {
+      "epoch": 0.8374027715350099,
+      "grad_norm": 0.5906192660331726,
+      "learning_rate": 8e-05,
+      "loss": 1.3344,
+      "num_input_tokens_seen": 1937758596,
+      "step": 13740
+    },
+    {
+      "epoch": 0.8380122349786306,
+      "grad_norm": 0.5197456479072571,
+      "learning_rate": 8e-05,
+      "loss": 1.4107,
+      "num_input_tokens_seen": 1939192700,
+      "step": 13750
+    },
+    {
+      "epoch": 0.8386216984222515,
+      "grad_norm": 0.5945485830307007,
+      "learning_rate": 8e-05,
+      "loss": 1.31,
+      "num_input_tokens_seen": 1940593480,
+      "step": 13760
+    },
+    {
+      "epoch": 0.8392311618658723,
+      "grad_norm": 0.4964558184146881,
+      "learning_rate": 8e-05,
+      "loss": 1.3275,
+      "num_input_tokens_seen": 1941979224,
+      "step": 13770
+    },
+    {
+      "epoch": 0.8398406253094931,
+      "grad_norm": 0.540803074836731,
+      "learning_rate": 8e-05,
+      "loss": 1.3242,
+      "num_input_tokens_seen": 1943362060,
+      "step": 13780
+    },
+    {
+      "epoch": 0.840450088753114,
+      "grad_norm": 0.5920431017875671,
+      "learning_rate": 8e-05,
+      "loss": 1.2768,
+      "num_input_tokens_seen": 1944773336,
+      "step": 13790
+    },
+    {
+      "epoch": 0.8410595521967348,
+      "grad_norm": 0.5019993185997009,
+      "learning_rate": 8e-05,
+      "loss": 1.3502,
+      "num_input_tokens_seen": 1946151796,
+      "step": 13800
+    },
+    {
+      "epoch": 0.8416690156403556,
+      "grad_norm": 0.44807812571525574,
+      "learning_rate": 8e-05,
+      "loss": 1.2786,
+      "num_input_tokens_seen": 1947532968,
+      "step": 13810
+    },
+    {
+      "epoch": 0.8422784790839765,
+      "grad_norm": 0.4444776773452759,
+      "learning_rate": 8e-05,
+      "loss": 1.2909,
+      "num_input_tokens_seen": 1948944796,
+      "step": 13820
+    },
+    {
+      "epoch": 0.8428879425275972,
+      "grad_norm": 0.5267406702041626,
+      "learning_rate": 8e-05,
+      "loss": 1.4207,
+      "num_input_tokens_seen": 1950341252,
+      "step": 13830
+    },
+    {
+      "epoch": 0.8434974059712181,
+      "grad_norm": 0.5206403732299805,
+      "learning_rate": 8e-05,
+      "loss": 1.3525,
+      "num_input_tokens_seen": 1951752800,
+      "step": 13840
+    },
+    {
+      "epoch": 0.8441068694148389,
+      "grad_norm": 0.4875728487968445,
+      "learning_rate": 8e-05,
+      "loss": 1.2995,
+      "num_input_tokens_seen": 1953159488,
+      "step": 13850
+    },
+    {
+      "epoch": 0.8447163328584597,
+      "grad_norm": 0.49969223141670227,
+      "learning_rate": 8e-05,
+      "loss": 1.3496,
+      "num_input_tokens_seen": 1954546568,
+      "step": 13860
+    },
+    {
+      "epoch": 0.8453257963020806,
+      "grad_norm": 0.44394856691360474,
+      "learning_rate": 8e-05,
+      "loss": 1.3643,
+      "num_input_tokens_seen": 1955979156,
+      "step": 13870
+    },
+    {
+      "epoch": 0.8459352597457014,
+      "grad_norm": 0.4543069303035736,
+      "learning_rate": 8e-05,
+      "loss": 1.2678,
+      "num_input_tokens_seen": 1957369568,
+      "step": 13880
+    },
+    {
+      "epoch": 0.8465447231893222,
+      "grad_norm": 0.5392143726348877,
+      "learning_rate": 8e-05,
+      "loss": 1.2553,
+      "num_input_tokens_seen": 1958773732,
+      "step": 13890
+    },
+    {
+      "epoch": 0.8471541866329431,
+      "grad_norm": 0.48405200242996216,
+      "learning_rate": 8e-05,
+      "loss": 1.3331,
+      "num_input_tokens_seen": 1960203416,
+      "step": 13900
+    },
+    {
+      "epoch": 0.8477636500765638,
+      "grad_norm": 0.44527801871299744,
+      "learning_rate": 8e-05,
+      "loss": 1.3401,
+      "num_input_tokens_seen": 1961626884,
+      "step": 13910
+    },
+    {
+      "epoch": 0.8483731135201846,
+      "grad_norm": 0.5325603485107422,
+      "learning_rate": 8e-05,
+      "loss": 1.3624,
+      "num_input_tokens_seen": 1963044460,
+      "step": 13920
+    },
+    {
+      "epoch": 0.8489825769638055,
+      "grad_norm": 0.531408429145813,
+      "learning_rate": 8e-05,
+      "loss": 1.345,
+      "num_input_tokens_seen": 1964494136,
+      "step": 13930
+    },
+    {
+      "epoch": 0.8495920404074263,
+      "grad_norm": 0.49144402146339417,
+      "learning_rate": 8e-05,
+      "loss": 1.3735,
+      "num_input_tokens_seen": 1965888548,
+      "step": 13940
+    },
+    {
+      "epoch": 0.8502015038510471,
+      "grad_norm": 0.5098019242286682,
+      "learning_rate": 8e-05,
+      "loss": 1.3549,
+      "num_input_tokens_seen": 1967290036,
+      "step": 13950
+    },
+    {
+      "epoch": 0.850810967294668,
+      "grad_norm": 0.5794479250907898,
+      "learning_rate": 8e-05,
+      "loss": 1.3227,
+      "num_input_tokens_seen": 1968705020,
+      "step": 13960
+    },
+    {
+      "epoch": 0.8514204307382888,
+      "grad_norm": 0.4568016529083252,
+      "learning_rate": 8e-05,
+      "loss": 1.26,
+      "num_input_tokens_seen": 1970114156,
+      "step": 13970
+    },
+    {
+      "epoch": 0.8520298941819096,
+      "grad_norm": 0.5260335206985474,
+      "learning_rate": 8e-05,
+      "loss": 1.356,
+      "num_input_tokens_seen": 1971510264,
+      "step": 13980
+    },
+    {
+      "epoch": 0.8526393576255304,
+      "grad_norm": 0.45923447608947754,
+      "learning_rate": 8e-05,
+      "loss": 1.3072,
+      "num_input_tokens_seen": 1972916944,
+      "step": 13990
+    },
+    {
+      "epoch": 0.8532488210691512,
+      "grad_norm": 0.6013538837432861,
+      "learning_rate": 8e-05,
+      "loss": 1.3758,
+      "num_input_tokens_seen": 1974342624,
+      "step": 14000
+    },
+    {
+      "epoch": 0.853858284512772,
+      "grad_norm": 0.5479147434234619,
+      "learning_rate": 8e-05,
+      "loss": 1.2949,
+      "num_input_tokens_seen": 1975743100,
+      "step": 14010
+    },
+    {
+      "epoch": 0.8544677479563929,
+      "grad_norm": 0.47358494997024536,
+      "learning_rate": 8e-05,
+      "loss": 1.3132,
+      "num_input_tokens_seen": 1977148904,
+      "step": 14020
+    },
+    {
+      "epoch": 0.8550772114000137,
+      "grad_norm": 0.5430836081504822,
+      "learning_rate": 8e-05,
+      "loss": 1.3641,
+      "num_input_tokens_seen": 1978557708,
+      "step": 14030
+    },
+    {
+      "epoch": 0.8556866748436346,
+      "grad_norm": 0.5587384700775146,
+      "learning_rate": 8e-05,
+      "loss": 1.2717,
+      "num_input_tokens_seen": 1979948296,
+      "step": 14040
+    },
+    {
+      "epoch": 0.8562961382872554,
+      "grad_norm": 0.5155408978462219,
+      "learning_rate": 8e-05,
+      "loss": 1.2849,
+      "num_input_tokens_seen": 1981315776,
+      "step": 14050
+    },
+    {
+      "epoch": 0.8569056017308762,
+      "grad_norm": 0.4776671528816223,
+      "learning_rate": 8e-05,
+      "loss": 1.3477,
+      "num_input_tokens_seen": 1982710884,
+      "step": 14060
+    },
+    {
+      "epoch": 0.857515065174497,
+      "grad_norm": 0.5615227222442627,
+      "learning_rate": 8e-05,
+      "loss": 1.3669,
+      "num_input_tokens_seen": 1984128224,
+      "step": 14070
+    },
+    {
+      "epoch": 0.8581245286181178,
+      "grad_norm": 0.5377513766288757,
+      "learning_rate": 8e-05,
+      "loss": 1.2847,
+      "num_input_tokens_seen": 1985557920,
+      "step": 14080
+    },
+    {
+      "epoch": 0.8587339920617386,
+      "grad_norm": 0.4939589202404022,
+      "learning_rate": 8e-05,
+      "loss": 1.3306,
+      "num_input_tokens_seen": 1986944272,
+      "step": 14090
+    },
+    {
+      "epoch": 0.8593434555053595,
+      "grad_norm": 0.5589674115180969,
+      "learning_rate": 8e-05,
+      "loss": 1.2835,
+      "num_input_tokens_seen": 1988374288,
+      "step": 14100
+    },
+    {
+      "epoch": 0.8599529189489803,
+      "grad_norm": 0.5040762424468994,
+      "learning_rate": 8e-05,
+      "loss": 1.2794,
+      "num_input_tokens_seen": 1989791720,
+      "step": 14110
+    },
+    {
+      "epoch": 0.8605623823926011,
+      "grad_norm": 0.40734902024269104,
+      "learning_rate": 8e-05,
+      "loss": 1.2372,
+      "num_input_tokens_seen": 1991197028,
+      "step": 14120
+    },
+    {
+      "epoch": 0.861171845836222,
+      "grad_norm": 0.49047330021858215,
+      "learning_rate": 8e-05,
+      "loss": 1.3511,
+      "num_input_tokens_seen": 1992584332,
+      "step": 14130
+    },
+    {
+      "epoch": 0.8617813092798428,
+      "grad_norm": 0.5605955719947815,
+      "learning_rate": 8e-05,
+      "loss": 1.2739,
+      "num_input_tokens_seen": 1993990784,
+      "step": 14140
+    },
+    {
+      "epoch": 0.8623907727234635,
+      "grad_norm": 0.4485302269458771,
+      "learning_rate": 8e-05,
+      "loss": 1.3499,
+      "num_input_tokens_seen": 1995433372,
+      "step": 14150
+    },
+    {
+      "epoch": 0.8630002361670844,
+      "grad_norm": 0.5024114847183228,
+      "learning_rate": 8e-05,
+      "loss": 1.3728,
+      "num_input_tokens_seen": 1996822904,
+      "step": 14160
+    },
+    {
+      "epoch": 0.8636096996107052,
+      "grad_norm": 0.5688835978507996,
+      "learning_rate": 8e-05,
+      "loss": 1.3265,
+      "num_input_tokens_seen": 1998261028,
+      "step": 14170
+    },
+    {
+      "epoch": 0.864219163054326,
+      "grad_norm": 0.5042539834976196,
+      "learning_rate": 8e-05,
+      "loss": 1.3042,
+      "num_input_tokens_seen": 1999646144,
+      "step": 14180
+    },
+    {
+      "epoch": 0.8648286264979469,
+      "grad_norm": 0.4236133098602295,
+      "learning_rate": 8e-05,
+      "loss": 1.3615,
+      "num_input_tokens_seen": 2001077220,
+      "step": 14190
+    },
+    {
+      "epoch": 0.8654380899415677,
+      "grad_norm": 0.5273798108100891,
+      "learning_rate": 8e-05,
+      "loss": 1.3866,
+      "num_input_tokens_seen": 2002474020,
+      "step": 14200
+    },
+    {
+      "epoch": 0.8660475533851885,
+      "grad_norm": 0.5134631991386414,
+      "learning_rate": 8e-05,
+      "loss": 1.3165,
+      "num_input_tokens_seen": 2003842724,
+      "step": 14210
+    },
+    {
+      "epoch": 0.8666570168288094,
+      "grad_norm": 0.6235056519508362,
+      "learning_rate": 8e-05,
+      "loss": 1.3191,
+      "num_input_tokens_seen": 2005293516,
+      "step": 14220
+    },
+    {
+      "epoch": 0.8672664802724301,
+      "grad_norm": 0.6013771295547485,
+      "learning_rate": 8e-05,
+      "loss": 1.2849,
+      "num_input_tokens_seen": 2006708784,
+      "step": 14230
+    },
+    {
+      "epoch": 0.867875943716051,
+      "grad_norm": 0.5487123131752014,
+      "learning_rate": 8e-05,
+      "loss": 1.2884,
+      "num_input_tokens_seen": 2008066768,
+      "step": 14240
+    },
+    {
+      "epoch": 0.8684854071596718,
+      "grad_norm": 0.5143485069274902,
+      "learning_rate": 8e-05,
+      "loss": 1.3233,
+      "num_input_tokens_seen": 2009474988,
+      "step": 14250
+    },
+    {
+      "epoch": 0.8690948706032926,
+      "grad_norm": 0.6184853911399841,
+      "learning_rate": 8e-05,
+      "loss": 1.406,
+      "num_input_tokens_seen": 2010864984,
+      "step": 14260
+    },
+    {
+      "epoch": 0.8697043340469135,
+      "grad_norm": 0.43288683891296387,
+      "learning_rate": 8e-05,
+      "loss": 1.3188,
+      "num_input_tokens_seen": 2012273664,
+      "step": 14270
+    },
+    {
+      "epoch": 0.8703137974905343,
+      "grad_norm": 0.5103833079338074,
+      "learning_rate": 8e-05,
+      "loss": 1.3006,
+      "num_input_tokens_seen": 2013731756,
+      "step": 14280
+    },
+    {
+      "epoch": 0.8709232609341551,
+      "grad_norm": 0.46680596470832825,
+      "learning_rate": 8e-05,
+      "loss": 1.306,
+      "num_input_tokens_seen": 2015108620,
+      "step": 14290
+    },
+    {
+      "epoch": 0.871532724377776,
+      "grad_norm": 0.46874338388442993,
+      "learning_rate": 8e-05,
+      "loss": 1.2844,
+      "num_input_tokens_seen": 2016514404,
+      "step": 14300
+    },
+    {
+      "epoch": 0.8721421878213967,
+      "grad_norm": 0.4423303008079529,
+      "learning_rate": 8e-05,
+      "loss": 1.3593,
+      "num_input_tokens_seen": 2017955308,
+      "step": 14310
+    },
+    {
+      "epoch": 0.8727516512650175,
+      "grad_norm": 0.507398784160614,
+      "learning_rate": 8e-05,
+      "loss": 1.2901,
+      "num_input_tokens_seen": 2019400760,
+      "step": 14320
+    },
+    {
+      "epoch": 0.8733611147086384,
+      "grad_norm": 0.46312689781188965,
+      "learning_rate": 8e-05,
+      "loss": 1.2778,
+      "num_input_tokens_seen": 2020825292,
+      "step": 14330
+    },
+    {
+      "epoch": 0.8739705781522592,
+      "grad_norm": 0.45070067048072815,
+      "learning_rate": 8e-05,
+      "loss": 1.2942,
+      "num_input_tokens_seen": 2022226416,
+      "step": 14340
+    },
+    {
+      "epoch": 0.87458004159588,
+      "grad_norm": 0.458053857088089,
+      "learning_rate": 8e-05,
+      "loss": 1.3677,
+      "num_input_tokens_seen": 2023626112,
+      "step": 14350
+    },
+    {
+      "epoch": 0.8751895050395009,
+      "grad_norm": 0.47325587272644043,
+      "learning_rate": 8e-05,
+      "loss": 1.3597,
+      "num_input_tokens_seen": 2024990412,
+      "step": 14360
+    },
+    {
+      "epoch": 0.8757989684831217,
+      "grad_norm": 0.5431790947914124,
+      "learning_rate": 8e-05,
+      "loss": 1.3117,
+      "num_input_tokens_seen": 2026355792,
+      "step": 14370
+    },
+    {
+      "epoch": 0.8764084319267424,
+      "grad_norm": 0.45192548632621765,
+      "learning_rate": 8e-05,
+      "loss": 1.3317,
+      "num_input_tokens_seen": 2027764676,
+      "step": 14380
+    },
+    {
+      "epoch": 0.8770178953703633,
+      "grad_norm": 0.541114866733551,
+      "learning_rate": 8e-05,
+      "loss": 1.2806,
+      "num_input_tokens_seen": 2029153492,
+      "step": 14390
+    },
+    {
+      "epoch": 0.8776273588139841,
+      "grad_norm": 0.4860060513019562,
+      "learning_rate": 8e-05,
+      "loss": 1.4211,
+      "num_input_tokens_seen": 2030588204,
+      "step": 14400
+    },
+    {
+      "epoch": 0.8782368222576049,
+      "grad_norm": 0.4681771695613861,
+      "learning_rate": 8e-05,
+      "loss": 1.3697,
+      "num_input_tokens_seen": 2032011484,
+      "step": 14410
+    },
+    {
+      "epoch": 0.8788462857012258,
+      "grad_norm": 0.4607865810394287,
+      "learning_rate": 8e-05,
+      "loss": 1.3697,
+      "num_input_tokens_seen": 2033420856,
+      "step": 14420
+    },
+    {
+      "epoch": 0.8794557491448466,
+      "grad_norm": 0.6460906863212585,
+      "learning_rate": 8e-05,
+      "loss": 1.3781,
+      "num_input_tokens_seen": 2034852316,
+      "step": 14430
+    },
+    {
+      "epoch": 0.8800652125884675,
+      "grad_norm": 0.5076463222503662,
+      "learning_rate": 8e-05,
+      "loss": 1.3139,
+      "num_input_tokens_seen": 2036238376,
+      "step": 14440
+    },
+    {
+      "epoch": 0.8806746760320883,
+      "grad_norm": 0.4806708097457886,
+      "learning_rate": 8e-05,
+      "loss": 1.3171,
+      "num_input_tokens_seen": 2037656080,
+      "step": 14450
+    },
+    {
+      "epoch": 0.881284139475709,
+      "grad_norm": 0.6310375332832336,
+      "learning_rate": 8e-05,
+      "loss": 1.2628,
+      "num_input_tokens_seen": 2039101276,
+      "step": 14460
+    },
+    {
+      "epoch": 0.8818936029193299,
+      "grad_norm": 0.5229341387748718,
+      "learning_rate": 8e-05,
+      "loss": 1.368,
+      "num_input_tokens_seen": 2040493272,
+      "step": 14470
+    },
+    {
+      "epoch": 0.8825030663629507,
+      "grad_norm": 0.5170985460281372,
+      "learning_rate": 8e-05,
+      "loss": 1.3502,
+      "num_input_tokens_seen": 2041931884,
+      "step": 14480
+    },
+    {
+      "epoch": 0.8831125298065715,
+      "grad_norm": 0.5730608105659485,
+      "learning_rate": 8e-05,
+      "loss": 1.3101,
+      "num_input_tokens_seen": 2043339212,
+      "step": 14490
+    },
+    {
+      "epoch": 0.8837219932501924,
+      "grad_norm": 0.4748365879058838,
+      "learning_rate": 8e-05,
+      "loss": 1.3242,
+      "num_input_tokens_seen": 2044736972,
+      "step": 14500
+    },
+    {
+      "epoch": 0.8843314566938132,
+      "grad_norm": 0.4900893270969391,
+      "learning_rate": 8e-05,
+      "loss": 1.2727,
+      "num_input_tokens_seen": 2046111152,
+      "step": 14510
+    },
+    {
+      "epoch": 0.884940920137434,
+      "grad_norm": 0.5116603374481201,
+      "learning_rate": 8e-05,
+      "loss": 1.298,
+      "num_input_tokens_seen": 2047511888,
+      "step": 14520
+    },
+    {
+      "epoch": 0.8855503835810549,
+      "grad_norm": 0.4778316617012024,
+      "learning_rate": 8e-05,
+      "loss": 1.2757,
+      "num_input_tokens_seen": 2048942212,
+      "step": 14530
+    },
+    {
+      "epoch": 0.8861598470246757,
+      "grad_norm": 0.4536020755767822,
+      "learning_rate": 8e-05,
+      "loss": 1.3739,
+      "num_input_tokens_seen": 2050337400,
+      "step": 14540
+    },
+    {
+      "epoch": 0.8867693104682964,
+      "grad_norm": 0.4740245044231415,
+      "learning_rate": 8e-05,
+      "loss": 1.2835,
+      "num_input_tokens_seen": 2051768816,
+      "step": 14550
+    },
+    {
+      "epoch": 0.8873787739119173,
+      "grad_norm": 0.4206888675689697,
+      "learning_rate": 8e-05,
+      "loss": 1.3625,
+      "num_input_tokens_seen": 2053167920,
+      "step": 14560
+    },
+    {
+      "epoch": 0.8879882373555381,
+      "grad_norm": 0.5243297815322876,
+      "learning_rate": 8e-05,
+      "loss": 1.338,
+      "num_input_tokens_seen": 2054565312,
+      "step": 14570
+    },
+    {
+      "epoch": 0.8885977007991589,
+      "grad_norm": 0.4735357165336609,
+      "learning_rate": 8e-05,
+      "loss": 1.3339,
+      "num_input_tokens_seen": 2056014664,
+      "step": 14580
+    },
+    {
+      "epoch": 0.8892071642427798,
+      "grad_norm": 0.43609538674354553,
+      "learning_rate": 8e-05,
+      "loss": 1.318,
+      "num_input_tokens_seen": 2057450624,
+      "step": 14590
+    },
+    {
+      "epoch": 0.8898166276864006,
+      "grad_norm": 0.5434485673904419,
+      "learning_rate": 8e-05,
+      "loss": 1.2862,
+      "num_input_tokens_seen": 2058888776,
+      "step": 14600
+    },
+    {
+      "epoch": 0.8904260911300214,
+      "grad_norm": 0.516051709651947,
+      "learning_rate": 8e-05,
+      "loss": 1.2762,
+      "num_input_tokens_seen": 2060339988,
+      "step": 14610
+    },
+    {
+      "epoch": 0.8910355545736423,
+      "grad_norm": 0.4696006178855896,
+      "learning_rate": 8e-05,
+      "loss": 1.3784,
+      "num_input_tokens_seen": 2061765008,
+      "step": 14620
+    },
+    {
+      "epoch": 0.891645018017263,
+      "grad_norm": 0.5371540188789368,
+      "learning_rate": 8e-05,
+      "loss": 1.2606,
+      "num_input_tokens_seen": 2063161640,
+      "step": 14630
+    },
+    {
+      "epoch": 0.8922544814608838,
+      "grad_norm": 0.463405042886734,
+      "learning_rate": 8e-05,
+      "loss": 1.3241,
+      "num_input_tokens_seen": 2064557948,
+      "step": 14640
+    },
+    {
+      "epoch": 0.8928639449045047,
+      "grad_norm": 0.48568493127822876,
+      "learning_rate": 8e-05,
+      "loss": 1.3389,
+      "num_input_tokens_seen": 2065990172,
+      "step": 14650
+    },
+    {
+      "epoch": 0.8934734083481255,
+      "grad_norm": 0.5302020311355591,
+      "learning_rate": 8e-05,
+      "loss": 1.3282,
+      "num_input_tokens_seen": 2067415172,
+      "step": 14660
+    },
+    {
+      "epoch": 0.8940828717917464,
+      "grad_norm": 0.4722622036933899,
+      "learning_rate": 8e-05,
+      "loss": 1.3124,
+      "num_input_tokens_seen": 2068826280,
+      "step": 14670
+    },
+    {
+      "epoch": 0.8946923352353672,
+      "grad_norm": 0.5686400532722473,
+      "learning_rate": 8e-05,
+      "loss": 1.2964,
+      "num_input_tokens_seen": 2070285916,
+      "step": 14680
+    },
+    {
+      "epoch": 0.895301798678988,
+      "grad_norm": 0.5032939314842224,
+      "learning_rate": 8e-05,
+      "loss": 1.3056,
+      "num_input_tokens_seen": 2071687308,
+      "step": 14690
+    },
+    {
+      "epoch": 0.8959112621226089,
+      "grad_norm": 0.43025267124176025,
+      "learning_rate": 8e-05,
+      "loss": 1.2774,
+      "num_input_tokens_seen": 2073089308,
+      "step": 14700
+    },
+    {
+      "epoch": 0.8965207255662296,
+      "grad_norm": 0.5931614637374878,
+      "learning_rate": 8e-05,
+      "loss": 1.2295,
+      "num_input_tokens_seen": 2074487708,
+      "step": 14710
+    },
+    {
+      "epoch": 0.8971301890098504,
+      "grad_norm": 0.5101529359817505,
+      "learning_rate": 8e-05,
+      "loss": 1.3418,
+      "num_input_tokens_seen": 2075901036,
+      "step": 14720
+    },
+    {
+      "epoch": 0.8977396524534713,
+      "grad_norm": 0.46563971042633057,
+      "learning_rate": 8e-05,
+      "loss": 1.2945,
+      "num_input_tokens_seen": 2077316540,
+      "step": 14730
+    },
+    {
+      "epoch": 0.8983491158970921,
+      "grad_norm": 0.535887598991394,
+      "learning_rate": 8e-05,
+      "loss": 1.3194,
+      "num_input_tokens_seen": 2078729452,
+      "step": 14740
+    },
+    {
+      "epoch": 0.8989585793407129,
+      "grad_norm": 0.5296595096588135,
+      "learning_rate": 8e-05,
+      "loss": 1.3476,
+      "num_input_tokens_seen": 2080099128,
+      "step": 14750
+    },
+    {
+      "epoch": 0.8995680427843338,
+      "grad_norm": 0.5821929574012756,
+      "learning_rate": 8e-05,
+      "loss": 1.3395,
+      "num_input_tokens_seen": 2081541188,
+      "step": 14760
+    },
+    {
+      "epoch": 0.9001775062279546,
+      "grad_norm": 0.5599145293235779,
+      "learning_rate": 8e-05,
+      "loss": 1.4025,
+      "num_input_tokens_seen": 2082915756,
+      "step": 14770
+    },
+    {
+      "epoch": 0.9007869696715753,
+      "grad_norm": 0.548435389995575,
+      "learning_rate": 8e-05,
+      "loss": 1.2736,
+      "num_input_tokens_seen": 2084374868,
+      "step": 14780
+    },
+    {
+      "epoch": 0.9013964331151962,
+      "grad_norm": 0.5481672286987305,
+      "learning_rate": 8e-05,
+      "loss": 1.3607,
+      "num_input_tokens_seen": 2085774468,
+      "step": 14790
+    },
+    {
+      "epoch": 0.902005896558817,
+      "grad_norm": 0.5384652018547058,
+      "learning_rate": 8e-05,
+      "loss": 1.2555,
+      "num_input_tokens_seen": 2087183164,
+      "step": 14800
+    },
+    {
+      "epoch": 0.9026153600024378,
+      "grad_norm": 0.546398937702179,
+      "learning_rate": 8e-05,
+      "loss": 1.3217,
+      "num_input_tokens_seen": 2088598132,
+      "step": 14810
+    },
+    {
+      "epoch": 0.9032248234460587,
+      "grad_norm": 0.4481549561023712,
+      "learning_rate": 8e-05,
+      "loss": 1.2731,
+      "num_input_tokens_seen": 2089956332,
+      "step": 14820
+    },
+    {
+      "epoch": 0.9038342868896795,
+      "grad_norm": 0.4861372709274292,
+      "learning_rate": 8e-05,
+      "loss": 1.3456,
+      "num_input_tokens_seen": 2091342172,
+      "step": 14830
+    },
+    {
+      "epoch": 0.9044437503333003,
+      "grad_norm": 0.5424238443374634,
+      "learning_rate": 8e-05,
+      "loss": 1.3,
+      "num_input_tokens_seen": 2092764024,
+      "step": 14840
+    },
+    {
+      "epoch": 0.9050532137769212,
+      "grad_norm": 0.4749182462692261,
+      "learning_rate": 8e-05,
+      "loss": 1.2833,
+      "num_input_tokens_seen": 2094144480,
+      "step": 14850
+    },
+    {
+      "epoch": 0.905662677220542,
+      "grad_norm": 0.4868445098400116,
+      "learning_rate": 8e-05,
+      "loss": 1.3079,
+      "num_input_tokens_seen": 2095575144,
+      "step": 14860
+    },
+    {
+      "epoch": 0.9062721406641628,
+      "grad_norm": 0.48875224590301514,
+      "learning_rate": 8e-05,
+      "loss": 1.3415,
+      "num_input_tokens_seen": 2096968984,
+      "step": 14870
+    },
+    {
+      "epoch": 0.9068816041077836,
+      "grad_norm": 0.6021813750267029,
+      "learning_rate": 8e-05,
+      "loss": 1.3031,
+      "num_input_tokens_seen": 2098395432,
+      "step": 14880
+    },
+    {
+      "epoch": 0.9074910675514044,
+      "grad_norm": 0.5280013084411621,
+      "learning_rate": 8e-05,
+      "loss": 1.2774,
+      "num_input_tokens_seen": 2099787376,
+      "step": 14890
+    },
+    {
+      "epoch": 0.9081005309950253,
+      "grad_norm": 0.44756680727005005,
+      "learning_rate": 8e-05,
+      "loss": 1.415,
+      "num_input_tokens_seen": 2101178184,
+      "step": 14900
+    },
+    {
+      "epoch": 0.9087099944386461,
+      "grad_norm": 0.5173446536064148,
+      "learning_rate": 8e-05,
+      "loss": 1.2678,
+      "num_input_tokens_seen": 2102550216,
+      "step": 14910
+    },
+    {
+      "epoch": 0.9093194578822669,
+      "grad_norm": 0.5057575106620789,
+      "learning_rate": 8e-05,
+      "loss": 1.2572,
+      "num_input_tokens_seen": 2103991892,
+      "step": 14920
+    },
+    {
+      "epoch": 0.9099289213258878,
+      "grad_norm": 0.4518465995788574,
+      "learning_rate": 8e-05,
+      "loss": 1.2135,
+      "num_input_tokens_seen": 2105393196,
+      "step": 14930
+    },
+    {
+      "epoch": 0.9105383847695085,
+      "grad_norm": 0.5331540107727051,
+      "learning_rate": 8e-05,
+      "loss": 1.225,
+      "num_input_tokens_seen": 2106853740,
+      "step": 14940
+    },
+    {
+      "epoch": 0.9111478482131293,
+      "grad_norm": 0.567085862159729,
+      "learning_rate": 8e-05,
+      "loss": 1.3343,
+      "num_input_tokens_seen": 2108256916,
+      "step": 14950
+    },
+    {
+      "epoch": 0.9117573116567502,
+      "grad_norm": 0.4467783570289612,
+      "learning_rate": 8e-05,
+      "loss": 1.3246,
+      "num_input_tokens_seen": 2109676832,
+      "step": 14960
+    },
+    {
+      "epoch": 0.912366775100371,
+      "grad_norm": 0.4943903684616089,
+      "learning_rate": 8e-05,
+      "loss": 1.3044,
+      "num_input_tokens_seen": 2111041288,
+      "step": 14970
+    },
+    {
+      "epoch": 0.9129762385439918,
+      "grad_norm": 0.5459105372428894,
+      "learning_rate": 8e-05,
+      "loss": 1.3449,
+      "num_input_tokens_seen": 2112428456,
+      "step": 14980
+    },
+    {
+      "epoch": 0.9135857019876127,
+      "grad_norm": 0.6656093597412109,
+      "learning_rate": 8e-05,
+      "loss": 1.3407,
+      "num_input_tokens_seen": 2113839204,
+      "step": 14990
+    },
+    {
+      "epoch": 0.9141951654312335,
+      "grad_norm": 0.5155654549598694,
+      "learning_rate": 8e-05,
+      "loss": 1.3348,
+      "num_input_tokens_seen": 2115240548,
+      "step": 15000
+    },
+    {
+      "epoch": 0.9148046288748543,
+      "grad_norm": 0.5123623013496399,
+      "learning_rate": 8e-05,
+      "loss": 1.2716,
+      "num_input_tokens_seen": 2116649340,
+      "step": 15010
+    },
+    {
+      "epoch": 0.9154140923184751,
+      "grad_norm": 0.46583035588264465,
+      "learning_rate": 8e-05,
+      "loss": 1.2999,
+      "num_input_tokens_seen": 2118049608,
+      "step": 15020
+    },
+    {
+      "epoch": 0.9160235557620959,
+      "grad_norm": 0.5250749588012695,
+      "learning_rate": 8e-05,
+      "loss": 1.3072,
+      "num_input_tokens_seen": 2119496856,
+      "step": 15030
+    },
+    {
+      "epoch": 0.9166330192057167,
+      "grad_norm": 0.4794974625110626,
+      "learning_rate": 8e-05,
+      "loss": 1.3763,
+      "num_input_tokens_seen": 2120930156,
+      "step": 15040
+    },
+    {
+      "epoch": 0.9172424826493376,
+      "grad_norm": 0.4998897314071655,
+      "learning_rate": 8e-05,
+      "loss": 1.3722,
+      "num_input_tokens_seen": 2122320520,
+      "step": 15050
+    },
+    {
+      "epoch": 0.9178519460929584,
+      "grad_norm": 0.4267031252384186,
+      "learning_rate": 8e-05,
+      "loss": 1.3729,
+      "num_input_tokens_seen": 2123749292,
+      "step": 15060
+    },
+    {
+      "epoch": 0.9184614095365793,
+      "grad_norm": 0.5466004610061646,
+      "learning_rate": 8e-05,
+      "loss": 1.2821,
+      "num_input_tokens_seen": 2125150732,
+      "step": 15070
+    },
+    {
+      "epoch": 0.9190708729802001,
+      "grad_norm": 0.4843430519104004,
+      "learning_rate": 8e-05,
+      "loss": 1.3095,
+      "num_input_tokens_seen": 2126553680,
+      "step": 15080
+    },
+    {
+      "epoch": 0.9196803364238209,
+      "grad_norm": 0.4891015887260437,
+      "learning_rate": 8e-05,
+      "loss": 1.2743,
+      "num_input_tokens_seen": 2127954848,
+      "step": 15090
+    },
+    {
+      "epoch": 0.9202897998674417,
+      "grad_norm": 0.4860542416572571,
+      "learning_rate": 8e-05,
+      "loss": 1.2681,
+      "num_input_tokens_seen": 2129337336,
+      "step": 15100
+    },
+    {
+      "epoch": 0.9208992633110625,
+      "grad_norm": 0.5063233375549316,
+      "learning_rate": 8e-05,
+      "loss": 1.2805,
+      "num_input_tokens_seen": 2130776264,
+      "step": 15110
+    },
+    {
+      "epoch": 0.9215087267546833,
+      "grad_norm": 0.513878345489502,
+      "learning_rate": 8e-05,
+      "loss": 1.3024,
+      "num_input_tokens_seen": 2132183500,
+      "step": 15120
+    },
+    {
+      "epoch": 0.9221181901983042,
+      "grad_norm": 0.6026753783226013,
+      "learning_rate": 8e-05,
+      "loss": 1.3195,
+      "num_input_tokens_seen": 2133618296,
+      "step": 15130
+    },
+    {
+      "epoch": 0.922727653641925,
+      "grad_norm": 0.6205632090568542,
+      "learning_rate": 8e-05,
+      "loss": 1.2973,
+      "num_input_tokens_seen": 2135035140,
+      "step": 15140
+    },
+    {
+      "epoch": 0.9233371170855458,
+      "grad_norm": 0.5202974677085876,
+      "learning_rate": 8e-05,
+      "loss": 1.3003,
+      "num_input_tokens_seen": 2136445656,
+      "step": 15150
+    },
+    {
+      "epoch": 0.9239465805291667,
+      "grad_norm": 0.4942576587200165,
+      "learning_rate": 8e-05,
+      "loss": 1.3261,
+      "num_input_tokens_seen": 2137871872,
+      "step": 15160
+    },
+    {
+      "epoch": 0.9245560439727875,
+      "grad_norm": 0.4791765511035919,
+      "learning_rate": 8e-05,
+      "loss": 1.3159,
+      "num_input_tokens_seen": 2139281680,
+      "step": 15170
+    },
+    {
+      "epoch": 0.9251655074164082,
+      "grad_norm": 0.5186665654182434,
+      "learning_rate": 8e-05,
+      "loss": 1.2736,
+      "num_input_tokens_seen": 2140719180,
+      "step": 15180
+    },
+    {
+      "epoch": 0.9257749708600291,
+      "grad_norm": 0.5142242312431335,
+      "learning_rate": 8e-05,
+      "loss": 1.2888,
+      "num_input_tokens_seen": 2142122972,
+      "step": 15190
+    },
+    {
+      "epoch": 0.9263844343036499,
+      "grad_norm": 0.4749925136566162,
+      "learning_rate": 8e-05,
+      "loss": 1.3422,
+      "num_input_tokens_seen": 2143543944,
+      "step": 15200
+    },
+    {
+      "epoch": 0.9269938977472707,
+      "grad_norm": 0.4524226784706116,
+      "learning_rate": 8e-05,
+      "loss": 1.3178,
+      "num_input_tokens_seen": 2144948304,
+      "step": 15210
+    },
+    {
+      "epoch": 0.9276033611908916,
+      "grad_norm": 0.5263559222221375,
+      "learning_rate": 8e-05,
+      "loss": 1.2454,
+      "num_input_tokens_seen": 2146366520,
+      "step": 15220
+    },
+    {
+      "epoch": 0.9282128246345124,
+      "grad_norm": 0.571079432964325,
+      "learning_rate": 8e-05,
+      "loss": 1.3552,
+      "num_input_tokens_seen": 2147785372,
+      "step": 15230
+    },
+    {
+      "epoch": 0.9288222880781332,
+      "grad_norm": 0.5034895539283752,
+      "learning_rate": 8e-05,
+      "loss": 1.325,
+      "num_input_tokens_seen": 2149217276,
+      "step": 15240
+    },
+    {
+      "epoch": 0.9294317515217541,
+      "grad_norm": 0.45907917618751526,
+      "learning_rate": 8e-05,
+      "loss": 1.2877,
+      "num_input_tokens_seen": 2150613728,
+      "step": 15250
+    },
+    {
+      "epoch": 0.9300412149653748,
+      "grad_norm": 0.4312871992588043,
+      "learning_rate": 8e-05,
+      "loss": 1.3095,
+      "num_input_tokens_seen": 2151987964,
+      "step": 15260
+    },
+    {
+      "epoch": 0.9306506784089957,
+      "grad_norm": 0.5046421885490417,
+      "learning_rate": 8e-05,
+      "loss": 1.3279,
+      "num_input_tokens_seen": 2153424756,
+      "step": 15270
+    },
+    {
+      "epoch": 0.9312601418526165,
+      "grad_norm": 0.5745195150375366,
+      "learning_rate": 8e-05,
+      "loss": 1.2899,
+      "num_input_tokens_seen": 2154852656,
+      "step": 15280
+    },
+    {
+      "epoch": 0.9318696052962373,
+      "grad_norm": 0.4521108865737915,
+      "learning_rate": 8e-05,
+      "loss": 1.2494,
+      "num_input_tokens_seen": 2156242808,
+      "step": 15290
+    },
+    {
+      "epoch": 0.9324790687398582,
+      "grad_norm": 0.43681252002716064,
+      "learning_rate": 8e-05,
+      "loss": 1.2328,
+      "num_input_tokens_seen": 2157698388,
+      "step": 15300
+    },
+    {
+      "epoch": 0.933088532183479,
+      "grad_norm": 0.43972909450531006,
+      "learning_rate": 8e-05,
+      "loss": 1.3337,
+      "num_input_tokens_seen": 2159079792,
+      "step": 15310
+    },
+    {
+      "epoch": 0.9336979956270998,
+      "grad_norm": 0.4632760286331177,
+      "learning_rate": 8e-05,
+      "loss": 1.2879,
+      "num_input_tokens_seen": 2160506712,
+      "step": 15320
+    },
+    {
+      "epoch": 0.9343074590707207,
+      "grad_norm": 0.4858049154281616,
+      "learning_rate": 8e-05,
+      "loss": 1.3715,
+      "num_input_tokens_seen": 2161935332,
+      "step": 15330
+    },
+    {
+      "epoch": 0.9349169225143414,
+      "grad_norm": 0.502547025680542,
+      "learning_rate": 8e-05,
+      "loss": 1.2783,
+      "num_input_tokens_seen": 2163308116,
+      "step": 15340
+    },
+    {
+      "epoch": 0.9355263859579622,
+      "grad_norm": 0.5862429141998291,
+      "learning_rate": 8e-05,
+      "loss": 1.3081,
+      "num_input_tokens_seen": 2164680144,
+      "step": 15350
+    },
+    {
+      "epoch": 0.9361358494015831,
+      "grad_norm": 0.5465579032897949,
+      "learning_rate": 8e-05,
+      "loss": 1.3302,
+      "num_input_tokens_seen": 2166084000,
+      "step": 15360
+    },
+    {
+      "epoch": 0.9367453128452039,
+      "grad_norm": 0.4558219909667969,
+      "learning_rate": 8e-05,
+      "loss": 1.3208,
+      "num_input_tokens_seen": 2167539648,
+      "step": 15370
+    },
+    {
+      "epoch": 0.9373547762888247,
+      "grad_norm": 0.4179956614971161,
+      "learning_rate": 8e-05,
+      "loss": 1.2822,
+      "num_input_tokens_seen": 2168925220,
+      "step": 15380
+    },
+    {
+      "epoch": 0.9379642397324456,
+      "grad_norm": 0.5408328175544739,
+      "learning_rate": 8e-05,
+      "loss": 1.311,
+      "num_input_tokens_seen": 2170353320,
+      "step": 15390
+    },
+    {
+      "epoch": 0.9385737031760664,
+      "grad_norm": 0.45750945806503296,
+      "learning_rate": 8e-05,
+      "loss": 1.3182,
+      "num_input_tokens_seen": 2171767016,
+      "step": 15400
+    },
+    {
+      "epoch": 0.9391831666196871,
+      "grad_norm": 0.5293141007423401,
+      "learning_rate": 8e-05,
+      "loss": 1.3439,
+      "num_input_tokens_seen": 2173170916,
+      "step": 15410
+    },
+    {
+      "epoch": 0.939792630063308,
+      "grad_norm": 0.4048777222633362,
+      "learning_rate": 8e-05,
+      "loss": 1.2898,
+      "num_input_tokens_seen": 2174582080,
+      "step": 15420
+    },
+    {
+      "epoch": 0.9404020935069288,
+      "grad_norm": 0.5126138925552368,
+      "learning_rate": 8e-05,
+      "loss": 1.3273,
+      "num_input_tokens_seen": 2176029604,
+      "step": 15430
+    },
+    {
+      "epoch": 0.9410115569505496,
+      "grad_norm": 0.5170230865478516,
+      "learning_rate": 8e-05,
+      "loss": 1.2763,
+      "num_input_tokens_seen": 2177398412,
+      "step": 15440
+    },
+    {
+      "epoch": 0.9416210203941705,
+      "grad_norm": 0.47761601209640503,
+      "learning_rate": 8e-05,
+      "loss": 1.3317,
+      "num_input_tokens_seen": 2178777544,
+      "step": 15450
+    },
+    {
+      "epoch": 0.9422304838377913,
+      "grad_norm": 0.4851064682006836,
+      "learning_rate": 8e-05,
+      "loss": 1.2528,
+      "num_input_tokens_seen": 2180193028,
+      "step": 15460
+    },
+    {
+      "epoch": 0.9428399472814121,
+      "grad_norm": 0.445537805557251,
+      "learning_rate": 8e-05,
+      "loss": 1.3249,
+      "num_input_tokens_seen": 2181639340,
+      "step": 15470
+    },
+    {
+      "epoch": 0.943449410725033,
+      "grad_norm": 0.4982694983482361,
+      "learning_rate": 8e-05,
+      "loss": 1.257,
+      "num_input_tokens_seen": 2183099028,
+      "step": 15480
+    },
+    {
+      "epoch": 0.9440588741686537,
+      "grad_norm": 0.4679562747478485,
+      "learning_rate": 8e-05,
+      "loss": 1.2946,
+      "num_input_tokens_seen": 2184447260,
+      "step": 15490
+    },
+    {
+      "epoch": 0.9446683376122746,
+      "grad_norm": 0.48683637380599976,
+      "learning_rate": 8e-05,
+      "loss": 1.3197,
+      "num_input_tokens_seen": 2185818896,
+      "step": 15500
+    },
+    {
+      "epoch": 0.9452778010558954,
+      "grad_norm": 0.6176589131355286,
+      "learning_rate": 8e-05,
+      "loss": 1.3705,
+      "num_input_tokens_seen": 2187242616,
+      "step": 15510
+    },
+    {
+      "epoch": 0.9458872644995162,
+      "grad_norm": 0.3919903635978699,
+      "learning_rate": 8e-05,
+      "loss": 1.2758,
+      "num_input_tokens_seen": 2188635656,
+      "step": 15520
+    },
+    {
+      "epoch": 0.9464967279431371,
+      "grad_norm": 0.48860692977905273,
+      "learning_rate": 8e-05,
+      "loss": 1.2734,
+      "num_input_tokens_seen": 2190004480,
+      "step": 15530
+    },
+    {
+      "epoch": 0.9471061913867579,
+      "grad_norm": 0.547088623046875,
+      "learning_rate": 8e-05,
+      "loss": 1.2956,
+      "num_input_tokens_seen": 2191428232,
+      "step": 15540
+    },
+    {
+      "epoch": 0.9477156548303787,
+      "grad_norm": 0.46737053990364075,
+      "learning_rate": 8e-05,
+      "loss": 1.3075,
+      "num_input_tokens_seen": 2192866976,
+      "step": 15550
+    },
+    {
+      "epoch": 0.9483251182739996,
+      "grad_norm": 0.5125867128372192,
+      "learning_rate": 8e-05,
+      "loss": 1.3354,
+      "num_input_tokens_seen": 2194303344,
+      "step": 15560
+    },
+    {
+      "epoch": 0.9489345817176204,
+      "grad_norm": 0.45754584670066833,
+      "learning_rate": 8e-05,
+      "loss": 1.3479,
+      "num_input_tokens_seen": 2195684020,
+      "step": 15570
+    },
+    {
+      "epoch": 0.9495440451612411,
+      "grad_norm": 0.4521014094352722,
+      "learning_rate": 8e-05,
+      "loss": 1.3423,
+      "num_input_tokens_seen": 2197122540,
+      "step": 15580
+    },
+    {
+      "epoch": 0.950153508604862,
+      "grad_norm": 0.44286853075027466,
+      "learning_rate": 8e-05,
+      "loss": 1.3658,
+      "num_input_tokens_seen": 2198552600,
+      "step": 15590
+    },
+    {
+      "epoch": 0.9507629720484828,
+      "grad_norm": 0.489595502614975,
+      "learning_rate": 8e-05,
+      "loss": 1.3535,
+      "num_input_tokens_seen": 2200004540,
+      "step": 15600
+    },
+    {
+      "epoch": 0.9513724354921036,
+      "grad_norm": 0.5072488188743591,
+      "learning_rate": 8e-05,
+      "loss": 1.2492,
+      "num_input_tokens_seen": 2201446580,
+      "step": 15610
+    },
+    {
+      "epoch": 0.9519818989357245,
+      "grad_norm": 0.5579765439033508,
+      "learning_rate": 8e-05,
+      "loss": 1.3371,
+      "num_input_tokens_seen": 2202855424,
+      "step": 15620
+    },
+    {
+      "epoch": 0.9525913623793453,
+      "grad_norm": 0.49142131209373474,
+      "learning_rate": 8e-05,
+      "loss": 1.3132,
+      "num_input_tokens_seen": 2204239896,
+      "step": 15630
+    },
+    {
+      "epoch": 0.9532008258229661,
+      "grad_norm": 0.49964556097984314,
+      "learning_rate": 8e-05,
+      "loss": 1.3217,
+      "num_input_tokens_seen": 2205656712,
+      "step": 15640
+    },
+    {
+      "epoch": 0.953810289266587,
+      "grad_norm": 0.493123322725296,
+      "learning_rate": 8e-05,
+      "loss": 1.2907,
+      "num_input_tokens_seen": 2207060424,
+      "step": 15650
+    },
+    {
+      "epoch": 0.9544197527102077,
+      "grad_norm": 0.5087452530860901,
+      "learning_rate": 8e-05,
+      "loss": 1.3345,
+      "num_input_tokens_seen": 2208448188,
+      "step": 15660
+    },
+    {
+      "epoch": 0.9550292161538285,
+      "grad_norm": 0.5435436367988586,
+      "learning_rate": 8e-05,
+      "loss": 1.3228,
+      "num_input_tokens_seen": 2209862512,
+      "step": 15670
+    },
+    {
+      "epoch": 0.9556386795974494,
+      "grad_norm": 0.5179306268692017,
+      "learning_rate": 8e-05,
+      "loss": 1.3164,
+      "num_input_tokens_seen": 2211270724,
+      "step": 15680
+    },
+    {
+      "epoch": 0.9562481430410702,
+      "grad_norm": 0.4974176287651062,
+      "learning_rate": 8e-05,
+      "loss": 1.2702,
+      "num_input_tokens_seen": 2212669556,
+      "step": 15690
+    },
+    {
+      "epoch": 0.9568576064846911,
+      "grad_norm": 0.5859740972518921,
+      "learning_rate": 8e-05,
+      "loss": 1.2829,
+      "num_input_tokens_seen": 2214058020,
+      "step": 15700
+    },
+    {
+      "epoch": 0.9574670699283119,
+      "grad_norm": 0.5075061321258545,
+      "learning_rate": 8e-05,
+      "loss": 1.32,
+      "num_input_tokens_seen": 2215513552,
+      "step": 15710
+    },
+    {
+      "epoch": 0.9580765333719327,
+      "grad_norm": 0.5155842304229736,
+      "learning_rate": 8e-05,
+      "loss": 1.2568,
+      "num_input_tokens_seen": 2216879340,
+      "step": 15720
+    },
+    {
+      "epoch": 0.9586859968155536,
+      "grad_norm": 0.5818564295768738,
+      "learning_rate": 8e-05,
+      "loss": 1.3082,
+      "num_input_tokens_seen": 2218291808,
+      "step": 15730
+    },
+    {
+      "epoch": 0.9592954602591743,
+      "grad_norm": 0.534583568572998,
+      "learning_rate": 8e-05,
+      "loss": 1.3433,
+      "num_input_tokens_seen": 2219696636,
+      "step": 15740
+    },
+    {
+      "epoch": 0.9599049237027951,
+      "grad_norm": 0.5481460690498352,
+      "learning_rate": 8e-05,
+      "loss": 1.2681,
+      "num_input_tokens_seen": 2221124380,
+      "step": 15750
+    },
+    {
+      "epoch": 0.960514387146416,
+      "grad_norm": 0.4679185450077057,
+      "learning_rate": 8e-05,
+      "loss": 1.3098,
+      "num_input_tokens_seen": 2222547488,
+      "step": 15760
+    },
+    {
+      "epoch": 0.9611238505900368,
+      "grad_norm": 0.576784074306488,
+      "learning_rate": 8e-05,
+      "loss": 1.3268,
+      "num_input_tokens_seen": 2223963556,
+      "step": 15770
+    },
+    {
+      "epoch": 0.9617333140336576,
+      "grad_norm": 0.48329129815101624,
+      "learning_rate": 8e-05,
+      "loss": 1.2158,
+      "num_input_tokens_seen": 2225362452,
+      "step": 15780
+    },
+    {
+      "epoch": 0.9623427774772785,
+      "grad_norm": 0.5717592835426331,
+      "learning_rate": 8e-05,
+      "loss": 1.3256,
+      "num_input_tokens_seen": 2226820220,
+      "step": 15790
+    },
+    {
+      "epoch": 0.9629522409208993,
+      "grad_norm": 0.4110204875469208,
+      "learning_rate": 8e-05,
+      "loss": 1.3205,
+      "num_input_tokens_seen": 2228219256,
+      "step": 15800
+    },
+    {
+      "epoch": 0.96356170436452,
+      "grad_norm": 0.5517961382865906,
+      "learning_rate": 8e-05,
+      "loss": 1.3158,
+      "num_input_tokens_seen": 2229620876,
+      "step": 15810
+    },
+    {
+      "epoch": 0.9641711678081409,
+      "grad_norm": 0.6182602047920227,
+      "learning_rate": 8e-05,
+      "loss": 1.3214,
+      "num_input_tokens_seen": 2231002716,
+      "step": 15820
+    },
+    {
+      "epoch": 0.9647806312517617,
+      "grad_norm": 0.5628819465637207,
+      "learning_rate": 8e-05,
+      "loss": 1.3393,
+      "num_input_tokens_seen": 2232410556,
+      "step": 15830
+    },
+    {
+      "epoch": 0.9653900946953825,
+      "grad_norm": 0.4238695502281189,
+      "learning_rate": 8e-05,
+      "loss": 1.2732,
+      "num_input_tokens_seen": 2233815396,
+      "step": 15840
+    },
+    {
+      "epoch": 0.9659995581390034,
+      "grad_norm": 0.5082485675811768,
+      "learning_rate": 8e-05,
+      "loss": 1.3654,
+      "num_input_tokens_seen": 2235185328,
+      "step": 15850
+    },
+    {
+      "epoch": 0.9666090215826242,
+      "grad_norm": 0.550703227519989,
+      "learning_rate": 8e-05,
+      "loss": 1.3748,
+      "num_input_tokens_seen": 2236596760,
+      "step": 15860
+    },
+    {
+      "epoch": 0.967218485026245,
+      "grad_norm": 0.47868087887763977,
+      "learning_rate": 8e-05,
+      "loss": 1.3216,
+      "num_input_tokens_seen": 2238020480,
+      "step": 15870
+    },
+    {
+      "epoch": 0.9678279484698659,
+      "grad_norm": 0.42936983704566956,
+      "learning_rate": 8e-05,
+      "loss": 1.2932,
+      "num_input_tokens_seen": 2239422468,
+      "step": 15880
+    },
+    {
+      "epoch": 0.9684374119134866,
+      "grad_norm": 0.5375044941902161,
+      "learning_rate": 8e-05,
+      "loss": 1.2576,
+      "num_input_tokens_seen": 2240839504,
+      "step": 15890
+    },
+    {
+      "epoch": 0.9690468753571075,
+      "grad_norm": 0.44315773248672485,
+      "learning_rate": 8e-05,
+      "loss": 1.2574,
+      "num_input_tokens_seen": 2242293840,
+      "step": 15900
+    },
+    {
+      "epoch": 0.9696563388007283,
+      "grad_norm": 0.550308346748352,
+      "learning_rate": 8e-05,
+      "loss": 1.2451,
+      "num_input_tokens_seen": 2243656524,
+      "step": 15910
+    },
+    {
+      "epoch": 0.9702658022443491,
+      "grad_norm": 0.47337761521339417,
+      "learning_rate": 8e-05,
+      "loss": 1.3502,
+      "num_input_tokens_seen": 2245041628,
+      "step": 15920
+    },
+    {
+      "epoch": 0.97087526568797,
+      "grad_norm": 0.49625223875045776,
+      "learning_rate": 8e-05,
+      "loss": 1.2356,
+      "num_input_tokens_seen": 2246441132,
+      "step": 15930
+    },
+    {
+      "epoch": 0.9714847291315908,
+      "grad_norm": 0.5873345732688904,
+      "learning_rate": 8e-05,
+      "loss": 1.3175,
+      "num_input_tokens_seen": 2247890244,
+      "step": 15940
+    },
+    {
+      "epoch": 0.9720941925752116,
+      "grad_norm": 0.5385427474975586,
+      "learning_rate": 8e-05,
+      "loss": 1.3387,
+      "num_input_tokens_seen": 2249271220,
+      "step": 15950
+    },
+    {
+      "epoch": 0.9727036560188325,
+      "grad_norm": 0.6146961450576782,
+      "learning_rate": 8e-05,
+      "loss": 1.3295,
+      "num_input_tokens_seen": 2250717056,
+      "step": 15960
+    },
+    {
+      "epoch": 0.9733131194624532,
+      "grad_norm": 0.4284912347793579,
+      "learning_rate": 8e-05,
+      "loss": 1.3742,
+      "num_input_tokens_seen": 2252132980,
+      "step": 15970
+    },
+    {
+      "epoch": 0.973922582906074,
+      "grad_norm": 0.5247157216072083,
+      "learning_rate": 8e-05,
+      "loss": 1.391,
+      "num_input_tokens_seen": 2253553588,
+      "step": 15980
+    },
+    {
+      "epoch": 0.9745320463496949,
+      "grad_norm": 0.5365486741065979,
+      "learning_rate": 8e-05,
+      "loss": 1.3256,
+      "num_input_tokens_seen": 2254999908,
+      "step": 15990
+    },
+    {
+      "epoch": 0.9751415097933157,
+      "grad_norm": 0.5320517420768738,
+      "learning_rate": 8e-05,
+      "loss": 1.3291,
+      "num_input_tokens_seen": 2256440196,
+      "step": 16000
+    },
+    {
+      "epoch": 0.9757509732369365,
+      "grad_norm": 0.5396883487701416,
+      "learning_rate": 8e-05,
+      "loss": 1.2364,
+      "num_input_tokens_seen": 2257840344,
+      "step": 16010
+    },
+    {
+      "epoch": 0.9763604366805574,
+      "grad_norm": 0.5101752281188965,
+      "learning_rate": 8e-05,
+      "loss": 1.2918,
+      "num_input_tokens_seen": 2259258160,
+      "step": 16020
+    },
+    {
+      "epoch": 0.9769699001241782,
+      "grad_norm": 0.46001750230789185,
+      "learning_rate": 8e-05,
+      "loss": 1.3095,
+      "num_input_tokens_seen": 2260655960,
+      "step": 16030
+    },
+    {
+      "epoch": 0.977579363567799,
+      "grad_norm": 0.5177302360534668,
+      "learning_rate": 8e-05,
+      "loss": 1.2653,
+      "num_input_tokens_seen": 2262030540,
+      "step": 16040
+    },
+    {
+      "epoch": 0.9781888270114198,
+      "grad_norm": 0.5023877620697021,
+      "learning_rate": 8e-05,
+      "loss": 1.3001,
+      "num_input_tokens_seen": 2263437256,
+      "step": 16050
+    },
+    {
+      "epoch": 0.9787982904550406,
+      "grad_norm": 0.5150201320648193,
+      "learning_rate": 8e-05,
+      "loss": 1.2895,
+      "num_input_tokens_seen": 2264836292,
+      "step": 16060
+    },
+    {
+      "epoch": 0.9794077538986614,
+      "grad_norm": 0.5193420052528381,
+      "learning_rate": 8e-05,
+      "loss": 1.2073,
+      "num_input_tokens_seen": 2266236256,
+      "step": 16070
+    },
+    {
+      "epoch": 0.9800172173422823,
+      "grad_norm": 0.5201960802078247,
+      "learning_rate": 8e-05,
+      "loss": 1.3091,
+      "num_input_tokens_seen": 2267626184,
+      "step": 16080
+    },
+    {
+      "epoch": 0.9806266807859031,
+      "grad_norm": 0.5707736015319824,
+      "learning_rate": 8e-05,
+      "loss": 1.2807,
+      "num_input_tokens_seen": 2269065944,
+      "step": 16090
+    },
+    {
+      "epoch": 0.981236144229524,
+      "grad_norm": 0.5105067491531372,
+      "learning_rate": 8e-05,
+      "loss": 1.3169,
+      "num_input_tokens_seen": 2270432792,
+      "step": 16100
+    },
+    {
+      "epoch": 0.9818456076731448,
+      "grad_norm": 0.5395560264587402,
+      "learning_rate": 8e-05,
+      "loss": 1.3133,
+      "num_input_tokens_seen": 2271812640,
+      "step": 16110
+    },
+    {
+      "epoch": 0.9824550711167656,
+      "grad_norm": 0.43431755900382996,
+      "learning_rate": 8e-05,
+      "loss": 1.3047,
+      "num_input_tokens_seen": 2273222960,
+      "step": 16120
+    },
+    {
+      "epoch": 0.9830645345603864,
+      "grad_norm": 0.5516846179962158,
+      "learning_rate": 8e-05,
+      "loss": 1.2922,
+      "num_input_tokens_seen": 2274609328,
+      "step": 16130
+    },
+    {
+      "epoch": 0.9836739980040072,
+      "grad_norm": 0.46326059103012085,
+      "learning_rate": 8e-05,
+      "loss": 1.2601,
+      "num_input_tokens_seen": 2276053984,
+      "step": 16140
+    },
+    {
+      "epoch": 0.984283461447628,
+      "grad_norm": 0.5828720331192017,
+      "learning_rate": 8e-05,
+      "loss": 1.2963,
+      "num_input_tokens_seen": 2277490232,
+      "step": 16150
+    },
+    {
+      "epoch": 0.9848929248912489,
+      "grad_norm": 0.5485450029373169,
+      "learning_rate": 8e-05,
+      "loss": 1.2844,
+      "num_input_tokens_seen": 2278878972,
+      "step": 16160
+    },
+    {
+      "epoch": 0.9855023883348697,
+      "grad_norm": 0.4985129237174988,
+      "learning_rate": 8e-05,
+      "loss": 1.271,
+      "num_input_tokens_seen": 2280327752,
+      "step": 16170
+    },
+    {
+      "epoch": 0.9861118517784905,
+      "grad_norm": 0.47291767597198486,
+      "learning_rate": 8e-05,
+      "loss": 1.2697,
+      "num_input_tokens_seen": 2281761044,
+      "step": 16180
+    },
+    {
+      "epoch": 0.9867213152221114,
+      "grad_norm": 0.43932777643203735,
+      "learning_rate": 8e-05,
+      "loss": 1.283,
+      "num_input_tokens_seen": 2283171296,
+      "step": 16190
+    },
+    {
+      "epoch": 0.9873307786657322,
+      "grad_norm": 0.5212329030036926,
+      "learning_rate": 8e-05,
+      "loss": 1.2226,
+      "num_input_tokens_seen": 2284634396,
+      "step": 16200
+    },
+    {
+      "epoch": 0.9879402421093529,
+      "grad_norm": 0.468487024307251,
+      "learning_rate": 8e-05,
+      "loss": 1.4401,
+      "num_input_tokens_seen": 2286043496,
+      "step": 16210
+    },
+    {
+      "epoch": 0.9885497055529738,
+      "grad_norm": 0.6297067403793335,
+      "learning_rate": 8e-05,
+      "loss": 1.313,
+      "num_input_tokens_seen": 2287450120,
+      "step": 16220
+    },
+    {
+      "epoch": 0.9891591689965946,
+      "grad_norm": 0.44979387521743774,
+      "learning_rate": 8e-05,
+      "loss": 1.2554,
+      "num_input_tokens_seen": 2288841740,
+      "step": 16230
+    },
+    {
+      "epoch": 0.9897686324402154,
+      "grad_norm": 0.4660492241382599,
+      "learning_rate": 8e-05,
+      "loss": 1.2291,
+      "num_input_tokens_seen": 2290243272,
+      "step": 16240
+    },
+    {
+      "epoch": 0.9903780958838363,
+      "grad_norm": 0.4804689884185791,
+      "learning_rate": 8e-05,
+      "loss": 1.3183,
+      "num_input_tokens_seen": 2291631656,
+      "step": 16250
+    },
+    {
+      "epoch": 0.9909875593274571,
+      "grad_norm": 0.5002725124359131,
+      "learning_rate": 8e-05,
+      "loss": 1.2312,
+      "num_input_tokens_seen": 2293014940,
+      "step": 16260
+    },
+    {
+      "epoch": 0.9915970227710779,
+      "grad_norm": 0.49899348616600037,
+      "learning_rate": 8e-05,
+      "loss": 1.1467,
+      "num_input_tokens_seen": 2294401564,
+      "step": 16270
+    },
+    {
+      "epoch": 0.9922064862146988,
+      "grad_norm": 0.4643469750881195,
+      "learning_rate": 8e-05,
+      "loss": 1.3575,
+      "num_input_tokens_seen": 2295835048,
+      "step": 16280
+    },
+    {
+      "epoch": 0.9928159496583195,
+      "grad_norm": 0.5447330474853516,
+      "learning_rate": 8e-05,
+      "loss": 1.2759,
+      "num_input_tokens_seen": 2297276424,
+      "step": 16290
+    },
+    {
+      "epoch": 0.9934254131019403,
+      "grad_norm": 0.556800901889801,
+      "learning_rate": 8e-05,
+      "loss": 1.2833,
+      "num_input_tokens_seen": 2298712204,
+      "step": 16300
+    },
+    {
+      "epoch": 0.9940348765455612,
+      "grad_norm": 0.5475782752037048,
+      "learning_rate": 8e-05,
+      "loss": 1.346,
+      "num_input_tokens_seen": 2300152748,
+      "step": 16310
+    },
+    {
+      "epoch": 0.994644339989182,
+      "grad_norm": 0.47940051555633545,
+      "learning_rate": 8e-05,
+      "loss": 1.2262,
+      "num_input_tokens_seen": 2301556176,
+      "step": 16320
+    },
+    {
+      "epoch": 0.9952538034328029,
+      "grad_norm": 0.45238494873046875,
+      "learning_rate": 8e-05,
+      "loss": 1.3266,
+      "num_input_tokens_seen": 2303010552,
+      "step": 16330
+    },
+    {
+      "epoch": 0.9958632668764237,
+      "grad_norm": 0.5712950825691223,
+      "learning_rate": 8e-05,
+      "loss": 1.3051,
+      "num_input_tokens_seen": 2304415944,
+      "step": 16340
+    },
+    {
+      "epoch": 0.9964727303200445,
+      "grad_norm": 0.5340079069137573,
+      "learning_rate": 8e-05,
+      "loss": 1.2415,
+      "num_input_tokens_seen": 2305805212,
+      "step": 16350
+    },
+    {
+      "epoch": 0.9970821937636654,
+      "grad_norm": 0.5300690531730652,
+      "learning_rate": 8e-05,
+      "loss": 1.295,
+      "num_input_tokens_seen": 2307230964,
+      "step": 16360
+    },
+    {
+      "epoch": 0.9976916572072861,
+      "grad_norm": 0.5516910552978516,
+      "learning_rate": 8e-05,
+      "loss": 1.2448,
+      "num_input_tokens_seen": 2308634668,
+      "step": 16370
+    },
+    {
+      "epoch": 0.9983011206509069,
+      "grad_norm": 0.48180052638053894,
+      "learning_rate": 8e-05,
+      "loss": 1.2914,
+      "num_input_tokens_seen": 2310022092,
+      "step": 16380
+    },
+    {
+      "epoch": 0.9989105840945278,
+      "grad_norm": 0.48212724924087524,
+      "learning_rate": 8e-05,
+      "loss": 1.3808,
+      "num_input_tokens_seen": 2311466628,
+      "step": 16390
+    },
+    {
+      "epoch": 0.9995200475381486,
+      "grad_norm": 0.435754656791687,
+      "learning_rate": 8e-05,
+      "loss": 1.3024,
+      "num_input_tokens_seen": 2312932112,
+      "step": 16400
+    },
+    {
+      "epoch": 0.9999466719486831,
+      "num_input_tokens_seen": 2313940996,
+      "step": 16407,
+      "total_flos": 9.029436409798197e+18,
+      "train_loss": 0.7368571982491552,
+      "train_runtime": 130509.2663,
+      "train_samples_per_second": 32.185,
+      "train_steps_per_second": 0.126
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 16407,
+  "num_input_tokens_seen": 2313940996,
+  "num_train_epochs": 1,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 9.029436409798197e+18,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}