{
  "best_metric": 0.5924356579780579,
  "best_model_checkpoint": "ckpt/llama2_13b_other/linguistics_puzzles_no_sys/checkpoint-500",
  "epoch": 5.0,
  "eval_steps": 100,
  "global_step": 950,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.05263157894736842,
      "grad_norm": 1.5050264596939087,
      "learning_rate": 2.5e-05,
      "loss": 2.5922,
      "step": 10
    },
    {
      "epoch": 0.10526315789473684,
      "grad_norm": 1.5525988340377808,
      "learning_rate": 5e-05,
      "loss": 2.3206,
      "step": 20
    },
    {
      "epoch": 0.15789473684210525,
      "grad_norm": 1.7404705286026,
      "learning_rate": 4.998573727324295e-05,
      "loss": 1.7229,
      "step": 30
    },
    {
      "epoch": 0.21052631578947367,
      "grad_norm": 1.8962088823318481,
      "learning_rate": 4.994296536700177e-05,
      "loss": 1.3729,
      "step": 40
    },
    {
      "epoch": 0.2631578947368421,
      "grad_norm": 1.776729941368103,
      "learning_rate": 4.987173308479738e-05,
      "loss": 1.3635,
      "step": 50
    },
    {
      "epoch": 0.3157894736842105,
      "grad_norm": 11.020795822143555,
      "learning_rate": 4.977212170395598e-05,
      "loss": 1.3315,
      "step": 60
    },
    {
      "epoch": 0.3684210526315789,
      "grad_norm": 2.192176580429077,
      "learning_rate": 4.964424488287009e-05,
      "loss": 1.2515,
      "step": 70
    },
    {
      "epoch": 0.42105263157894735,
      "grad_norm": 2.4063496589660645,
      "learning_rate": 4.948824853131236e-05,
      "loss": 1.1872,
      "step": 80
    },
    {
      "epoch": 0.47368421052631576,
      "grad_norm": 2.7862613201141357,
      "learning_rate": 4.930431064394977e-05,
      "loss": 1.1552,
      "step": 90
    },
    {
      "epoch": 0.5263157894736842,
      "grad_norm": 3.5330026149749756,
      "learning_rate": 4.909264109724853e-05,
      "loss": 1.1276,
      "step": 100
    },
    {
      "epoch": 0.5263157894736842,
      "eval_loss": 1.0876480340957642,
      "eval_runtime": 1.9022,
      "eval_samples_per_second": 42.057,
      "eval_steps_per_second": 5.257,
      "step": 100
    },
    {
      "epoch": 0.5789473684210527,
      "grad_norm": 2.4774415493011475,
      "learning_rate": 4.885348141000122e-05,
      "loss": 1.1756,
      "step": 110
    },
    {
      "epoch": 0.631578947368421,
      "grad_norm": 2.380500555038452,
      "learning_rate": 4.858710446774951e-05,
      "loss": 1.1106,
      "step": 120
    },
    {
      "epoch": 0.6842105263157895,
      "grad_norm": 3.0656540393829346,
      "learning_rate": 4.829381421141671e-05,
      "loss": 1.0175,
      "step": 130
    },
    {
      "epoch": 0.7368421052631579,
      "grad_norm": 5.256251811981201,
      "learning_rate": 4.7973945290505766e-05,
      "loss": 0.9733,
      "step": 140
    },
    {
      "epoch": 0.7894736842105263,
      "grad_norm": 2.674135446548462,
      "learning_rate": 4.7627862681258037e-05,
      "loss": 0.9907,
      "step": 150
    },
    {
      "epoch": 0.8421052631578947,
      "grad_norm": 3.5206069946289062,
      "learning_rate": 4.725596127020879e-05,
      "loss": 0.9312,
      "step": 160
    },
    {
      "epoch": 0.8947368421052632,
      "grad_norm": 3.4086978435516357,
      "learning_rate": 4.685866540361456e-05,
      "loss": 0.9586,
      "step": 170
    },
    {
      "epoch": 0.9473684210526315,
      "grad_norm": 4.591642379760742,
      "learning_rate": 4.643642840326627e-05,
      "loss": 0.9595,
      "step": 180
    },
    {
      "epoch": 1.0,
      "grad_norm": 2.8823249340057373,
      "learning_rate": 4.598973204924097e-05,
      "loss": 0.8331,
      "step": 190
    },
    {
      "epoch": 1.0526315789473684,
      "grad_norm": 3.7064428329467773,
      "learning_rate": 4.551908603018191e-05,
      "loss": 0.8128,
      "step": 200
    },
    {
      "epoch": 1.0526315789473684,
      "eval_loss": 0.8153461217880249,
      "eval_runtime": 1.9192,
      "eval_samples_per_second": 41.684,
      "eval_steps_per_second": 5.21,
      "step": 200
    },
    {
      "epoch": 1.1052631578947367,
      "grad_norm": 4.2386274337768555,
      "learning_rate": 4.502502736173462e-05,
      "loss": 0.8186,
      "step": 210
    },
    {
      "epoch": 1.1578947368421053,
      "grad_norm": 3.1767256259918213,
      "learning_rate": 4.45081197738023e-05,
      "loss": 0.6895,
      "step": 220
    },
    {
      "epoch": 1.2105263157894737,
      "grad_norm": 3.748518466949463,
      "learning_rate": 4.3968953067319777e-05,
      "loss": 0.7901,
      "step": 230
    },
    {
      "epoch": 1.263157894736842,
      "grad_norm": 3.807053565979004,
      "learning_rate": 4.340814244127993e-05,
      "loss": 0.704,
      "step": 240
    },
    {
      "epoch": 1.3157894736842106,
      "grad_norm": 5.013542175292969,
      "learning_rate": 4.282632779078051e-05,
      "loss": 0.6879,
      "step": 250
    },
    {
      "epoch": 1.368421052631579,
      "grad_norm": 4.752715110778809,
      "learning_rate": 4.222417297689217e-05,
      "loss": 0.7563,
      "step": 260
    },
    {
      "epoch": 1.4210526315789473,
      "grad_norm": 3.6476950645446777,
      "learning_rate": 4.160236506918098e-05,
      "loss": 0.6846,
      "step": 270
    },
    {
      "epoch": 1.4736842105263157,
      "grad_norm": 3.8758108615875244,
      "learning_rate": 4.096161356174959e-05,
      "loss": 0.7155,
      "step": 280
    },
    {
      "epoch": 1.526315789473684,
      "grad_norm": 4.166601657867432,
      "learning_rate": 4.030264956369157e-05,
      "loss": 0.8037,
      "step": 290
    },
    {
      "epoch": 1.5789473684210527,
      "grad_norm": 4.603171348571777,
      "learning_rate": 3.962622496488269e-05,
      "loss": 0.6705,
      "step": 300
    },
    {
      "epoch": 1.5789473684210527,
      "eval_loss": 0.6891714930534363,
      "eval_runtime": 1.9174,
      "eval_samples_per_second": 41.724,
      "eval_steps_per_second": 5.216,
      "step": 300
    },
    {
      "epoch": 1.631578947368421,
      "grad_norm": 3.820142984390259,
      "learning_rate": 3.893311157806091e-05,
      "loss": 0.6389,
      "step": 310
    },
    {
      "epoch": 1.6842105263157894,
      "grad_norm": 5.900814533233643,
      "learning_rate": 3.822410025817406e-05,
      "loss": 0.7223,
      "step": 320
    },
    {
      "epoch": 1.736842105263158,
      "grad_norm": 4.315140724182129,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 0.6948,
      "step": 330
    },
    {
      "epoch": 1.7894736842105263,
      "grad_norm": 4.747324466705322,
      "learning_rate": 3.67616370150689e-05,
      "loss": 0.6658,
      "step": 340
    },
    {
      "epoch": 1.8421052631578947,
      "grad_norm": 3.504014492034912,
      "learning_rate": 3.600985378894086e-05,
      "loss": 0.643,
      "step": 350
    },
    {
      "epoch": 1.8947368421052633,
      "grad_norm": 5.181077480316162,
      "learning_rate": 3.5245508119914687e-05,
      "loss": 0.6537,
      "step": 360
    },
    {
      "epoch": 1.9473684210526314,
      "grad_norm": 5.073149681091309,
      "learning_rate": 3.44694721402644e-05,
      "loss": 0.641,
      "step": 370
    },
    {
      "epoch": 2.0,
      "grad_norm": 5.070895671844482,
      "learning_rate": 3.3682631321120504e-05,
      "loss": 0.6708,
      "step": 380
    },
    {
      "epoch": 2.0526315789473686,
      "grad_norm": 5.305852890014648,
      "learning_rate": 3.2885883462131394e-05,
      "loss": 0.5061,
      "step": 390
    },
    {
      "epoch": 2.1052631578947367,
      "grad_norm": 6.452213287353516,
      "learning_rate": 3.2080137667057595e-05,
      "loss": 0.4876,
      "step": 400
    },
    {
      "epoch": 2.1052631578947367,
      "eval_loss": 0.6224929690361023,
      "eval_runtime": 1.9167,
      "eval_samples_per_second": 41.739,
      "eval_steps_per_second": 5.217,
      "step": 400
    },
    {
      "epoch": 2.1578947368421053,
      "grad_norm": 3.6080775260925293,
      "learning_rate": 3.126631330646802e-05,
      "loss": 0.485,
      "step": 410
    },
    {
      "epoch": 2.2105263157894735,
      "grad_norm": 2.2630574703216553,
      "learning_rate": 3.0445338968721287e-05,
      "loss": 0.536,
      "step": 420
    },
    {
      "epoch": 2.263157894736842,
      "grad_norm": 4.616273880004883,
      "learning_rate": 2.961815140042974e-05,
      "loss": 0.4493,
      "step": 430
    },
    {
      "epoch": 2.3157894736842106,
      "grad_norm": 4.5297956466674805,
      "learning_rate": 2.878569443761442e-05,
      "loss": 0.4806,
      "step": 440
    },
    {
      "epoch": 2.3684210526315788,
      "grad_norm": 4.910376071929932,
      "learning_rate": 2.7948917928771158e-05,
      "loss": 0.4642,
      "step": 450
    },
    {
      "epoch": 2.4210526315789473,
      "grad_norm": 4.3276801109313965,
      "learning_rate": 2.7108776651076118e-05,
      "loss": 0.4857,
      "step": 460
    },
    {
      "epoch": 2.473684210526316,
      "grad_norm": 3.657116413116455,
      "learning_rate": 2.6266229220967818e-05,
      "loss": 0.4604,
      "step": 470
    },
    {
      "epoch": 2.526315789473684,
      "grad_norm": 4.7539896965026855,
      "learning_rate": 2.5422237000348276e-05,
      "loss": 0.4294,
      "step": 480
    },
    {
      "epoch": 2.5789473684210527,
      "grad_norm": 4.227921962738037,
      "learning_rate": 2.4577762999651726e-05,
      "loss": 0.436,
      "step": 490
    },
    {
      "epoch": 2.6315789473684212,
      "grad_norm": 6.821872234344482,
      "learning_rate": 2.3733770779032184e-05,
      "loss": 0.4435,
      "step": 500
    },
    {
      "epoch": 2.6315789473684212,
      "eval_loss": 0.5924356579780579,
      "eval_runtime": 1.9193,
      "eval_samples_per_second": 41.683,
      "eval_steps_per_second": 5.21,
      "step": 500
    },
    {
      "epoch": 2.6842105263157894,
      "grad_norm": 4.023755073547363,
      "learning_rate": 2.2891223348923884e-05,
      "loss": 0.4128,
      "step": 510
    },
    {
      "epoch": 2.736842105263158,
      "grad_norm": 4.245009899139404,
      "learning_rate": 2.2051082071228854e-05,
      "loss": 0.4201,
      "step": 520
    },
    {
      "epoch": 2.7894736842105265,
      "grad_norm": 7.485212326049805,
      "learning_rate": 2.1214305562385592e-05,
      "loss": 0.4144,
      "step": 530
    },
    {
      "epoch": 2.8421052631578947,
      "grad_norm": 3.890044689178467,
      "learning_rate": 2.0381848599570276e-05,
      "loss": 0.4325,
      "step": 540
    },
    {
      "epoch": 2.8947368421052633,
      "grad_norm": 5.785126686096191,
      "learning_rate": 1.9554661031278712e-05,
      "loss": 0.4539,
      "step": 550
    },
    {
      "epoch": 2.9473684210526314,
      "grad_norm": 3.959681272506714,
      "learning_rate": 1.8733686693531985e-05,
      "loss": 0.3898,
      "step": 560
    },
    {
      "epoch": 3.0,
      "grad_norm": 6.1470160484313965,
      "learning_rate": 1.79198623329424e-05,
      "loss": 0.4347,
      "step": 570
    },
    {
      "epoch": 3.0526315789473686,
      "grad_norm": 6.080893039703369,
      "learning_rate": 1.711411653786861e-05,
      "loss": 0.2771,
      "step": 580
    },
    {
      "epoch": 3.1052631578947367,
      "grad_norm": 3.995936155319214,
      "learning_rate": 1.6317368678879495e-05,
      "loss": 0.2786,
      "step": 590
    },
    {
      "epoch": 3.1578947368421053,
      "grad_norm": 4.9943084716796875,
      "learning_rate": 1.55305278597356e-05,
      "loss": 0.2743,
      "step": 600
    },
    {
      "epoch": 3.1578947368421053,
      "eval_loss": 0.6151354908943176,
      "eval_runtime": 1.9185,
      "eval_samples_per_second": 41.7,
      "eval_steps_per_second": 5.212,
      "step": 600
    },
    {
      "epoch": 3.2105263157894735,
      "grad_norm": 3.650193452835083,
      "learning_rate": 1.475449188008532e-05,
      "loss": 0.2611,
      "step": 610
    },
    {
      "epoch": 3.263157894736842,
      "grad_norm": 3.5425643920898438,
      "learning_rate": 1.399014621105914e-05,
      "loss": 0.237,
      "step": 620
    },
    {
      "epoch": 3.3157894736842106,
      "grad_norm": 4.187167644500732,
      "learning_rate": 1.3238362984931113e-05,
      "loss": 0.2319,
      "step": 630
    },
    {
      "epoch": 3.3684210526315788,
      "grad_norm": 3.7174108028411865,
      "learning_rate": 1.2500000000000006e-05,
      "loss": 0.2785,
      "step": 640
    },
    {
      "epoch": 3.4210526315789473,
      "grad_norm": 4.665218353271484,
      "learning_rate": 1.1775899741825947e-05,
      "loss": 0.3323,
      "step": 650
    },
    {
      "epoch": 3.473684210526316,
      "grad_norm": 6.711315631866455,
      "learning_rate": 1.1066888421939093e-05,
      "loss": 0.2762,
      "step": 660
    },
    {
      "epoch": 3.526315789473684,
      "grad_norm": 4.101406097412109,
      "learning_rate": 1.0373775035117305e-05,
      "loss": 0.2982,
      "step": 670
    },
    {
      "epoch": 3.5789473684210527,
      "grad_norm": 3.3571157455444336,
      "learning_rate": 9.697350436308427e-06,
      "loss": 0.2338,
      "step": 680
    },
    {
      "epoch": 3.6315789473684212,
      "grad_norm": 7.152629852294922,
      "learning_rate": 9.038386438250415e-06,
      "loss": 0.2962,
      "step": 690
    },
    {
      "epoch": 3.6842105263157894,
      "grad_norm": 5.147871971130371,
      "learning_rate": 8.397634930819021e-06,
      "loss": 0.2846,
      "step": 700
    },
    {
      "epoch": 3.6842105263157894,
      "eval_loss": 0.6083844900131226,
      "eval_runtime": 1.9199,
      "eval_samples_per_second": 41.67,
      "eval_steps_per_second": 5.209,
      "step": 700
    },
    {
      "epoch": 3.736842105263158,
      "grad_norm": 3.984264373779297,
      "learning_rate": 7.775827023107835e-06,
      "loss": 0.2895,
      "step": 710
    },
    {
      "epoch": 3.7894736842105265,
      "grad_norm": 6.230710983276367,
      "learning_rate": 7.173672209219495e-06,
      "loss": 0.3261,
      "step": 720
    },
    {
      "epoch": 3.8421052631578947,
      "grad_norm": 3.685063362121582,
      "learning_rate": 6.591857558720071e-06,
      "loss": 0.2358,
      "step": 730
    },
    {
      "epoch": 3.8947368421052633,
      "grad_norm": 4.337435245513916,
      "learning_rate": 6.031046932680229e-06,
      "loss": 0.2723,
      "step": 740
    },
    {
      "epoch": 3.9473684210526314,
      "grad_norm": 4.504445552825928,
      "learning_rate": 5.491880226197707e-06,
      "loss": 0.2941,
      "step": 750
    },
    {
      "epoch": 4.0,
      "grad_norm": 4.7959442138671875,
      "learning_rate": 4.9749726382653905e-06,
      "loss": 0.2721,
      "step": 760
    },
    {
      "epoch": 4.052631578947368,
      "grad_norm": 2.663322925567627,
      "learning_rate": 4.480913969818098e-06,
      "loss": 0.1677,
      "step": 770
    },
    {
      "epoch": 4.105263157894737,
      "grad_norm": 5.704188346862793,
      "learning_rate": 4.010267950759025e-06,
      "loss": 0.2291,
      "step": 780
    },
    {
      "epoch": 4.157894736842105,
      "grad_norm": 4.857370853424072,
      "learning_rate": 3.5635715967337223e-06,
      "loss": 0.1991,
      "step": 790
    },
    {
      "epoch": 4.2105263157894735,
      "grad_norm": 2.6290528774261475,
      "learning_rate": 3.141334596385448e-06,
      "loss": 0.2069,
      "step": 800
    },
    {
      "epoch": 4.2105263157894735,
      "eval_loss": 0.6427180767059326,
      "eval_runtime": 1.9195,
      "eval_samples_per_second": 41.677,
      "eval_steps_per_second": 5.21,
      "step": 800
    },
    {
      "epoch": 4.2631578947368425,
      "grad_norm": 6.7939558029174805,
      "learning_rate": 2.7440387297912123e-06,
      "loss": 0.2213,
      "step": 810
    },
    {
      "epoch": 4.315789473684211,
      "grad_norm": 5.425328731536865,
      "learning_rate": 2.372137318741968e-06,
      "loss": 0.2008,
      "step": 820
    },
    {
      "epoch": 4.368421052631579,
      "grad_norm": 3.0159809589385986,
      "learning_rate": 2.026054709494235e-06,
      "loss": 0.2178,
      "step": 830
    },
    {
      "epoch": 4.421052631578947,
      "grad_norm": 4.54276704788208,
      "learning_rate": 1.7061857885832893e-06,
      "loss": 0.1878,
      "step": 840
    },
    {
      "epoch": 4.473684210526316,
      "grad_norm": 4.1157755851745605,
      "learning_rate": 1.4128955322504966e-06,
      "loss": 0.1733,
      "step": 850
    },
    {
      "epoch": 4.526315789473684,
      "grad_norm": 4.860106945037842,
      "learning_rate": 1.1465185899987797e-06,
      "loss": 0.193,
      "step": 860
    },
    {
      "epoch": 4.578947368421053,
      "grad_norm": 4.945047378540039,
      "learning_rate": 9.073589027514789e-07,
      "loss": 0.1802,
      "step": 870
    },
    {
      "epoch": 4.631578947368421,
      "grad_norm": 2.316741943359375,
      "learning_rate": 6.956893560502359e-07,
      "loss": 0.1736,
      "step": 880
    },
    {
      "epoch": 4.684210526315789,
      "grad_norm": 4.012813091278076,
      "learning_rate": 5.117514686876379e-07,
      "loss": 0.1761,
      "step": 890
    },
    {
      "epoch": 4.7368421052631575,
      "grad_norm": 5.301681995391846,
      "learning_rate": 3.557551171299051e-07,
      "loss": 0.172,
      "step": 900
    },
    {
      "epoch": 4.7368421052631575,
      "eval_loss": 0.6494551301002502,
      "eval_runtime": 1.9201,
      "eval_samples_per_second": 41.665,
      "eval_steps_per_second": 5.208,
      "step": 900
    },
    {
      "epoch": 4.7894736842105265,
      "grad_norm": 3.559140205383301,
      "learning_rate": 2.27878296044029e-07,
      "loss": 0.1734,
      "step": 910
    },
    {
      "epoch": 4.842105263157895,
      "grad_norm": 7.743849277496338,
      "learning_rate": 1.2826691520262114e-07,
      "loss": 0.1954,
      "step": 920
    },
    {
      "epoch": 4.894736842105263,
      "grad_norm": 3.5408854484558105,
      "learning_rate": 5.7034632998231865e-08,
      "loss": 0.1744,
      "step": 930
    },
    {
      "epoch": 4.947368421052632,
      "grad_norm": 2.413121461868286,
      "learning_rate": 1.4262726757049982e-08,
      "loss": 0.1778,
      "step": 940
    },
    {
      "epoch": 5.0,
      "grad_norm": 2.56962513923645,
      "learning_rate": 0.0,
      "loss": 0.1836,
      "step": 950
    },
    {
      "epoch": 5.0,
      "step": 950,
      "total_flos": 2.0275085174217114e+17,
      "train_loss": 0.5822552880487945,
      "train_runtime": 660.0352,
      "train_samples_per_second": 11.515,
      "train_steps_per_second": 1.439
    }
  ],
  "logging_steps": 10,
  "max_steps": 950,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 100,
  "total_flos": 2.0275085174217114e+17,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}