{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.3708019191226868,
  "eval_steps": 500,
  "global_step": 2000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "entropy": 0.9619973443448544,
      "epoch": 0.006854009595613434,
      "grad_norm": 0.8562721610069275,
      "learning_rate": 0.0001995887594242632,
      "loss": 0.7973,
      "mean_token_accuracy": 0.7519877135753632,
      "num_tokens": 15771.0,
      "step": 10
    },
    {
      "entropy": 0.7034977793693542,
      "epoch": 0.013708019191226868,
      "grad_norm": 0.5451128482818604,
      "learning_rate": 0.0001991318254512223,
      "loss": 0.5986,
      "mean_token_accuracy": 0.8309322476387024,
      "num_tokens": 33062.0,
      "step": 20
    },
    {
      "entropy": 0.6603402759879827,
      "epoch": 0.0205620287868403,
      "grad_norm": 0.5171676278114319,
      "learning_rate": 0.00019867489147818142,
      "loss": 0.633,
      "mean_token_accuracy": 0.8433935061097145,
      "num_tokens": 48936.0,
      "step": 30
    },
    {
      "entropy": 0.6830728624016047,
      "epoch": 0.027416038382453736,
      "grad_norm": 0.4969835877418518,
      "learning_rate": 0.0001982179575051405,
      "loss": 0.6773,
      "mean_token_accuracy": 0.8266744241118431,
      "num_tokens": 61449.0,
      "step": 40
    },
    {
      "entropy": 0.5286078054457903,
      "epoch": 0.03427004797806717,
      "grad_norm": 0.44698312878608704,
      "learning_rate": 0.00019776102353209963,
      "loss": 0.5558,
      "mean_token_accuracy": 0.8533428102731705,
      "num_tokens": 77104.0,
      "step": 50
    },
    {
      "entropy": 0.5590948283672332,
      "epoch": 0.0411240575736806,
      "grad_norm": 0.38724300265312195,
      "learning_rate": 0.00019730408955905872,
      "loss": 0.5771,
      "mean_token_accuracy": 0.8514142513275147,
      "num_tokens": 91558.0,
      "step": 60
    },
    {
      "entropy": 0.599293502047658,
      "epoch": 0.047978067169294036,
      "grad_norm": 0.5922872424125671,
      "learning_rate": 0.00019684715558601783,
      "loss": 0.5309,
      "mean_token_accuracy": 0.851950392127037,
      "num_tokens": 105756.0,
      "step": 70
    },
    {
      "entropy": 0.6024694256484509,
      "epoch": 0.05483207676490747,
      "grad_norm": 0.5078150629997253,
      "learning_rate": 0.00019639022161297692,
      "loss": 0.6727,
      "mean_token_accuracy": 0.8480887472629547,
      "num_tokens": 120163.0,
      "step": 80
    },
    {
      "entropy": 0.5648054199293255,
      "epoch": 0.061686086360520906,
      "grad_norm": 0.29077377915382385,
      "learning_rate": 0.00019593328763993604,
      "loss": 0.5509,
      "mean_token_accuracy": 0.8548661589622497,
      "num_tokens": 138293.0,
      "step": 90
    },
    {
      "entropy": 0.5659369576722384,
      "epoch": 0.06854009595613433,
      "grad_norm": 0.3394547700881958,
      "learning_rate": 0.00019547635366689515,
      "loss": 0.5729,
      "mean_token_accuracy": 0.8460367009043693,
      "num_tokens": 157530.0,
      "step": 100
    },
    {
      "entropy": 0.6021960902959108,
      "epoch": 0.07539410555174778,
      "grad_norm": 0.42912933230400085,
      "learning_rate": 0.00019501941969385424,
      "loss": 0.5465,
      "mean_token_accuracy": 0.851969163119793,
      "num_tokens": 170607.0,
      "step": 110
    },
    {
      "entropy": 0.610968679189682,
      "epoch": 0.0822481151473612,
      "grad_norm": 0.3759806752204895,
      "learning_rate": 0.00019456248572081335,
      "loss": 0.632,
      "mean_token_accuracy": 0.83554507791996,
      "num_tokens": 185672.0,
      "step": 120
    },
    {
      "entropy": 0.5211818940937519,
      "epoch": 0.08910212474297464,
      "grad_norm": 0.503212034702301,
      "learning_rate": 0.00019410555174777244,
      "loss": 0.547,
      "mean_token_accuracy": 0.8610908895730972,
      "num_tokens": 200482.0,
      "step": 130
    },
    {
      "entropy": 0.4536220826208591,
      "epoch": 0.09595613433858807,
      "grad_norm": 0.7268697619438171,
      "learning_rate": 0.00019364861777473156,
      "loss": 0.4726,
      "mean_token_accuracy": 0.870763523876667,
      "num_tokens": 216537.0,
      "step": 140
    },
    {
      "entropy": 0.5031640276312828,
      "epoch": 0.10281014393420151,
      "grad_norm": 0.33594396710395813,
      "learning_rate": 0.00019319168380169065,
      "loss": 0.5923,
      "mean_token_accuracy": 0.8628711074590683,
      "num_tokens": 232400.0,
      "step": 150
    },
    {
      "entropy": 0.6555169004946947,
      "epoch": 0.10966415352981494,
      "grad_norm": 0.5894250869750977,
      "learning_rate": 0.00019273474982864976,
      "loss": 0.5634,
      "mean_token_accuracy": 0.838917362689972,
      "num_tokens": 244273.0,
      "step": 160
    },
    {
      "entropy": 0.6185528110712767,
      "epoch": 0.11651816312542837,
      "grad_norm": 0.5221670269966125,
      "learning_rate": 0.00019227781585560888,
      "loss": 0.6818,
      "mean_token_accuracy": 0.8385803163051605,
      "num_tokens": 262927.0,
      "step": 170
    },
    {
      "entropy": 0.43512718714773657,
      "epoch": 0.12337217272104181,
      "grad_norm": 0.4728280007839203,
      "learning_rate": 0.00019182088188256796,
      "loss": 0.5178,
      "mean_token_accuracy": 0.8747259676456451,
      "num_tokens": 271716.0,
      "step": 180
    },
    {
      "entropy": 0.6649946108460426,
      "epoch": 0.13022618231665525,
      "grad_norm": 0.47320684790611267,
      "learning_rate": 0.00019136394790952708,
      "loss": 0.6651,
      "mean_token_accuracy": 0.8223798260092735,
      "num_tokens": 287518.0,
      "step": 190
    },
    {
      "entropy": 0.44864910580217837,
      "epoch": 0.13708019191226867,
      "grad_norm": 0.4356485903263092,
      "learning_rate": 0.00019090701393648617,
      "loss": 0.4631,
      "mean_token_accuracy": 0.8835410609841347,
      "num_tokens": 306299.0,
      "step": 200
    },
    {
      "entropy": 0.5033049076795578,
      "epoch": 0.1439342015078821,
      "grad_norm": 0.48287737369537354,
      "learning_rate": 0.00019045007996344528,
      "loss": 0.5933,
      "mean_token_accuracy": 0.8669374987483025,
      "num_tokens": 321955.0,
      "step": 210
    },
    {
      "entropy": 0.43575111888349055,
      "epoch": 0.15078821110349555,
      "grad_norm": 0.5973707437515259,
      "learning_rate": 0.0001899931459904044,
      "loss": 0.4856,
      "mean_token_accuracy": 0.8797904253005981,
      "num_tokens": 332265.0,
      "step": 220
    },
    {
      "entropy": 0.6193726476281881,
      "epoch": 0.157642220699109,
      "grad_norm": 0.28756600618362427,
      "learning_rate": 0.0001895362120173635,
      "loss": 0.6056,
      "mean_token_accuracy": 0.8307039767503739,
      "num_tokens": 346377.0,
      "step": 230
    },
    {
      "entropy": 0.5753613166511059,
      "epoch": 0.1644962302947224,
      "grad_norm": 0.4320402145385742,
      "learning_rate": 0.0001890792780443226,
      "loss": 0.5834,
      "mean_token_accuracy": 0.8549783885478973,
      "num_tokens": 362964.0,
      "step": 240
    },
    {
      "entropy": 0.5963706407696009,
      "epoch": 0.17135023989033585,
      "grad_norm": 0.4648321866989136,
      "learning_rate": 0.0001886223440712817,
      "loss": 0.6272,
      "mean_token_accuracy": 0.8493530780076981,
      "num_tokens": 375717.0,
      "step": 250
    },
    {
      "entropy": 0.5467700261622668,
      "epoch": 0.1782042494859493,
      "grad_norm": 0.3487129211425781,
      "learning_rate": 0.00018816541009824083,
      "loss": 0.5449,
      "mean_token_accuracy": 0.8526464059948922,
      "num_tokens": 394586.0,
      "step": 260
    },
    {
      "entropy": 0.4246529323980212,
      "epoch": 0.1850582590815627,
      "grad_norm": 0.7286052703857422,
      "learning_rate": 0.00018770847612519992,
      "loss": 0.45,
      "mean_token_accuracy": 0.8814342901110649,
      "num_tokens": 411636.0,
      "step": 270
    },
    {
      "entropy": 0.5389048531651497,
      "epoch": 0.19191226867717615,
      "grad_norm": 0.3287123441696167,
      "learning_rate": 0.00018725154215215904,
      "loss": 0.5138,
      "mean_token_accuracy": 0.8506704963743686,
      "num_tokens": 427077.0,
      "step": 280
    },
    {
      "entropy": 0.5297756217420101,
      "epoch": 0.1987662782727896,
      "grad_norm": 0.5151430368423462,
      "learning_rate": 0.00018679460817911812,
      "loss": 0.5953,
      "mean_token_accuracy": 0.8586609676480293,
      "num_tokens": 442257.0,
      "step": 290
    },
    {
      "entropy": 0.5314730744808912,
      "epoch": 0.20562028786840303,
      "grad_norm": 0.9860548377037048,
      "learning_rate": 0.00018633767420607724,
      "loss": 0.5143,
      "mean_token_accuracy": 0.8650062039494515,
      "num_tokens": 458093.0,
      "step": 300
    },
    {
      "entropy": 0.5666845880448819,
      "epoch": 0.21247429746401644,
      "grad_norm": 0.8684160113334656,
      "learning_rate": 0.00018588074023303635,
      "loss": 0.5487,
      "mean_token_accuracy": 0.8509581357240676,
      "num_tokens": 471306.0,
      "step": 310
    },
    {
      "entropy": 0.5297997735440731,
      "epoch": 0.21932830705962988,
      "grad_norm": 0.3815328776836395,
      "learning_rate": 0.00018542380625999544,
      "loss": 0.6052,
      "mean_token_accuracy": 0.8568633005023003,
      "num_tokens": 488461.0,
      "step": 320
    },
    {
      "entropy": 0.5316725082695484,
      "epoch": 0.22618231665524333,
      "grad_norm": 0.5312303900718689,
      "learning_rate": 0.00018496687228695456,
      "loss": 0.6035,
      "mean_token_accuracy": 0.858753177523613,
      "num_tokens": 503665.0,
      "step": 330
    },
    {
      "entropy": 0.6088610142469406,
      "epoch": 0.23303632625085674,
      "grad_norm": 0.40660324692726135,
      "learning_rate": 0.00018450993831391365,
      "loss": 0.6232,
      "mean_token_accuracy": 0.8444906592369079,
      "num_tokens": 521925.0,
      "step": 340
    },
    {
      "entropy": 0.6339217025786639,
      "epoch": 0.23989033584647018,
      "grad_norm": 0.5640454888343811,
      "learning_rate": 0.00018405300434087276,
      "loss": 0.6188,
      "mean_token_accuracy": 0.8382566079497338,
      "num_tokens": 535970.0,
      "step": 350
    },
    {
      "entropy": 0.5411492632701993,
      "epoch": 0.24674434544208362,
      "grad_norm": 0.42631176114082336,
      "learning_rate": 0.00018359607036783185,
      "loss": 0.5528,
      "mean_token_accuracy": 0.8523587495088577,
      "num_tokens": 551676.0,
      "step": 360
    },
    {
      "entropy": 0.5561750333756208,
      "epoch": 0.25359835503769707,
      "grad_norm": 0.5579405426979065,
      "learning_rate": 0.00018313913639479097,
      "loss": 0.5793,
      "mean_token_accuracy": 0.8607801914215087,
      "num_tokens": 568488.0,
      "step": 370
    },
    {
      "entropy": 0.5319446712732315,
      "epoch": 0.2604523646333105,
      "grad_norm": 0.8342606425285339,
      "learning_rate": 0.00018268220242175008,
      "loss": 0.4994,
      "mean_token_accuracy": 0.863979734480381,
      "num_tokens": 582963.0,
      "step": 380
    },
    {
      "entropy": 0.5910112973302603,
      "epoch": 0.2673063742289239,
      "grad_norm": 0.5433372259140015,
      "learning_rate": 0.00018222526844870917,
      "loss": 0.6669,
      "mean_token_accuracy": 0.8433835208415985,
      "num_tokens": 598471.0,
      "step": 390
    },
    {
      "entropy": 0.46995992250740526,
      "epoch": 0.27416038382453733,
      "grad_norm": 0.26409879326820374,
      "learning_rate": 0.00018176833447566828,
      "loss": 0.5199,
      "mean_token_accuracy": 0.87328050583601,
      "num_tokens": 614036.0,
      "step": 400
    },
    {
      "entropy": 0.5400116696953774,
      "epoch": 0.2810143934201508,
      "grad_norm": 0.3498149514198303,
      "learning_rate": 0.00018131140050262737,
      "loss": 0.5902,
      "mean_token_accuracy": 0.8512750566005707,
      "num_tokens": 630937.0,
      "step": 410
    },
    {
      "entropy": 0.45603593066334724,
      "epoch": 0.2878684030157642,
      "grad_norm": 0.6973631978034973,
      "learning_rate": 0.0001808544665295865,
      "loss": 0.484,
      "mean_token_accuracy": 0.8728810593485832,
      "num_tokens": 642492.0,
      "step": 420
    },
    {
      "entropy": 0.5664497867226601,
      "epoch": 0.29472241261137766,
      "grad_norm": 0.4047413170337677,
      "learning_rate": 0.0001803975325565456,
      "loss": 0.5107,
      "mean_token_accuracy": 0.8518401965498924,
      "num_tokens": 656785.0,
      "step": 430
    },
    {
      "entropy": 0.5749023761600256,
      "epoch": 0.3015764222069911,
      "grad_norm": 0.5084949135780334,
      "learning_rate": 0.0001799405985835047,
      "loss": 0.5558,
      "mean_token_accuracy": 0.8492432355880737,
      "num_tokens": 671870.0,
      "step": 440
    },
    {
      "entropy": 0.4889295015484095,
      "epoch": 0.30843043180260454,
      "grad_norm": 0.42546579241752625,
      "learning_rate": 0.0001794836646104638,
      "loss": 0.5416,
      "mean_token_accuracy": 0.8649413183331489,
      "num_tokens": 685980.0,
      "step": 450
    },
    {
      "entropy": 0.5743775438517332,
      "epoch": 0.315284441398218,
      "grad_norm": 0.3708641231060028,
      "learning_rate": 0.0001790267306374229,
      "loss": 0.5976,
      "mean_token_accuracy": 0.8467541456222534,
      "num_tokens": 699287.0,
      "step": 460
    },
    {
      "entropy": 0.5913191799074411,
      "epoch": 0.32213845099383137,
      "grad_norm": 0.37332257628440857,
      "learning_rate": 0.000178569796664382,
      "loss": 0.5695,
      "mean_token_accuracy": 0.8441656738519668,
      "num_tokens": 714803.0,
      "step": 470
    },
    {
      "entropy": 0.45778534524142744,
      "epoch": 0.3289924605894448,
      "grad_norm": 0.5047005414962769,
      "learning_rate": 0.0001781128626913411,
      "loss": 0.4778,
      "mean_token_accuracy": 0.8752694010734559,
      "num_tokens": 732120.0,
      "step": 480
    },
    {
      "entropy": 0.5643713362514973,
      "epoch": 0.33584647018505825,
      "grad_norm": 0.4013417065143585,
      "learning_rate": 0.0001776559287183002,
      "loss": 0.5366,
      "mean_token_accuracy": 0.8520827397704125,
      "num_tokens": 745974.0,
      "step": 490
    },
    {
      "entropy": 0.4815288335084915,
      "epoch": 0.3427004797806717,
      "grad_norm": 0.3859888017177582,
      "learning_rate": 0.00017719899474525933,
      "loss": 0.5521,
      "mean_token_accuracy": 0.8687581121921539,
      "num_tokens": 759499.0,
      "step": 500
    },
    {
      "entropy": 0.5015339620411396,
      "epoch": 0.34955448937628514,
      "grad_norm": 0.6697199940681458,
      "learning_rate": 0.00017674206077221842,
      "loss": 0.4885,
      "mean_token_accuracy": 0.8649638772010804,
      "num_tokens": 773698.0,
      "step": 510
    },
    {
      "entropy": 0.6096027113497258,
      "epoch": 0.3564084989718986,
      "grad_norm": 1.021246075630188,
      "learning_rate": 0.00017628512679917753,
      "loss": 0.6121,
      "mean_token_accuracy": 0.8394175350666047,
      "num_tokens": 788221.0,
      "step": 520
    },
    {
      "entropy": 0.5186641301959753,
      "epoch": 0.363262508567512,
      "grad_norm": 0.8043237924575806,
      "learning_rate": 0.00017582819282613662,
      "loss": 0.5592,
      "mean_token_accuracy": 0.8656348437070847,
      "num_tokens": 802330.0,
      "step": 530
    },
    {
      "entropy": 0.6064855309203268,
      "epoch": 0.3701165181631254,
      "grad_norm": 0.43781760334968567,
      "learning_rate": 0.00017537125885309574,
      "loss": 0.5874,
      "mean_token_accuracy": 0.8404153436422348,
      "num_tokens": 819615.0,
      "step": 540
    },
    {
      "entropy": 0.5140635691583156,
      "epoch": 0.37697052775873885,
      "grad_norm": 0.7001516819000244,
      "learning_rate": 0.00017491432488005482,
      "loss": 0.5547,
      "mean_token_accuracy": 0.8677607625722885,
      "num_tokens": 835616.0,
      "step": 550
    },
    {
      "entropy": 0.4637599032372236,
      "epoch": 0.3838245373543523,
      "grad_norm": 0.38044923543930054,
      "learning_rate": 0.00017445739090701394,
      "loss": 0.5021,
      "mean_token_accuracy": 0.87646614164114,
      "num_tokens": 851065.0,
      "step": 560
    },
    {
      "entropy": 0.5091348428279161,
      "epoch": 0.39067854694996573,
      "grad_norm": 0.476380318403244,
      "learning_rate": 0.00017400045693397305,
      "loss": 0.7145,
      "mean_token_accuracy": 0.8637179903686046,
      "num_tokens": 866992.0,
      "step": 570
    },
    {
      "entropy": 0.5213964153081179,
      "epoch": 0.3975325565455792,
      "grad_norm": 0.3205454647541046,
      "learning_rate": 0.00017354352296093214,
      "loss": 0.5395,
      "mean_token_accuracy": 0.8607370749115943,
      "num_tokens": 881803.0,
      "step": 580
    },
    {
      "entropy": 0.619412742741406,
      "epoch": 0.4043865661411926,
      "grad_norm": 0.6068571209907532,
      "learning_rate": 0.00017308658898789126,
      "loss": 0.7504,
      "mean_token_accuracy": 0.8409741953015327,
      "num_tokens": 896283.0,
      "step": 590
    },
    {
      "entropy": 0.41873827911913397,
      "epoch": 0.41124057573680606,
      "grad_norm": 0.7212440371513367,
      "learning_rate": 0.00017262965501485035,
      "loss": 0.4416,
      "mean_token_accuracy": 0.8842875167727471,
      "num_tokens": 912255.0,
      "step": 600
    },
    {
      "entropy": 0.5398109834641218,
      "epoch": 0.41809458533241944,
      "grad_norm": 0.5380451083183289,
      "learning_rate": 0.00017217272104180946,
      "loss": 0.6092,
      "mean_token_accuracy": 0.8535096302628518,
      "num_tokens": 924648.0,
      "step": 610
    },
    {
      "entropy": 0.5522895563393831,
      "epoch": 0.4249485949280329,
      "grad_norm": 0.3073669373989105,
      "learning_rate": 0.00017171578706876858,
      "loss": 0.5943,
      "mean_token_accuracy": 0.8548912346363068,
      "num_tokens": 942485.0,
      "step": 620
    },
    {
      "entropy": 0.6571722824126482,
      "epoch": 0.4318026045236463,
      "grad_norm": 0.32408949732780457,
      "learning_rate": 0.00017125885309572766,
      "loss": 0.6485,
      "mean_token_accuracy": 0.8336619213223457,
      "num_tokens": 959239.0,
      "step": 630
    },
    {
      "entropy": 0.44382771104574203,
      "epoch": 0.43865661411925977,
      "grad_norm": 0.3714044988155365,
      "learning_rate": 0.00017080191912268678,
      "loss": 0.451,
      "mean_token_accuracy": 0.8748382180929184,
      "num_tokens": 971292.0,
      "step": 640
    },
    {
      "entropy": 0.5132732756435872,
      "epoch": 0.4455106237148732,
      "grad_norm": 0.39163199067115784,
      "learning_rate": 0.00017034498514964587,
      "loss": 0.5782,
      "mean_token_accuracy": 0.8603680655360222,
      "num_tokens": 986609.0,
      "step": 650
    },
    {
      "entropy": 0.6249308317899704,
      "epoch": 0.45236463331048665,
      "grad_norm": 0.9571526646614075,
      "learning_rate": 0.00016988805117660498,
      "loss": 0.6102,
      "mean_token_accuracy": 0.8412078201770783,
      "num_tokens": 1001886.0,
      "step": 660
    },
    {
      "entropy": 0.5879610646516085,
      "epoch": 0.4592186429061001,
      "grad_norm": 0.5797366499900818,
      "learning_rate": 0.00016943111720356407,
      "loss": 0.564,
      "mean_token_accuracy": 0.8438061460852623,
      "num_tokens": 1016026.0,
      "step": 670
    },
    {
      "entropy": 0.5628054179251194,
      "epoch": 0.4660726525017135,
      "grad_norm": 1.196199893951416,
      "learning_rate": 0.0001689741832305232,
      "loss": 0.5893,
      "mean_token_accuracy": 0.8494728982448578,
      "num_tokens": 1027357.0,
      "step": 680
    },
    {
      "entropy": 0.47301769629120827,
      "epoch": 0.4729266620973269,
      "grad_norm": 0.41828563809394836,
      "learning_rate": 0.0001685172492574823,
      "loss": 0.5013,
      "mean_token_accuracy": 0.8767685040831565,
      "num_tokens": 1042347.0,
      "step": 690
    },
    {
      "entropy": 0.5087582165375352,
      "epoch": 0.47978067169294036,
      "grad_norm": 0.36651521921157837,
      "learning_rate": 0.00016806031528444142,
      "loss": 0.5749,
      "mean_token_accuracy": 0.8544954568147659,
      "num_tokens": 1061719.0,
      "step": 700
    },
    {
      "entropy": 0.5259541615843772,
      "epoch": 0.4866346812885538,
      "grad_norm": 0.5560138821601868,
      "learning_rate": 0.00016760338131140053,
      "loss": 0.5649,
      "mean_token_accuracy": 0.860032057762146,
      "num_tokens": 1075339.0,
      "step": 710
    },
    {
      "entropy": 0.4655290380120277,
      "epoch": 0.49348869088416725,
      "grad_norm": 0.37081801891326904,
      "learning_rate": 0.00016714644733835962,
      "loss": 0.5218,
      "mean_token_accuracy": 0.8754615411162376,
      "num_tokens": 1089873.0,
      "step": 720
    },
    {
      "entropy": 0.5236123736947775,
      "epoch": 0.5003427004797807,
      "grad_norm": 0.4931930899620056,
      "learning_rate": 0.00016668951336531874,
      "loss": 0.5333,
      "mean_token_accuracy": 0.8595968760550022,
      "num_tokens": 1107924.0,
      "step": 730
    },
    {
      "entropy": 0.5846156869083643,
      "epoch": 0.5071967100753941,
      "grad_norm": 0.47382351756095886,
      "learning_rate": 0.00016623257939227782,
      "loss": 0.6035,
      "mean_token_accuracy": 0.8439710319042206,
      "num_tokens": 1123755.0,
      "step": 740
    },
    {
      "entropy": 0.5500952435657382,
      "epoch": 0.5140507196710076,
      "grad_norm": 0.5050795674324036,
      "learning_rate": 0.00016577564541923694,
      "loss": 0.6073,
      "mean_token_accuracy": 0.8468502178788185,
      "num_tokens": 1138565.0,
      "step": 750
    },
    {
      "entropy": 0.5514631005004048,
      "epoch": 0.520904729266621,
      "grad_norm": 0.5656992793083191,
      "learning_rate": 0.00016531871144619603,
      "loss": 0.5075,
      "mean_token_accuracy": 0.8544194102287292,
      "num_tokens": 1157676.0,
      "step": 760
    },
    {
      "entropy": 0.49982974790036677,
      "epoch": 0.5277587388622345,
      "grad_norm": 0.4955768585205078,
      "learning_rate": 0.00016486177747315514,
      "loss": 0.5041,
      "mean_token_accuracy": 0.8634250342845917,
      "num_tokens": 1174042.0,
      "step": 770
    },
    {
      "entropy": 0.5862449683248997,
      "epoch": 0.5346127484578478,
      "grad_norm": 0.5951958298683167,
      "learning_rate": 0.00016440484350011426,
      "loss": 0.5804,
      "mean_token_accuracy": 0.8448848068714142,
      "num_tokens": 1190184.0,
      "step": 780
    },
    {
      "entropy": 0.4969827888533473,
      "epoch": 0.5414667580534612,
      "grad_norm": 0.4006407856941223,
      "learning_rate": 0.00016394790952707335,
      "loss": 0.5379,
      "mean_token_accuracy": 0.8700526058673859,
      "num_tokens": 1205954.0,
      "step": 790
    },
    {
      "entropy": 0.5849012348800897,
      "epoch": 0.5483207676490747,
      "grad_norm": 0.624742329120636,
      "learning_rate": 0.00016349097555403246,
      "loss": 0.6361,
      "mean_token_accuracy": 0.8400089010596276,
      "num_tokens": 1220942.0,
      "step": 800
    },
    {
      "entropy": 0.5625248458236456,
      "epoch": 0.5551747772446881,
      "grad_norm": 1.40684974193573,
      "learning_rate": 0.00016303404158099155,
      "loss": 0.6314,
      "mean_token_accuracy": 0.8519342541694641,
      "num_tokens": 1236644.0,
      "step": 810
    },
    {
      "entropy": 0.49897886253893375,
      "epoch": 0.5620287868403016,
      "grad_norm": 0.5541409850120544,
      "learning_rate": 0.00016257710760795067,
      "loss": 0.5443,
      "mean_token_accuracy": 0.8650717407464981,
      "num_tokens": 1252654.0,
      "step": 820
    },
    {
      "entropy": 0.4627639502286911,
      "epoch": 0.568882796435915,
      "grad_norm": 0.48005378246307373,
      "learning_rate": 0.00016212017363490975,
      "loss": 0.5117,
      "mean_token_accuracy": 0.8780093342065811,
      "num_tokens": 1267484.0,
      "step": 830
    },
    {
      "entropy": 0.5190326888114214,
      "epoch": 0.5757368060315284,
      "grad_norm": 0.6706176400184631,
      "learning_rate": 0.00016166323966186887,
      "loss": 0.5251,
      "mean_token_accuracy": 0.8629909038543702,
      "num_tokens": 1282022.0,
      "step": 840
    },
    {
      "entropy": 0.5338688423857093,
      "epoch": 0.5825908156271419,
      "grad_norm": 0.5225579738616943,
      "learning_rate": 0.00016120630568882798,
      "loss": 0.5929,
      "mean_token_accuracy": 0.856892392039299,
      "num_tokens": 1296087.0,
      "step": 850
    },
    {
      "entropy": 0.49688454922288655,
      "epoch": 0.5894448252227553,
      "grad_norm": 0.26188451051712036,
      "learning_rate": 0.00016074937171578707,
      "loss": 0.6414,
      "mean_token_accuracy": 0.8633167922496796,
      "num_tokens": 1313509.0,
      "step": 860
    },
    {
      "entropy": 0.5256685543805361,
      "epoch": 0.5962988348183688,
      "grad_norm": 0.6064833998680115,
      "learning_rate": 0.0001602924377427462,
      "loss": 0.4999,
      "mean_token_accuracy": 0.8630939826369286,
      "num_tokens": 1328689.0,
      "step": 870
    },
    {
      "entropy": 0.5231131825596094,
      "epoch": 0.6031528444139822,
      "grad_norm": 0.4005596339702606,
      "learning_rate": 0.00015983550376970528,
      "loss": 0.5431,
      "mean_token_accuracy": 0.8552977308630944,
      "num_tokens": 1345534.0,
      "step": 880
    },
    {
      "entropy": 0.5158163897693158,
      "epoch": 0.6100068540095956,
      "grad_norm": 0.8037022948265076,
      "learning_rate": 0.0001593785697966644,
      "loss": 0.6514,
      "mean_token_accuracy": 0.8581083044409752,
      "num_tokens": 1361497.0,
      "step": 890
    },
    {
      "entropy": 0.4985586106777191,
      "epoch": 0.6168608636052091,
      "grad_norm": 0.38318124413490295,
      "learning_rate": 0.0001589216358236235,
      "loss": 0.5226,
      "mean_token_accuracy": 0.8698930114507675,
      "num_tokens": 1376582.0,
      "step": 900
    },
    {
      "entropy": 0.505715012550354,
      "epoch": 0.6237148732008225,
      "grad_norm": 0.39991825819015503,
      "learning_rate": 0.0001584647018505826,
      "loss": 0.5799,
      "mean_token_accuracy": 0.8601746618747711,
      "num_tokens": 1393099.0,
      "step": 910
    },
    {
      "entropy": 0.5512417580932378,
      "epoch": 0.630568882796436,
      "grad_norm": 0.3564458191394806,
      "learning_rate": 0.0001580077678775417,
      "loss": 0.6042,
      "mean_token_accuracy": 0.8542352899909019,
      "num_tokens": 1405413.0,
      "step": 920
    },
    {
      "entropy": 0.6023517435416579,
      "epoch": 0.6374228923920493,
      "grad_norm": 0.5077884197235107,
      "learning_rate": 0.0001575508339045008,
      "loss": 0.5491,
      "mean_token_accuracy": 0.8345923721790314,
      "num_tokens": 1420630.0,
      "step": 930
    },
    {
      "entropy": 0.45138914659619334,
      "epoch": 0.6442769019876627,
      "grad_norm": 0.526972234249115,
      "learning_rate": 0.00015709389993145991,
      "loss": 0.4937,
      "mean_token_accuracy": 0.8770761311054229,
      "num_tokens": 1434843.0,
      "step": 940
    },
    {
      "entropy": 0.5025592448189855,
      "epoch": 0.6511309115832762,
      "grad_norm": 0.33170071244239807,
      "learning_rate": 0.000156636965958419,
      "loss": 0.6154,
      "mean_token_accuracy": 0.8634001970291137,
      "num_tokens": 1451288.0,
      "step": 950
    },
    {
      "entropy": 0.4892792083323002,
      "epoch": 0.6579849211788896,
      "grad_norm": 0.5489900708198547,
      "learning_rate": 0.00015618003198537812,
      "loss": 0.4766,
      "mean_token_accuracy": 0.8682720705866813,
      "num_tokens": 1461993.0,
      "step": 960
    },
    {
      "entropy": 0.45295149497687814,
      "epoch": 0.6648389307745031,
      "grad_norm": 0.4360350966453552,
      "learning_rate": 0.00015572309801233723,
      "loss": 0.4526,
      "mean_token_accuracy": 0.8701859056949616,
      "num_tokens": 1476768.0,
      "step": 970
    },
    {
      "entropy": 0.5168599784374237,
      "epoch": 0.6716929403701165,
      "grad_norm": 0.6106163263320923,
      "learning_rate": 0.00015526616403929632,
      "loss": 0.5304,
      "mean_token_accuracy": 0.8633713632822037,
      "num_tokens": 1490815.0,
      "step": 980
    },
    {
      "entropy": 0.5374733801931143,
      "epoch": 0.67854694996573,
      "grad_norm": 0.6627058982849121,
      "learning_rate": 0.00015480923006625544,
      "loss": 0.6159,
      "mean_token_accuracy": 0.8490797847509384,
      "num_tokens": 1509112.0,
      "step": 990
    },
    {
      "entropy": 0.4429541861638427,
      "epoch": 0.6854009595613434,
      "grad_norm": 0.4334024488925934,
      "learning_rate": 0.00015435229609321452,
      "loss": 0.4791,
      "mean_token_accuracy": 0.8743316605687141,
      "num_tokens": 1525280.0,
      "step": 1000
    },
    {
      "entropy": 0.5753512477502227,
      "epoch": 0.6922549691569568,
      "grad_norm": 0.4165889322757721,
      "learning_rate": 0.00015389536212017364,
      "loss": 0.6037,
      "mean_token_accuracy": 0.8437389314174653,
      "num_tokens": 1541391.0,
      "step": 1010
    },
    {
      "entropy": 0.4632994903251529,
      "epoch": 0.6991089787525703,
      "grad_norm": 0.562599241733551,
      "learning_rate": 0.00015343842814713273,
      "loss": 0.4878,
      "mean_token_accuracy": 0.8699111506342888,
      "num_tokens": 1558296.0,
      "step": 1020
    },
    {
      "entropy": 0.5592806939035654,
      "epoch": 0.7059629883481837,
      "grad_norm": 0.5945786833763123,
      "learning_rate": 0.00015298149417409184,
      "loss": 0.5674,
      "mean_token_accuracy": 0.8522913366556167,
      "num_tokens": 1572653.0,
      "step": 1030
    },
    {
      "entropy": 0.4662055611610413,
      "epoch": 0.7128169979437972,
      "grad_norm": 0.5158401131629944,
      "learning_rate": 0.00015252456020105096,
      "loss": 0.5197,
      "mean_token_accuracy": 0.8687326833605766,
      "num_tokens": 1586887.0,
      "step": 1040
    },
    {
      "entropy": 0.5457549329847098,
      "epoch": 0.7196710075394106,
      "grad_norm": 0.5199631452560425,
      "learning_rate": 0.00015206762622801005,
      "loss": 0.5676,
      "mean_token_accuracy": 0.8544633775949478,
      "num_tokens": 1602235.0,
      "step": 1050
    },
    {
      "entropy": 0.5050226721912623,
      "epoch": 0.726525017135024,
      "grad_norm": 0.4335126578807831,
      "learning_rate": 0.00015161069225496916,
      "loss": 0.5754,
      "mean_token_accuracy": 0.859102988243103,
      "num_tokens": 1613841.0,
      "step": 1060
    },
    {
      "entropy": 0.532437426224351,
      "epoch": 0.7333790267306374,
      "grad_norm": 1.1267528533935547,
      "learning_rate": 0.00015115375828192825,
      "loss": 0.5769,
      "mean_token_accuracy": 0.8577256396412849,
      "num_tokens": 1630682.0,
      "step": 1070
    },
    {
      "entropy": 0.44471236430108546,
      "epoch": 0.7402330363262508,
      "grad_norm": 0.38670477271080017,
      "learning_rate": 0.00015069682430888737,
      "loss": 0.5022,
      "mean_token_accuracy": 0.8728992164134979,
      "num_tokens": 1644143.0,
      "step": 1080
    },
    {
      "entropy": 0.5494170293211937,
      "epoch": 0.7470870459218643,
      "grad_norm": 0.5468182563781738,
      "learning_rate": 0.00015023989033584648,
      "loss": 0.5275,
      "mean_token_accuracy": 0.8561429738998413,
      "num_tokens": 1655963.0,
      "step": 1090
    },
    {
      "entropy": 0.45474576093256475,
      "epoch": 0.7539410555174777,
      "grad_norm": 0.41850578784942627,
      "learning_rate": 0.00014978295636280557,
      "loss": 0.4832,
      "mean_token_accuracy": 0.8719681158661843,
      "num_tokens": 1671647.0,
      "step": 1100
    },
    {
      "entropy": 0.5208310384303332,
      "epoch": 0.7607950651130911,
      "grad_norm": 0.4063032865524292,
      "learning_rate": 0.00014932602238976468,
      "loss": 0.4723,
      "mean_token_accuracy": 0.8688164860010147,
      "num_tokens": 1686566.0,
      "step": 1110
    },
    {
      "entropy": 0.48968234546482564,
      "epoch": 0.7676490747087046,
      "grad_norm": 0.4851994216442108,
      "learning_rate": 0.0001488690884167238,
      "loss": 0.5295,
      "mean_token_accuracy": 0.8702765092253685,
      "num_tokens": 1701964.0,
      "step": 1120
    },
    {
      "entropy": 0.4356084756553173,
      "epoch": 0.774503084304318,
      "grad_norm": 0.3857946991920471,
      "learning_rate": 0.00014841215444368291,
      "loss": 0.4956,
      "mean_token_accuracy": 0.8722770050168037,
      "num_tokens": 1717179.0,
      "step": 1130
    },
    {
      "entropy": 0.4641384145244956,
      "epoch": 0.7813570938999315,
      "grad_norm": 0.6515609622001648,
      "learning_rate": 0.000147955220470642,
      "loss": 0.5409,
      "mean_token_accuracy": 0.8693532347679138,
      "num_tokens": 1731173.0,
      "step": 1140
    },
    {
      "entropy": 0.5151202550157905,
      "epoch": 0.7882111034955449,
      "grad_norm": 0.397576242685318,
      "learning_rate": 0.00014749828649760112,
      "loss": 0.5038,
      "mean_token_accuracy": 0.8632552906870842,
      "num_tokens": 1747319.0,
      "step": 1150
    },
    {
      "entropy": 0.5475881166756154,
      "epoch": 0.7950651130911583,
      "grad_norm": 0.6286647319793701,
      "learning_rate": 0.0001470413525245602,
      "loss": 0.5117,
      "mean_token_accuracy": 0.8562549948692322,
      "num_tokens": 1761875.0,
      "step": 1160
    },
    {
      "entropy": 0.529462856054306,
      "epoch": 0.8019191226867718,
      "grad_norm": 0.35311031341552734,
      "learning_rate": 0.00014658441855151932,
      "loss": 0.5344,
      "mean_token_accuracy": 0.8580666840076446,
      "num_tokens": 1775404.0,
      "step": 1170
    },
    {
      "entropy": 0.5058778524398804,
      "epoch": 0.8087731322823852,
      "grad_norm": 0.4609556496143341,
      "learning_rate": 0.00014612748457847844,
      "loss": 0.5985,
      "mean_token_accuracy": 0.8641025006771088,
      "num_tokens": 1792569.0,
      "step": 1180
    },
    {
      "entropy": 0.48940622098743913,
      "epoch": 0.8156271418779987,
      "grad_norm": 0.4840541481971741,
      "learning_rate": 0.00014567055060543752,
      "loss": 0.5918,
      "mean_token_accuracy": 0.8665370926260948,
      "num_tokens": 1809055.0,
      "step": 1190
    },
    {
      "entropy": 0.48019982874393463,
      "epoch": 0.8224811514736121,
      "grad_norm": 0.45010054111480713,
      "learning_rate": 0.00014521361663239664,
      "loss": 0.5367,
      "mean_token_accuracy": 0.8690081983804703,
      "num_tokens": 1822135.0,
      "step": 1200
    },
    {
      "entropy": 0.40929838865995405,
      "epoch": 0.8293351610692254,
      "grad_norm": 0.5522090196609497,
      "learning_rate": 0.00014475668265935573,
      "loss": 0.393,
      "mean_token_accuracy": 0.887654384970665,
      "num_tokens": 1838028.0,
      "step": 1210
    },
    {
      "entropy": 0.5324885908514261,
      "epoch": 0.8361891706648389,
      "grad_norm": 0.34172698855400085,
      "learning_rate": 0.00014429974868631484,
      "loss": 0.5601,
      "mean_token_accuracy": 0.8547317892313003,
      "num_tokens": 1852880.0,
      "step": 1220
    },
    {
      "entropy": 0.4797104798257351,
      "epoch": 0.8430431802604523,
      "grad_norm": 0.4826546311378479,
      "learning_rate": 0.00014384281471327393,
      "loss": 0.6045,
      "mean_token_accuracy": 0.8638340055942535,
      "num_tokens": 1863729.0,
      "step": 1230
    },
    {
      "entropy": 0.6502325102686882,
      "epoch": 0.8498971898560658,
      "grad_norm": 0.5058287382125854,
      "learning_rate": 0.00014338588074023305,
      "loss": 0.6526,
      "mean_token_accuracy": 0.8361533902585506,
      "num_tokens": 1880647.0,
      "step": 1240
    },
    {
      "entropy": 0.5473382823169232,
      "epoch": 0.8567511994516792,
      "grad_norm": 0.4426461458206177,
      "learning_rate": 0.00014292894676719216,
      "loss": 0.5567,
      "mean_token_accuracy": 0.8567781046032905,
      "num_tokens": 1897892.0,
      "step": 1250
    },
    {
      "entropy": 0.4634023506194353,
      "epoch": 0.8636052090472927,
      "grad_norm": 0.36186039447784424,
      "learning_rate": 0.00014247201279415125,
      "loss": 0.4819,
      "mean_token_accuracy": 0.8789055705070495,
      "num_tokens": 1914142.0,
      "step": 1260
    },
    {
      "entropy": 0.4072172209620476,
      "epoch": 0.8704592186429061,
      "grad_norm": 1.0110243558883667,
      "learning_rate": 0.00014201507882111037,
      "loss": 0.4309,
      "mean_token_accuracy": 0.8848265528678894,
      "num_tokens": 1928985.0,
      "step": 1270
    },
    {
      "entropy": 0.473388434574008,
      "epoch": 0.8773132282385195,
      "grad_norm": 0.4536840319633484,
      "learning_rate": 0.00014155814484806945,
      "loss": 0.5185,
      "mean_token_accuracy": 0.8669295862317086,
      "num_tokens": 1946333.0,
      "step": 1280
    },
    {
      "entropy": 0.4370680347084999,
      "epoch": 0.884167237834133,
      "grad_norm": 0.47696638107299805,
      "learning_rate": 0.00014110121087502857,
      "loss": 0.4653,
      "mean_token_accuracy": 0.8775255620479584,
      "num_tokens": 1960422.0,
      "step": 1290
    },
    {
      "entropy": 0.43862779829651116,
      "epoch": 0.8910212474297464,
      "grad_norm": 0.5211741924285889,
      "learning_rate": 0.00014064427690198766,
      "loss": 0.4933,
      "mean_token_accuracy": 0.8786178231239319,
      "num_tokens": 1975560.0,
      "step": 1300
    },
    {
      "entropy": 0.5589355498552322,
      "epoch": 0.8978752570253599,
      "grad_norm": 0.4128772020339966,
      "learning_rate": 0.00014018734292894677,
      "loss": 0.582,
      "mean_token_accuracy": 0.851758947968483,
      "num_tokens": 1989833.0,
      "step": 1310
    },
    {
      "entropy": 0.49184365533292296,
      "epoch": 0.9047292666209733,
      "grad_norm": 0.41796982288360596,
      "learning_rate": 0.0001397304089559059,
      "loss": 0.5851,
      "mean_token_accuracy": 0.8666749432682991,
      "num_tokens": 2007135.0,
      "step": 1320
    },
    {
      "entropy": 0.5475243296474218,
      "epoch": 0.9115832762165867,
      "grad_norm": 0.34859004616737366,
      "learning_rate": 0.00013927347498286498,
      "loss": 0.6538,
      "mean_token_accuracy": 0.8524001136422157,
      "num_tokens": 2024798.0,
      "step": 1330
    },
    {
      "entropy": 0.5540970169007778,
      "epoch": 0.9184372858122002,
      "grad_norm": 0.7016645669937134,
      "learning_rate": 0.0001388165410098241,
      "loss": 0.5823,
      "mean_token_accuracy": 0.8562126755714417,
      "num_tokens": 2039195.0,
      "step": 1340
    },
    {
      "entropy": 0.5045036463066935,
      "epoch": 0.9252912954078135,
      "grad_norm": 0.452700674533844,
      "learning_rate": 0.00013835960703678318,
      "loss": 0.5448,
      "mean_token_accuracy": 0.8633385166525841,
      "num_tokens": 2051288.0,
      "step": 1350
    },
    {
      "entropy": 0.4977139551192522,
      "epoch": 0.932145305003427,
      "grad_norm": 0.5281541347503662,
      "learning_rate": 0.0001379026730637423,
      "loss": 0.5267,
      "mean_token_accuracy": 0.8511179909110069,
      "num_tokens": 2066556.0,
      "step": 1360
    },
    {
      "entropy": 0.44772500842809676,
      "epoch": 0.9389993145990404,
      "grad_norm": 0.3925575017929077,
      "learning_rate": 0.0001374457390907014,
      "loss": 0.4809,
      "mean_token_accuracy": 0.8780866637825966,
      "num_tokens": 2081977.0,
      "step": 1370
    },
    {
      "entropy": 0.4240812066942453,
      "epoch": 0.9458533241946538,
      "grad_norm": 0.38476476073265076,
      "learning_rate": 0.0001369888051176605,
      "loss": 0.4692,
      "mean_token_accuracy": 0.8760198578238487,
      "num_tokens": 2096240.0,
      "step": 1380
    },
    {
      "entropy": 0.45738366320729257,
      "epoch": 0.9527073337902673,
      "grad_norm": 0.3925527036190033,
      "learning_rate": 0.00013653187114461961,
      "loss": 0.4463,
      "mean_token_accuracy": 0.8781074970960617,
      "num_tokens": 2112542.0,
      "step": 1390
    },
    {
      "entropy": 0.6118490558117629,
      "epoch": 0.9595613433858807,
      "grad_norm": 0.5079460144042969,
      "learning_rate": 0.0001360749371715787,
      "loss": 0.5996,
      "mean_token_accuracy": 0.8484726428985596,
      "num_tokens": 2128158.0,
      "step": 1400
    },
    {
      "entropy": 0.5192625004798174,
      "epoch": 0.9664153529814942,
      "grad_norm": 0.5811266899108887,
      "learning_rate": 0.00013561800319853782,
      "loss": 0.4973,
      "mean_token_accuracy": 0.8583417847752571,
      "num_tokens": 2144185.0,
      "step": 1410
    },
    {
      "entropy": 0.424515537545085,
      "epoch": 0.9732693625771076,
      "grad_norm": 0.530555784702301,
      "learning_rate": 0.0001351610692254969,
      "loss": 0.4075,
      "mean_token_accuracy": 0.8812750861048698,
      "num_tokens": 2156802.0,
      "step": 1420
    },
    {
      "entropy": 0.43874580282717945,
      "epoch": 0.980123372172721,
      "grad_norm": 0.37783336639404297,
      "learning_rate": 0.00013470413525245602,
      "loss": 0.4982,
      "mean_token_accuracy": 0.8769993513822556,
      "num_tokens": 2173093.0,
      "step": 1430
    },
    {
      "entropy": 0.5104522176086903,
      "epoch": 0.9869773817683345,
      "grad_norm": 0.3972949981689453,
      "learning_rate": 0.00013424720127941514,
      "loss": 0.5447,
      "mean_token_accuracy": 0.866170509159565,
      "num_tokens": 2186974.0,
      "step": 1440
    },
    {
      "entropy": 0.412432143650949,
      "epoch": 0.9938313913639479,
      "grad_norm": 0.4336175322532654,
      "learning_rate": 0.00013379026730637422,
      "loss": 0.4758,
      "mean_token_accuracy": 0.8839898869395256,
      "num_tokens": 2205666.0,
      "step": 1450
    },
    {
      "entropy": 0.43371021617203953,
      "epoch": 1.0006854009595614,
      "grad_norm": 0.35590001940727234,
      "learning_rate": 0.00013333333333333334,
      "loss": 0.4562,
      "mean_token_accuracy": 0.8767066270112991,
      "num_tokens": 2219486.0,
      "step": 1460
    },
    {
      "entropy": 0.5512058276683092,
      "epoch": 1.0075394105551747,
      "grad_norm": 0.5244994163513184,
      "learning_rate": 0.00013287639936029243,
      "loss": 0.5225,
      "mean_token_accuracy": 0.8718675941228866,
      "num_tokens": 2236772.0,
      "step": 1470
    },
    {
      "entropy": 0.4393087295815349,
      "epoch": 1.0143934201507883,
      "grad_norm": 0.4118908941745758,
      "learning_rate": 0.00013241946538725154,
      "loss": 0.4569,
      "mean_token_accuracy": 0.8839558228850365,
      "num_tokens": 2251853.0,
      "step": 1480
    },
    {
      "entropy": 0.41823681965470316,
      "epoch": 1.0212474297464016,
      "grad_norm": 0.3691245913505554,
      "learning_rate": 0.00013196253141421063,
      "loss": 0.394,
      "mean_token_accuracy": 0.882546167075634,
      "num_tokens": 2267899.0,
      "step": 1490
    },
    {
      "entropy": 0.4980747589841485,
      "epoch": 1.0281014393420151,
      "grad_norm": 0.37798306345939636,
      "learning_rate": 0.00013150559744116975,
      "loss": 0.4356,
      "mean_token_accuracy": 0.8722493484616279,
      "num_tokens": 2283255.0,
      "step": 1500
    },
    {
      "entropy": 0.4623454766348004,
      "epoch": 1.0349554489376285,
      "grad_norm": 0.579682469367981,
      "learning_rate": 0.00013104866346812886,
      "loss": 0.4834,
      "mean_token_accuracy": 0.8708335652947425,
      "num_tokens": 2300951.0,
      "step": 1510
    },
    {
      "entropy": 0.410476384870708,
      "epoch": 1.041809458533242,
      "grad_norm": 0.5235070586204529,
      "learning_rate": 0.00013059172949508795,
      "loss": 0.4059,
      "mean_token_accuracy": 0.8934275403618812,
      "num_tokens": 2318518.0,
      "step": 1520
    },
    {
      "entropy": 0.46547266785055397,
      "epoch": 1.0486634681288554,
      "grad_norm": 0.5329645872116089,
      "learning_rate": 0.00013013479552204707,
      "loss": 0.4878,
      "mean_token_accuracy": 0.8683855041861535,
      "num_tokens": 2333514.0,
      "step": 1530
    },
    {
      "entropy": 0.4657827939838171,
      "epoch": 1.055517477724469,
      "grad_norm": 0.4697127640247345,
      "learning_rate": 0.00012967786154900618,
      "loss": 0.5924,
      "mean_token_accuracy": 0.8728780604898929,
      "num_tokens": 2349021.0,
      "step": 1540
    },
    {
      "entropy": 0.4068853797391057,
      "epoch": 1.0623714873200822,
      "grad_norm": 0.5506519675254822,
      "learning_rate": 0.0001292209275759653,
      "loss": 0.4642,
      "mean_token_accuracy": 0.8888774290680885,
      "num_tokens": 2365215.0,
      "step": 1550
    },
    {
      "entropy": 0.46772981975227595,
      "epoch": 1.0692254969156956,
      "grad_norm": 0.4057336747646332,
      "learning_rate": 0.00012876399360292438,
      "loss": 0.5222,
      "mean_token_accuracy": 0.8738009124994278,
      "num_tokens": 2380948.0,
      "step": 1560
    },
    {
      "entropy": 0.42752092853188517,
      "epoch": 1.0760795065113091,
      "grad_norm": 0.47754740715026855,
      "learning_rate": 0.0001283070596298835,
      "loss": 0.4535,
      "mean_token_accuracy": 0.8837641194462776,
      "num_tokens": 2394485.0,
      "step": 1570
    },
    {
      "entropy": 0.4943581625819206,
      "epoch": 1.0829335161069225,
      "grad_norm": 1.1818515062332153,
      "learning_rate": 0.0001278501256568426,
      "loss": 0.4384,
      "mean_token_accuracy": 0.8675463929772377,
      "num_tokens": 2410923.0,
      "step": 1580
    },
    {
      "entropy": 0.4559463683515787,
      "epoch": 1.089787525702536,
      "grad_norm": 0.8639698624610901,
      "learning_rate": 0.0001273931916838017,
      "loss": 0.4358,
      "mean_token_accuracy": 0.874145744740963,
      "num_tokens": 2424582.0,
      "step": 1590
    },
    {
      "entropy": 0.48699437770992515,
      "epoch": 1.0966415352981493,
      "grad_norm": 0.5072070956230164,
      "learning_rate": 0.00012693625771076082,
      "loss": 0.4742,
      "mean_token_accuracy": 0.8715042397379875,
      "num_tokens": 2438857.0,
      "step": 1600
    },
    {
      "entropy": 0.37908787596970794,
      "epoch": 1.103495544893763,
      "grad_norm": 0.38422709703445435,
      "learning_rate": 0.0001264793237377199,
      "loss": 0.3833,
      "mean_token_accuracy": 0.8959223717451096,
      "num_tokens": 2450917.0,
      "step": 1610
    },
    {
      "entropy": 0.41005379147827625,
      "epoch": 1.1103495544893762,
      "grad_norm": 0.7326657176017761,
      "learning_rate": 0.00012602238976467902,
      "loss": 0.5058,
      "mean_token_accuracy": 0.8892273604869843,
      "num_tokens": 2467673.0,
      "step": 1620
    },
    {
      "entropy": 0.39053783323615787,
      "epoch": 1.1172035640849898,
      "grad_norm": 0.7487586140632629,
      "learning_rate": 0.0001255654557916381,
      "loss": 0.4545,
      "mean_token_accuracy": 0.8870472058653831,
      "num_tokens": 2479327.0,
      "step": 1630
    },
    {
      "entropy": 0.35662307422608136,
      "epoch": 1.124057573680603,
      "grad_norm": 0.4632146656513214,
      "learning_rate": 0.00012510852181859723,
      "loss": 0.4,
      "mean_token_accuracy": 0.9051155418157577,
      "num_tokens": 2492353.0,
      "step": 1640
    },
    {
      "entropy": 0.4149196095764637,
      "epoch": 1.1309115832762167,
      "grad_norm": 0.6427090167999268,
      "learning_rate": 0.00012465158784555634,
      "loss": 0.4389,
      "mean_token_accuracy": 0.8885096952319145,
      "num_tokens": 2506780.0,
      "step": 1650
    },
    {
      "entropy": 0.4400394601747394,
      "epoch": 1.13776559287183,
      "grad_norm": 1.0589375495910645,
      "learning_rate": 0.00012419465387251543,
      "loss": 0.4414,
      "mean_token_accuracy": 0.8822434306144714,
      "num_tokens": 2517981.0,
      "step": 1660
    },
    {
      "entropy": 0.39627051912248135,
      "epoch": 1.1446196024674435,
      "grad_norm": 0.5921131372451782,
      "learning_rate": 0.00012373771989947454,
      "loss": 0.4153,
      "mean_token_accuracy": 0.8856975853443145,
      "num_tokens": 2534627.0,
      "step": 1670
    },
    {
      "entropy": 0.3326151974499226,
      "epoch": 1.1514736120630569,
      "grad_norm": 0.3639609217643738,
      "learning_rate": 0.00012328078592643363,
      "loss": 0.3653,
      "mean_token_accuracy": 0.9019765496253968,
      "num_tokens": 2552034.0,
      "step": 1680
    },
    {
      "entropy": 0.3960498865693808,
      "epoch": 1.1583276216586702,
      "grad_norm": 0.47495174407958984,
      "learning_rate": 0.00012282385195339275,
      "loss": 0.3592,
      "mean_token_accuracy": 0.8895070508122445,
      "num_tokens": 2560423.0,
      "step": 1690
    },
    {
      "entropy": 0.47324896343052386,
      "epoch": 1.1651816312542838,
      "grad_norm": 0.46483132243156433,
      "learning_rate": 0.00012236691798035184,
      "loss": 0.4973,
      "mean_token_accuracy": 0.8739207655191421,
      "num_tokens": 2576604.0,
      "step": 1700
    },
    {
      "entropy": 0.4374368604272604,
      "epoch": 1.1720356408498973,
      "grad_norm": 0.4708305597305298,
      "learning_rate": 0.00012190998400731095,
      "loss": 0.5012,
      "mean_token_accuracy": 0.8828093633055687,
      "num_tokens": 2594330.0,
      "step": 1710
    },
    {
      "entropy": 0.4581823544576764,
      "epoch": 1.1788896504455106,
      "grad_norm": 0.32159796357154846,
      "learning_rate": 0.00012145305003427005,
      "loss": 0.4913,
      "mean_token_accuracy": 0.87235506772995,
      "num_tokens": 2610206.0,
      "step": 1720
    },
    {
      "entropy": 0.48096910435706375,
      "epoch": 1.185743660041124,
      "grad_norm": 0.4397423565387726,
      "learning_rate": 0.00012099611606122915,
      "loss": 0.4941,
      "mean_token_accuracy": 0.873574560880661,
      "num_tokens": 2625531.0,
      "step": 1730
    },
    {
      "entropy": 0.41527645140886305,
      "epoch": 1.1925976696367375,
      "grad_norm": 0.5944206118583679,
      "learning_rate": 0.00012053918208818826,
      "loss": 0.4278,
      "mean_token_accuracy": 0.8867389008402824,
      "num_tokens": 2642344.0,
      "step": 1740
    },
    {
      "entropy": 0.39353391956537964,
      "epoch": 1.1994516792323509,
      "grad_norm": 0.4932354688644409,
      "learning_rate": 0.00012008224811514737,
      "loss": 0.4537,
      "mean_token_accuracy": 0.8972290739417076,
      "num_tokens": 2656168.0,
      "step": 1750
    },
    {
      "entropy": 0.37847189866006375,
      "epoch": 1.2063056888279644,
      "grad_norm": 0.9315231442451477,
      "learning_rate": 0.00011962531414210647,
      "loss": 0.3621,
      "mean_token_accuracy": 0.8991809234023094,
      "num_tokens": 2671640.0,
      "step": 1760
    },
    {
      "entropy": 0.39046813510358336,
      "epoch": 1.2131596984235777,
      "grad_norm": 0.7334297895431519,
      "learning_rate": 0.00011916838016906557,
      "loss": 0.4152,
      "mean_token_accuracy": 0.8910555317997932,
      "num_tokens": 2688325.0,
      "step": 1770
    },
    {
      "entropy": 0.47341820131987333,
      "epoch": 1.2200137080191913,
      "grad_norm": 0.6043763160705566,
      "learning_rate": 0.00011871144619602468,
      "loss": 0.5287,
      "mean_token_accuracy": 0.8681931465864181,
      "num_tokens": 2701964.0,
      "step": 1780
    },
    {
      "entropy": 0.38153773359954357,
      "epoch": 1.2268677176148046,
      "grad_norm": 0.7347066402435303,
      "learning_rate": 0.00011825451222298378,
      "loss": 0.3988,
      "mean_token_accuracy": 0.8908339574933052,
      "num_tokens": 2717051.0,
      "step": 1790
    },
    {
      "entropy": 0.4114368222653866,
      "epoch": 1.2337217272104182,
      "grad_norm": 0.8183625936508179,
      "learning_rate": 0.00011779757824994288,
      "loss": 0.3782,
      "mean_token_accuracy": 0.8912013128399849,
      "num_tokens": 2729159.0,
      "step": 1800
    },
    {
      "entropy": 0.43790148589760064,
      "epoch": 1.2405757368060315,
      "grad_norm": 0.8269321322441101,
      "learning_rate": 0.000117340644276902,
      "loss": 0.4837,
      "mean_token_accuracy": 0.8785713255405426,
      "num_tokens": 2742211.0,
      "step": 1810
    },
    {
      "entropy": 0.44326872155070307,
      "epoch": 1.247429746401645,
      "grad_norm": 0.6595312356948853,
      "learning_rate": 0.0001168837103038611,
      "loss": 0.4289,
      "mean_token_accuracy": 0.878677237033844,
      "num_tokens": 2759715.0,
      "step": 1820
    },
    {
      "entropy": 0.4569962047040462,
      "epoch": 1.2542837559972584,
      "grad_norm": 0.6932032704353333,
      "learning_rate": 0.0001164267763308202,
      "loss": 0.4707,
      "mean_token_accuracy": 0.8800874799489975,
      "num_tokens": 2771679.0,
      "step": 1830
    },
    {
      "entropy": 0.414714134298265,
      "epoch": 1.261137765592872,
      "grad_norm": 0.33980950713157654,
      "learning_rate": 0.0001159698423577793,
      "loss": 0.4363,
      "mean_token_accuracy": 0.8790075898170471,
      "num_tokens": 2786737.0,
      "step": 1840
    },
    {
      "entropy": 0.39454092010855674,
      "epoch": 1.2679917751884853,
      "grad_norm": 0.98993319272995,
      "learning_rate": 0.0001155129083847384,
      "loss": 0.4543,
      "mean_token_accuracy": 0.8888913974165916,
      "num_tokens": 2799933.0,
      "step": 1850
    },
    {
      "entropy": 0.39810120295733215,
      "epoch": 1.2748457847840986,
      "grad_norm": 0.7984906435012817,
      "learning_rate": 0.0001150559744116975,
      "loss": 0.3674,
      "mean_token_accuracy": 0.8919711828231811,
      "num_tokens": 2817711.0,
      "step": 1860
    },
    {
      "entropy": 0.3886662429198623,
      "epoch": 1.2816997943797122,
      "grad_norm": 0.5177662372589111,
      "learning_rate": 0.00011459904043865662,
      "loss": 0.4052,
      "mean_token_accuracy": 0.8918895840644836,
      "num_tokens": 2833956.0,
      "step": 1870
    },
    {
      "entropy": 0.365813566558063,
      "epoch": 1.2885538039753255,
      "grad_norm": 0.6121789813041687,
      "learning_rate": 0.00011414210646561572,
      "loss": 0.3742,
      "mean_token_accuracy": 0.894109770655632,
      "num_tokens": 2849690.0,
      "step": 1880
    },
    {
      "entropy": 0.29113590624183416,
      "epoch": 1.295407813570939,
      "grad_norm": 0.4612710475921631,
      "learning_rate": 0.00011368517249257482,
      "loss": 0.325,
      "mean_token_accuracy": 0.9130572691559792,
      "num_tokens": 2863377.0,
      "step": 1890
    },
    {
      "entropy": 0.425050226226449,
      "epoch": 1.3022618231665524,
      "grad_norm": 0.5517224073410034,
      "learning_rate": 0.00011322823851953392,
      "loss": 0.4724,
      "mean_token_accuracy": 0.8836025670170784,
      "num_tokens": 2880706.0,
      "step": 1900
    },
    {
      "entropy": 0.3978625874966383,
      "epoch": 1.309115832762166,
      "grad_norm": 0.5483404994010925,
      "learning_rate": 0.00011277130454649303,
      "loss": 0.4356,
      "mean_token_accuracy": 0.8916041046380997,
      "num_tokens": 2899104.0,
      "step": 1910
    },
    {
      "entropy": 0.40184173360466957,
      "epoch": 1.3159698423577793,
      "grad_norm": 0.48635584115982056,
      "learning_rate": 0.00011231437057345213,
      "loss": 0.4383,
      "mean_token_accuracy": 0.8931035652756691,
      "num_tokens": 2914219.0,
      "step": 1920
    },
    {
      "entropy": 0.41671360936015844,
      "epoch": 1.3228238519533928,
      "grad_norm": 0.4210641384124756,
      "learning_rate": 0.00011185743660041123,
      "loss": 0.4472,
      "mean_token_accuracy": 0.8848306000232696,
      "num_tokens": 2929365.0,
      "step": 1930
    },
    {
      "entropy": 0.43334737941622736,
      "epoch": 1.3296778615490061,
      "grad_norm": 0.44904014468193054,
      "learning_rate": 0.00011140050262737035,
      "loss": 0.381,
      "mean_token_accuracy": 0.879435133934021,
      "num_tokens": 2942463.0,
      "step": 1940
    },
    {
      "entropy": 0.4951056869700551,
      "epoch": 1.3365318711446195,
      "grad_norm": 0.4268065392971039,
      "learning_rate": 0.00011094356865432945,
      "loss": 0.4273,
      "mean_token_accuracy": 0.8738623633980751,
      "num_tokens": 2956810.0,
      "step": 1950
    },
    {
      "entropy": 0.4137628199532628,
      "epoch": 1.343385880740233,
      "grad_norm": 0.5589777231216431,
      "learning_rate": 0.00011048663468128858,
      "loss": 0.461,
      "mean_token_accuracy": 0.8912892028689384,
      "num_tokens": 2971407.0,
      "step": 1960
    },
    {
      "entropy": 0.40699709448963406,
      "epoch": 1.3502398903358466,
      "grad_norm": 0.6441206336021423,
      "learning_rate": 0.00011002970070824768,
      "loss": 0.4126,
      "mean_token_accuracy": 0.8863832727074623,
      "num_tokens": 2986374.0,
      "step": 1970
    },
    {
      "entropy": 0.4124022351577878,
      "epoch": 1.35709389993146,
      "grad_norm": 0.4372117519378662,
      "learning_rate": 0.00010957276673520678,
      "loss": 0.4399,
      "mean_token_accuracy": 0.8908446505665779,
      "num_tokens": 3002339.0,
      "step": 1980
    },
    {
      "entropy": 0.3705425701104105,
      "epoch": 1.3639479095270732,
      "grad_norm": 0.8048654794692993,
      "learning_rate": 0.00010911583276216588,
      "loss": 0.3946,
      "mean_token_accuracy": 0.8922546565532684,
      "num_tokens": 3013169.0,
      "step": 1990
    },
    {
      "entropy": 0.3852098049595952,
      "epoch": 1.3708019191226868,
      "grad_norm": 0.7653835415840149,
      "learning_rate": 0.00010865889878912498,
      "loss": 0.3783,
      "mean_token_accuracy": 0.892442500591278,
      "num_tokens": 3026414.0,
      "step": 2000
    }
  ],
  "logging_steps": 10,
  "max_steps": 4377,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 2.4129091159209984e+16,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}