{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 3284,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0060901339829476245,
      "grad_norm": 23.5319766998291,
      "learning_rate": 1.2162162162162164e-05,
      "loss": 4.5905,
      "mean_token_accuracy": 0.3401473943144083,
      "num_tokens": 132681.0,
      "step": 10
    },
    {
      "epoch": 0.012180267965895249,
      "grad_norm": 6.916630744934082,
      "learning_rate": 2.5675675675675675e-05,
      "loss": 3.957,
      "mean_token_accuracy": 0.3799716055393219,
      "num_tokens": 264238.0,
      "step": 20
    },
    {
      "epoch": 0.018270401948842874,
      "grad_norm": 2.717982292175293,
      "learning_rate": 3.918918918918919e-05,
      "loss": 3.047,
      "mean_token_accuracy": 0.46811963245272636,
      "num_tokens": 401308.0,
      "step": 30
    },
    {
      "epoch": 0.024360535931790498,
      "grad_norm": 2.407865524291992,
      "learning_rate": 5.27027027027027e-05,
      "loss": 2.5322,
      "mean_token_accuracy": 0.5190828196704388,
      "num_tokens": 532444.0,
      "step": 40
    },
    {
      "epoch": 0.030450669914738125,
      "grad_norm": 1.018301248550415,
      "learning_rate": 6.621621621621621e-05,
      "loss": 2.1325,
      "mean_token_accuracy": 0.5782605841755867,
      "num_tokens": 660406.0,
      "step": 50
    },
    {
      "epoch": 0.03654080389768575,
      "grad_norm": 0.7315741181373596,
      "learning_rate": 7.972972972972974e-05,
      "loss": 1.9044,
      "mean_token_accuracy": 0.6264914631843567,
      "num_tokens": 795304.0,
      "step": 60
    },
    {
      "epoch": 0.04263093788063337,
      "grad_norm": 0.6652920246124268,
      "learning_rate": 9.324324324324324e-05,
      "loss": 1.633,
      "mean_token_accuracy": 0.6683938711881637,
      "num_tokens": 934543.0,
      "step": 70
    },
    {
      "epoch": 0.048721071863580996,
      "grad_norm": 0.6119660139083862,
      "learning_rate": 0.00010675675675675677,
      "loss": 1.543,
      "mean_token_accuracy": 0.6834091022610664,
      "num_tokens": 1070669.0,
      "step": 80
    },
    {
      "epoch": 0.05481120584652863,
      "grad_norm": 0.591424286365509,
      "learning_rate": 0.00012027027027027027,
      "loss": 1.4154,
      "mean_token_accuracy": 0.6991497233510018,
      "num_tokens": 1211114.0,
      "step": 90
    },
    {
      "epoch": 0.06090133982947625,
      "grad_norm": 0.5663530230522156,
      "learning_rate": 0.0001337837837837838,
      "loss": 1.3176,
      "mean_token_accuracy": 0.7089206710457802,
      "num_tokens": 1349584.0,
      "step": 100
    },
    {
      "epoch": 0.06699147381242387,
      "grad_norm": 0.5881878137588501,
      "learning_rate": 0.0001472972972972973,
      "loss": 1.2293,
      "mean_token_accuracy": 0.7254403859376908,
      "num_tokens": 1487515.0,
      "step": 110
    },
    {
      "epoch": 0.0730816077953715,
      "grad_norm": 0.7664394974708557,
      "learning_rate": 0.00016081081081081083,
      "loss": 1.1814,
      "mean_token_accuracy": 0.7306812778115273,
      "num_tokens": 1618603.0,
      "step": 120
    },
    {
      "epoch": 0.07917174177831912,
      "grad_norm": 0.6155670881271362,
      "learning_rate": 0.00017432432432432432,
      "loss": 1.1967,
      "mean_token_accuracy": 0.7284250959753991,
      "num_tokens": 1750466.0,
      "step": 130
    },
    {
      "epoch": 0.08526187576126674,
      "grad_norm": 0.5296258330345154,
      "learning_rate": 0.00018783783783783784,
      "loss": 1.0955,
      "mean_token_accuracy": 0.7472824215888977,
      "num_tokens": 1887913.0,
      "step": 140
    },
    {
      "epoch": 0.09135200974421437,
      "grad_norm": 0.5564976334571838,
      "learning_rate": 0.00019999998054550544,
      "loss": 1.118,
      "mean_token_accuracy": 0.7397311359643937,
      "num_tokens": 2018579.0,
      "step": 150
    },
    {
      "epoch": 0.09744214372716199,
      "grad_norm": 0.5301142930984497,
      "learning_rate": 0.00019999764601633156,
      "loss": 1.045,
      "mean_token_accuracy": 0.7519380420446395,
      "num_tokens": 2158851.0,
      "step": 160
    },
    {
      "epoch": 0.10353227771010962,
      "grad_norm": 0.5949111580848694,
      "learning_rate": 0.00019999142070388495,
      "loss": 1.0497,
      "mean_token_accuracy": 0.7520910769701004,
      "num_tokens": 2296715.0,
      "step": 170
    },
    {
      "epoch": 0.10962241169305725,
      "grad_norm": 0.6169262528419495,
      "learning_rate": 0.0001999813048772986,
      "loss": 1.0821,
      "mean_token_accuracy": 0.7406247839331627,
      "num_tokens": 2424756.0,
      "step": 180
    },
    {
      "epoch": 0.11571254567600488,
      "grad_norm": 0.58912593126297,
      "learning_rate": 0.00019996729897390057,
      "loss": 1.0286,
      "mean_token_accuracy": 0.7527454376220704,
      "num_tokens": 2559362.0,
      "step": 190
    },
    {
      "epoch": 0.1218026796589525,
      "grad_norm": 0.5084304213523865,
      "learning_rate": 0.00019994940359919483,
      "loss": 0.992,
      "mean_token_accuracy": 0.7640391126275062,
      "num_tokens": 2700231.0,
      "step": 200
    },
    {
      "epoch": 0.1278928136419001,
      "grad_norm": 0.5790796279907227,
      "learning_rate": 0.00019992761952683516,
      "loss": 1.0146,
      "mean_token_accuracy": 0.7554366230964661,
      "num_tokens": 2831324.0,
      "step": 210
    },
    {
      "epoch": 0.13398294762484775,
      "grad_norm": 0.5852051377296448,
      "learning_rate": 0.00019990194769859188,
      "loss": 0.978,
      "mean_token_accuracy": 0.7612502485513687,
      "num_tokens": 2967346.0,
      "step": 220
    },
    {
      "epoch": 0.14007308160779536,
      "grad_norm": 0.5102785229682922,
      "learning_rate": 0.00019987238922431088,
      "loss": 0.9616,
      "mean_token_accuracy": 0.7677591517567635,
      "num_tokens": 3110936.0,
      "step": 230
    },
    {
      "epoch": 0.146163215590743,
      "grad_norm": 0.5472669005393982,
      "learning_rate": 0.00019983894538186576,
      "loss": 0.9535,
      "mean_token_accuracy": 0.76737689524889,
      "num_tokens": 3247496.0,
      "step": 240
    },
    {
      "epoch": 0.15225334957369063,
      "grad_norm": 0.5611053109169006,
      "learning_rate": 0.0001998016176171026,
      "loss": 0.9577,
      "mean_token_accuracy": 0.7626092001795769,
      "num_tokens": 3384178.0,
      "step": 250
    },
    {
      "epoch": 0.15834348355663824,
      "grad_norm": 0.54055255651474,
      "learning_rate": 0.0001997604075437774,
      "loss": 0.9907,
      "mean_token_accuracy": 0.7575223430991173,
      "num_tokens": 3517617.0,
      "step": 260
    },
    {
      "epoch": 0.16443361753958588,
      "grad_norm": 0.558316707611084,
      "learning_rate": 0.0001997153169434864,
      "loss": 0.944,
      "mean_token_accuracy": 0.7664194419980049,
      "num_tokens": 3662878.0,
      "step": 270
    },
    {
      "epoch": 0.1705237515225335,
      "grad_norm": 0.49766939878463745,
      "learning_rate": 0.0001996663477655889,
      "loss": 0.9106,
      "mean_token_accuracy": 0.7760038167238236,
      "num_tokens": 3807411.0,
      "step": 280
    },
    {
      "epoch": 0.17661388550548113,
      "grad_norm": 0.4953667223453522,
      "learning_rate": 0.0001996135021271232,
      "loss": 0.9687,
      "mean_token_accuracy": 0.7605679705739021,
      "num_tokens": 3936840.0,
      "step": 290
    },
    {
      "epoch": 0.18270401948842874,
      "grad_norm": 0.5447947978973389,
      "learning_rate": 0.00019955678231271484,
      "loss": 0.9625,
      "mean_token_accuracy": 0.7603292793035508,
      "num_tokens": 4067826.0,
      "step": 300
    },
    {
      "epoch": 0.18879415347137637,
      "grad_norm": 0.4665842056274414,
      "learning_rate": 0.00019949619077447807,
      "loss": 0.9372,
      "mean_token_accuracy": 0.7676101759076118,
      "num_tokens": 4205887.0,
      "step": 310
    },
    {
      "epoch": 0.19488428745432398,
      "grad_norm": 0.515690267086029,
      "learning_rate": 0.00019943173013190965,
      "loss": 0.923,
      "mean_token_accuracy": 0.7708473294973374,
      "num_tokens": 4342894.0,
      "step": 320
    },
    {
      "epoch": 0.20097442143727162,
      "grad_norm": 0.5831382274627686,
      "learning_rate": 0.00019936340317177565,
      "loss": 0.9203,
      "mean_token_accuracy": 0.7708552837371826,
      "num_tokens": 4477651.0,
      "step": 330
    },
    {
      "epoch": 0.20706455542021923,
      "grad_norm": 0.6162773966789246,
      "learning_rate": 0.0001992912128479911,
      "loss": 0.916,
      "mean_token_accuracy": 0.7702088996767997,
      "num_tokens": 4610746.0,
      "step": 340
    },
    {
      "epoch": 0.21315468940316687,
      "grad_norm": 0.5172462463378906,
      "learning_rate": 0.00019921516228149207,
      "loss": 0.8942,
      "mean_token_accuracy": 0.7741821393370628,
      "num_tokens": 4751175.0,
      "step": 350
    },
    {
      "epoch": 0.2192448233861145,
      "grad_norm": 0.5890468955039978,
      "learning_rate": 0.0001991352547601009,
      "loss": 0.9229,
      "mean_token_accuracy": 0.7691043332219124,
      "num_tokens": 4882328.0,
      "step": 360
    },
    {
      "epoch": 0.22533495736906212,
      "grad_norm": 0.5522404909133911,
      "learning_rate": 0.00019905149373838408,
      "loss": 0.9294,
      "mean_token_accuracy": 0.7646071568131447,
      "num_tokens": 5012181.0,
      "step": 370
    },
    {
      "epoch": 0.23142509135200975,
      "grad_norm": 0.5349445939064026,
      "learning_rate": 0.0001989638828375028,
      "loss": 0.8797,
      "mean_token_accuracy": 0.7771721839904785,
      "num_tokens": 5151133.0,
      "step": 380
    },
    {
      "epoch": 0.23751522533495736,
      "grad_norm": 0.531052827835083,
      "learning_rate": 0.00019887242584505635,
      "loss": 0.9221,
      "mean_token_accuracy": 0.7678465083241462,
      "num_tokens": 5279790.0,
      "step": 390
    },
    {
      "epoch": 0.243605359317905,
      "grad_norm": 0.5126324892044067,
      "learning_rate": 0.00019877712671491864,
      "loss": 0.8862,
      "mean_token_accuracy": 0.7739894777536392,
      "num_tokens": 5412390.0,
      "step": 400
    },
    {
      "epoch": 0.2496954933008526,
      "grad_norm": 0.5111438632011414,
      "learning_rate": 0.00019867798956706693,
      "loss": 0.9005,
      "mean_token_accuracy": 0.7721902653574944,
      "num_tokens": 5545801.0,
      "step": 410
    },
    {
      "epoch": 0.2557856272838002,
      "grad_norm": 0.5488138794898987,
      "learning_rate": 0.00019857501868740402,
      "loss": 0.8988,
      "mean_token_accuracy": 0.7690282896161079,
      "num_tokens": 5673758.0,
      "step": 420
    },
    {
      "epoch": 0.2618757612667479,
      "grad_norm": 0.5497994422912598,
      "learning_rate": 0.0001984682185275727,
      "loss": 0.8802,
      "mean_token_accuracy": 0.7780183687806129,
      "num_tokens": 5813158.0,
      "step": 430
    },
    {
      "epoch": 0.2679658952496955,
      "grad_norm": 0.5478431582450867,
      "learning_rate": 0.0001983575937047635,
      "loss": 0.865,
      "mean_token_accuracy": 0.7785944610834121,
      "num_tokens": 5947367.0,
      "step": 440
    },
    {
      "epoch": 0.2740560292326431,
      "grad_norm": 0.5188766717910767,
      "learning_rate": 0.00019824314900151487,
      "loss": 0.8798,
      "mean_token_accuracy": 0.7752803862094879,
      "num_tokens": 6081060.0,
      "step": 450
    },
    {
      "epoch": 0.2801461632155907,
      "grad_norm": 0.530222475528717,
      "learning_rate": 0.00019812488936550666,
      "loss": 0.8628,
      "mean_token_accuracy": 0.7801630645990372,
      "num_tokens": 6217834.0,
      "step": 460
    },
    {
      "epoch": 0.2862362971985384,
      "grad_norm": 0.5987964868545532,
      "learning_rate": 0.00019800281990934614,
      "loss": 0.8775,
      "mean_token_accuracy": 0.7760324433445931,
      "num_tokens": 6350451.0,
      "step": 470
    },
    {
      "epoch": 0.292326431181486,
      "grad_norm": 0.5468559265136719,
      "learning_rate": 0.0001978769459103468,
      "loss": 0.8721,
      "mean_token_accuracy": 0.7794204503297806,
      "num_tokens": 6484738.0,
      "step": 480
    },
    {
      "epoch": 0.2984165651644336,
      "grad_norm": 0.5541098117828369,
      "learning_rate": 0.0001977472728103005,
      "loss": 0.8785,
      "mean_token_accuracy": 0.7767582029104233,
      "num_tokens": 6619313.0,
      "step": 490
    },
    {
      "epoch": 0.30450669914738127,
      "grad_norm": 0.5134281516075134,
      "learning_rate": 0.0001976138062152419,
      "loss": 0.8717,
      "mean_token_accuracy": 0.7752724394202233,
      "num_tokens": 6753195.0,
      "step": 500
    },
    {
      "epoch": 0.3105968331303289,
      "grad_norm": 0.49164435267448425,
      "learning_rate": 0.00019747655189520633,
      "loss": 0.8757,
      "mean_token_accuracy": 0.7768464118242264,
      "num_tokens": 6890448.0,
      "step": 510
    },
    {
      "epoch": 0.3166869671132765,
      "grad_norm": 0.5899345278739929,
      "learning_rate": 0.00019733551578398023,
      "loss": 0.8322,
      "mean_token_accuracy": 0.7859320402145386,
      "num_tokens": 7027488.0,
      "step": 520
    },
    {
      "epoch": 0.3227771010962241,
      "grad_norm": 0.6552841663360596,
      "learning_rate": 0.0001971907039788447,
      "loss": 0.861,
      "mean_token_accuracy": 0.7770532324910164,
      "num_tokens": 7161184.0,
      "step": 530
    },
    {
      "epoch": 0.32886723507917176,
      "grad_norm": 0.5038822889328003,
      "learning_rate": 0.0001970421227403117,
      "loss": 0.8825,
      "mean_token_accuracy": 0.775890800356865,
      "num_tokens": 7294399.0,
      "step": 540
    },
    {
      "epoch": 0.33495736906211937,
      "grad_norm": 0.5094267129898071,
      "learning_rate": 0.00019688977849185378,
      "loss": 0.8598,
      "mean_token_accuracy": 0.7817838475108146,
      "num_tokens": 7427183.0,
      "step": 550
    },
    {
      "epoch": 0.341047503045067,
      "grad_norm": 0.5282809138298035,
      "learning_rate": 0.00019673367781962594,
      "loss": 0.8463,
      "mean_token_accuracy": 0.7812959104776382,
      "num_tokens": 7561734.0,
      "step": 560
    },
    {
      "epoch": 0.3471376370280146,
      "grad_norm": 0.45355409383773804,
      "learning_rate": 0.00019657382747218123,
      "loss": 0.8207,
      "mean_token_accuracy": 0.7888262197375298,
      "num_tokens": 7706228.0,
      "step": 570
    },
    {
      "epoch": 0.35322777101096225,
      "grad_norm": 0.5162333846092224,
      "learning_rate": 0.00019641023436017883,
      "loss": 0.8235,
      "mean_token_accuracy": 0.7868947923183441,
      "num_tokens": 7846684.0,
      "step": 580
    },
    {
      "epoch": 0.35931790499390986,
      "grad_norm": 0.5194632411003113,
      "learning_rate": 0.00019624290555608526,
      "loss": 0.8129,
      "mean_token_accuracy": 0.7884069249033928,
      "num_tokens": 7986811.0,
      "step": 590
    },
    {
      "epoch": 0.3654080389768575,
      "grad_norm": 0.5494846701622009,
      "learning_rate": 0.00019607184829386882,
      "loss": 0.8084,
      "mean_token_accuracy": 0.7874000474810601,
      "num_tokens": 8124538.0,
      "step": 600
    },
    {
      "epoch": 0.37149817295980514,
      "grad_norm": 0.5368776917457581,
      "learning_rate": 0.0001958970699686866,
      "loss": 0.8225,
      "mean_token_accuracy": 0.783010233938694,
      "num_tokens": 8260529.0,
      "step": 610
    },
    {
      "epoch": 0.37758830694275275,
      "grad_norm": 0.6229024529457092,
      "learning_rate": 0.00019571857813656496,
      "loss": 0.8786,
      "mean_token_accuracy": 0.7753148928284646,
      "num_tokens": 8389042.0,
      "step": 620
    },
    {
      "epoch": 0.38367844092570036,
      "grad_norm": 0.5601000785827637,
      "learning_rate": 0.00019553638051407279,
      "loss": 0.8909,
      "mean_token_accuracy": 0.7745720192790031,
      "num_tokens": 8513603.0,
      "step": 630
    },
    {
      "epoch": 0.38976857490864797,
      "grad_norm": 0.438970685005188,
      "learning_rate": 0.0001953504849779879,
      "loss": 0.8085,
      "mean_token_accuracy": 0.7871840804815292,
      "num_tokens": 8652970.0,
      "step": 640
    },
    {
      "epoch": 0.39585870889159563,
      "grad_norm": 0.5505132079124451,
      "learning_rate": 0.00019516089956495648,
      "loss": 0.8102,
      "mean_token_accuracy": 0.7869585514068603,
      "num_tokens": 8792103.0,
      "step": 650
    },
    {
      "epoch": 0.40194884287454324,
      "grad_norm": 0.5447221398353577,
      "learning_rate": 0.00019496763247114581,
      "loss": 0.8336,
      "mean_token_accuracy": 0.7816034242510795,
      "num_tokens": 8926853.0,
      "step": 660
    },
    {
      "epoch": 0.40803897685749085,
      "grad_norm": 0.4652746915817261,
      "learning_rate": 0.00019477069205188965,
      "loss": 0.8383,
      "mean_token_accuracy": 0.7826304718852043,
      "num_tokens": 9059592.0,
      "step": 670
    },
    {
      "epoch": 0.41412911084043846,
      "grad_norm": 0.42363590002059937,
      "learning_rate": 0.00019457008682132726,
      "loss": 0.847,
      "mean_token_accuracy": 0.7810002073645592,
      "num_tokens": 9193062.0,
      "step": 680
    },
    {
      "epoch": 0.42021924482338613,
      "grad_norm": 0.5209478735923767,
      "learning_rate": 0.00019436582545203518,
      "loss": 0.8766,
      "mean_token_accuracy": 0.7733785718679428,
      "num_tokens": 9315805.0,
      "step": 690
    },
    {
      "epoch": 0.42630937880633374,
      "grad_norm": 0.5176642537117004,
      "learning_rate": 0.00019415791677465237,
      "loss": 0.8155,
      "mean_token_accuracy": 0.7869213685393334,
      "num_tokens": 9448863.0,
      "step": 700
    },
    {
      "epoch": 0.43239951278928135,
      "grad_norm": 0.4531058371067047,
      "learning_rate": 0.00019394636977749843,
      "loss": 0.8096,
      "mean_token_accuracy": 0.7903949975967407,
      "num_tokens": 9589382.0,
      "step": 710
    },
    {
      "epoch": 0.438489646772229,
      "grad_norm": 0.5651549100875854,
      "learning_rate": 0.000193731193606185,
      "loss": 0.8263,
      "mean_token_accuracy": 0.7823062822222709,
      "num_tokens": 9723562.0,
      "step": 720
    },
    {
      "epoch": 0.4445797807551766,
      "grad_norm": 0.5377989411354065,
      "learning_rate": 0.00019351239756322031,
      "loss": 0.7993,
      "mean_token_accuracy": 0.7908329650759697,
      "num_tokens": 9859255.0,
      "step": 730
    },
    {
      "epoch": 0.45066991473812423,
      "grad_norm": 0.5420868396759033,
      "learning_rate": 0.00019328999110760722,
      "loss": 0.8461,
      "mean_token_accuracy": 0.7780480548739434,
      "num_tokens": 9981578.0,
      "step": 740
    },
    {
      "epoch": 0.45676004872107184,
      "grad_norm": 0.4889216125011444,
      "learning_rate": 0.000193063983854434,
      "loss": 0.7652,
      "mean_token_accuracy": 0.7959530428051949,
      "num_tokens": 10122922.0,
      "step": 750
    },
    {
      "epoch": 0.4628501827040195,
      "grad_norm": 0.5044087767601013,
      "learning_rate": 0.00019283438557445893,
      "loss": 0.824,
      "mean_token_accuracy": 0.7845935523509979,
      "num_tokens": 10252854.0,
      "step": 760
    },
    {
      "epoch": 0.4689403166869671,
      "grad_norm": 0.5286466479301453,
      "learning_rate": 0.00019260120619368773,
      "loss": 0.815,
      "mean_token_accuracy": 0.7850656941533088,
      "num_tokens": 10385075.0,
      "step": 770
    },
    {
      "epoch": 0.47503045066991473,
      "grad_norm": 0.5441628694534302,
      "learning_rate": 0.00019236445579294437,
      "loss": 0.8048,
      "mean_token_accuracy": 0.7876680314540863,
      "num_tokens": 10520011.0,
      "step": 780
    },
    {
      "epoch": 0.48112058465286234,
      "grad_norm": 0.49002447724342346,
      "learning_rate": 0.0001921241446074355,
      "loss": 0.8059,
      "mean_token_accuracy": 0.7898563235998154,
      "num_tokens": 10652488.0,
      "step": 790
    },
    {
      "epoch": 0.48721071863581,
      "grad_norm": 0.4479144811630249,
      "learning_rate": 0.0001918802830263077,
      "loss": 0.7913,
      "mean_token_accuracy": 0.7928732186555862,
      "num_tokens": 10785974.0,
      "step": 800
    },
    {
      "epoch": 0.4933008526187576,
      "grad_norm": 0.5007497668266296,
      "learning_rate": 0.00019163288159219853,
      "loss": 0.8083,
      "mean_token_accuracy": 0.7893043681979179,
      "num_tokens": 10920950.0,
      "step": 810
    },
    {
      "epoch": 0.4993909866017052,
      "grad_norm": 0.5289483070373535,
      "learning_rate": 0.00019138195100078064,
      "loss": 0.8033,
      "mean_token_accuracy": 0.7864485770463944,
      "num_tokens": 11056380.0,
      "step": 820
    },
    {
      "epoch": 0.5054811205846529,
      "grad_norm": 0.5604159832000732,
      "learning_rate": 0.0001911275021002994,
      "loss": 0.7652,
      "mean_token_accuracy": 0.7946401730179786,
      "num_tokens": 11196074.0,
      "step": 830
    },
    {
      "epoch": 0.5115712545676004,
      "grad_norm": 0.43645399808883667,
      "learning_rate": 0.00019086954589110397,
      "loss": 0.7724,
      "mean_token_accuracy": 0.7990294560790062,
      "num_tokens": 11337990.0,
      "step": 840
    },
    {
      "epoch": 0.5176613885505481,
      "grad_norm": 0.43992146849632263,
      "learning_rate": 0.0001906080935251716,
      "loss": 0.7612,
      "mean_token_accuracy": 0.7999786615371705,
      "num_tokens": 11481565.0,
      "step": 850
    },
    {
      "epoch": 0.5237515225334958,
      "grad_norm": 0.5595120191574097,
      "learning_rate": 0.0001903431563056256,
      "loss": 0.8266,
      "mean_token_accuracy": 0.7859750911593437,
      "num_tokens": 11611714.0,
      "step": 860
    },
    {
      "epoch": 0.5298416565164433,
      "grad_norm": 0.5001987218856812,
      "learning_rate": 0.0001900747456862467,
      "loss": 0.8506,
      "mean_token_accuracy": 0.779585388302803,
      "num_tokens": 11736573.0,
      "step": 870
    },
    {
      "epoch": 0.535931790499391,
      "grad_norm": 0.430147647857666,
      "learning_rate": 0.00018980287327097784,
      "loss": 0.7707,
      "mean_token_accuracy": 0.795211361348629,
      "num_tokens": 11876859.0,
      "step": 880
    },
    {
      "epoch": 0.5420219244823387,
      "grad_norm": 0.5346289873123169,
      "learning_rate": 0.00018952755081342245,
      "loss": 0.8057,
      "mean_token_accuracy": 0.7871127843856811,
      "num_tokens": 12007654.0,
      "step": 890
    },
    {
      "epoch": 0.5481120584652862,
      "grad_norm": 0.46072253584861755,
      "learning_rate": 0.00018924879021633653,
      "loss": 0.7924,
      "mean_token_accuracy": 0.7913773030042648,
      "num_tokens": 12140520.0,
      "step": 900
    },
    {
      "epoch": 0.5542021924482339,
      "grad_norm": 0.4803653955459595,
      "learning_rate": 0.00018896660353111375,
      "loss": 0.8398,
      "mean_token_accuracy": 0.7807079553604126,
      "num_tokens": 12267219.0,
      "step": 910
    },
    {
      "epoch": 0.5602923264311814,
      "grad_norm": 0.5219636559486389,
      "learning_rate": 0.0001886810029572647,
      "loss": 0.7612,
      "mean_token_accuracy": 0.7993015512824059,
      "num_tokens": 12404646.0,
      "step": 920
    },
    {
      "epoch": 0.5663824604141291,
      "grad_norm": 0.501483142375946,
      "learning_rate": 0.00018839200084188936,
      "loss": 0.7953,
      "mean_token_accuracy": 0.787814213335514,
      "num_tokens": 12538219.0,
      "step": 930
    },
    {
      "epoch": 0.5724725943970768,
      "grad_norm": 0.47334522008895874,
      "learning_rate": 0.00018809960967914346,
      "loss": 0.789,
      "mean_token_accuracy": 0.7928574904799461,
      "num_tokens": 12673805.0,
      "step": 940
    },
    {
      "epoch": 0.5785627283800243,
      "grad_norm": 0.5057492852210999,
      "learning_rate": 0.00018780384210969806,
      "loss": 0.7746,
      "mean_token_accuracy": 0.7947553545236588,
      "num_tokens": 12811727.0,
      "step": 950
    },
    {
      "epoch": 0.584652862362972,
      "grad_norm": 0.5179910659790039,
      "learning_rate": 0.00018750471092019325,
      "loss": 0.7962,
      "mean_token_accuracy": 0.7905686929821968,
      "num_tokens": 12947641.0,
      "step": 960
    },
    {
      "epoch": 0.5907429963459196,
      "grad_norm": 0.45797088742256165,
      "learning_rate": 0.00018720222904268543,
      "loss": 0.7678,
      "mean_token_accuracy": 0.7969774708151818,
      "num_tokens": 13083869.0,
      "step": 970
    },
    {
      "epoch": 0.5968331303288672,
      "grad_norm": 0.48360612988471985,
      "learning_rate": 0.00018689640955408803,
      "loss": 0.7996,
      "mean_token_accuracy": 0.7885591968894005,
      "num_tokens": 13211807.0,
      "step": 980
    },
    {
      "epoch": 0.6029232643118149,
      "grad_norm": 0.4378497004508972,
      "learning_rate": 0.00018658726567560635,
      "loss": 0.7652,
      "mean_token_accuracy": 0.7969291344285011,
      "num_tokens": 13351856.0,
      "step": 990
    },
    {
      "epoch": 0.6090133982947625,
      "grad_norm": 0.4857536852359772,
      "learning_rate": 0.00018627481077216577,
      "loss": 0.7786,
      "mean_token_accuracy": 0.7914443418383599,
      "num_tokens": 13486443.0,
      "step": 1000
    },
    {
      "epoch": 0.6151035322777101,
      "grad_norm": 0.5233064293861389,
      "learning_rate": 0.0001859590583518343,
      "loss": 0.8241,
      "mean_token_accuracy": 0.7811850637197495,
      "num_tokens": 13612035.0,
      "step": 1010
    },
    {
      "epoch": 0.6211936662606578,
      "grad_norm": 0.5328738689422607,
      "learning_rate": 0.00018564002206523816,
      "loss": 0.7502,
      "mean_token_accuracy": 0.7993430674076081,
      "num_tokens": 13756509.0,
      "step": 1020
    },
    {
      "epoch": 0.6272838002436053,
      "grad_norm": 0.47962310910224915,
      "learning_rate": 0.000185317715704972,
      "loss": 0.7984,
      "mean_token_accuracy": 0.7864531084895134,
      "num_tokens": 13883033.0,
      "step": 1030
    },
    {
      "epoch": 0.633373934226553,
      "grad_norm": 0.5685893893241882,
      "learning_rate": 0.0001849921532050024,
      "loss": 0.7869,
      "mean_token_accuracy": 0.7909937381744385,
      "num_tokens": 14015234.0,
      "step": 1040
    },
    {
      "epoch": 0.6394640682095006,
      "grad_norm": 0.49146631360054016,
      "learning_rate": 0.00018466334864006566,
      "loss": 0.7952,
      "mean_token_accuracy": 0.7878949210047722,
      "num_tokens": 14149319.0,
      "step": 1050
    },
    {
      "epoch": 0.6455542021924482,
      "grad_norm": 0.5556225776672363,
      "learning_rate": 0.0001843313162250591,
      "loss": 0.7524,
      "mean_token_accuracy": 0.7994373366236687,
      "num_tokens": 14286868.0,
      "step": 1060
    },
    {
      "epoch": 0.6516443361753959,
      "grad_norm": 0.511379063129425,
      "learning_rate": 0.00018399607031442666,
      "loss": 0.7929,
      "mean_token_accuracy": 0.7921562284231186,
      "num_tokens": 14418354.0,
      "step": 1070
    },
    {
      "epoch": 0.6577344701583435,
      "grad_norm": 0.5019840598106384,
      "learning_rate": 0.00018365762540153836,
      "loss": 0.758,
      "mean_token_accuracy": 0.7989353060722351,
      "num_tokens": 14553174.0,
      "step": 1080
    },
    {
      "epoch": 0.6638246041412911,
      "grad_norm": 0.6032467484474182,
      "learning_rate": 0.00018331599611806366,
      "loss": 0.7888,
      "mean_token_accuracy": 0.7903819754719734,
      "num_tokens": 14681393.0,
      "step": 1090
    },
    {
      "epoch": 0.6699147381242387,
      "grad_norm": 0.5369830131530762,
      "learning_rate": 0.00018297119723333877,
      "loss": 0.765,
      "mean_token_accuracy": 0.7950262635946274,
      "num_tokens": 14814565.0,
      "step": 1100
    },
    {
      "epoch": 0.6760048721071864,
      "grad_norm": 0.5289803743362427,
      "learning_rate": 0.00018262324365372846,
      "loss": 0.7496,
      "mean_token_accuracy": 0.8032818242907525,
      "num_tokens": 14954351.0,
      "step": 1110
    },
    {
      "epoch": 0.682095006090134,
      "grad_norm": 0.5440439581871033,
      "learning_rate": 0.0001822721504219814,
      "loss": 0.7432,
      "mean_token_accuracy": 0.799126236140728,
      "num_tokens": 15094879.0,
      "step": 1120
    },
    {
      "epoch": 0.6881851400730816,
      "grad_norm": 0.46225935220718384,
      "learning_rate": 0.00018191793271657978,
      "loss": 0.7513,
      "mean_token_accuracy": 0.8022688791155815,
      "num_tokens": 15234906.0,
      "step": 1130
    },
    {
      "epoch": 0.6942752740560292,
      "grad_norm": 0.5592020750045776,
      "learning_rate": 0.0001815606058510833,
      "loss": 0.7583,
      "mean_token_accuracy": 0.7984497547149658,
      "num_tokens": 15373526.0,
      "step": 1140
    },
    {
      "epoch": 0.7003654080389768,
      "grad_norm": 0.525090217590332,
      "learning_rate": 0.00018120018527346702,
      "loss": 0.7254,
      "mean_token_accuracy": 0.8070619881153107,
      "num_tokens": 15516264.0,
      "step": 1150
    },
    {
      "epoch": 0.7064555420219245,
      "grad_norm": 0.5380759239196777,
      "learning_rate": 0.00018083668656545355,
      "loss": 0.8041,
      "mean_token_accuracy": 0.7866759791970253,
      "num_tokens": 15640444.0,
      "step": 1160
    },
    {
      "epoch": 0.7125456760048721,
      "grad_norm": 0.47815701365470886,
      "learning_rate": 0.00018047012544183938,
      "loss": 0.7604,
      "mean_token_accuracy": 0.796156468987465,
      "num_tokens": 15778070.0,
      "step": 1170
    },
    {
      "epoch": 0.7186358099878197,
      "grad_norm": 0.5380450487136841,
      "learning_rate": 0.00018010051774981553,
      "loss": 0.8135,
      "mean_token_accuracy": 0.7842124432325364,
      "num_tokens": 15899739.0,
      "step": 1180
    },
    {
      "epoch": 0.7247259439707674,
      "grad_norm": 0.5047502517700195,
      "learning_rate": 0.00017972787946828246,
      "loss": 0.7642,
      "mean_token_accuracy": 0.7989341139793396,
      "num_tokens": 16035805.0,
      "step": 1190
    },
    {
      "epoch": 0.730816077953715,
      "grad_norm": 0.5440967679023743,
      "learning_rate": 0.00017935222670715918,
      "loss": 0.735,
      "mean_token_accuracy": 0.8048294603824615,
      "num_tokens": 16172541.0,
      "step": 1200
    },
    {
      "epoch": 0.7369062119366626,
      "grad_norm": 0.4766077399253845,
      "learning_rate": 0.000178973575706687,
      "loss": 0.805,
      "mean_token_accuracy": 0.7871790423989296,
      "num_tokens": 16296988.0,
      "step": 1210
    },
    {
      "epoch": 0.7429963459196103,
      "grad_norm": 0.4153214991092682,
      "learning_rate": 0.00017859194283672704,
      "loss": 0.7635,
      "mean_token_accuracy": 0.7964595645666123,
      "num_tokens": 16432022.0,
      "step": 1220
    },
    {
      "epoch": 0.7490864799025578,
      "grad_norm": 0.4698518216609955,
      "learning_rate": 0.00017820734459605302,
      "loss": 0.7397,
      "mean_token_accuracy": 0.8046972885727882,
      "num_tokens": 16572880.0,
      "step": 1230
    },
    {
      "epoch": 0.7551766138855055,
      "grad_norm": 0.46101540327072144,
      "learning_rate": 0.00017781979761163756,
      "loss": 0.7174,
      "mean_token_accuracy": 0.8066875368356705,
      "num_tokens": 16714419.0,
      "step": 1240
    },
    {
      "epoch": 0.761266747868453,
      "grad_norm": 0.5313341021537781,
      "learning_rate": 0.00017742931863793358,
      "loss": 0.7797,
      "mean_token_accuracy": 0.7911526098847389,
      "num_tokens": 16838285.0,
      "step": 1250
    },
    {
      "epoch": 0.7673568818514007,
      "grad_norm": 0.4627362787723541,
      "learning_rate": 0.00017703592455614998,
      "loss": 0.7626,
      "mean_token_accuracy": 0.7970306649804115,
      "num_tokens": 16976065.0,
      "step": 1260
    },
    {
      "epoch": 0.7734470158343484,
      "grad_norm": 0.5429073572158813,
      "learning_rate": 0.00017663963237352177,
      "loss": 0.7398,
      "mean_token_accuracy": 0.8005403786897659,
      "num_tokens": 17112901.0,
      "step": 1270
    },
    {
      "epoch": 0.7795371498172959,
      "grad_norm": 0.6781270503997803,
      "learning_rate": 0.00017624045922257471,
      "loss": 0.7607,
      "mean_token_accuracy": 0.7946217939257622,
      "num_tokens": 17245480.0,
      "step": 1280
    },
    {
      "epoch": 0.7856272838002436,
      "grad_norm": 0.5227305293083191,
      "learning_rate": 0.00017583842236038483,
      "loss": 0.7217,
      "mean_token_accuracy": 0.8064659267663956,
      "num_tokens": 17387171.0,
      "step": 1290
    },
    {
      "epoch": 0.7917174177831913,
      "grad_norm": 0.49253156781196594,
      "learning_rate": 0.0001754335391678323,
      "loss": 0.7652,
      "mean_token_accuracy": 0.7960015773773194,
      "num_tokens": 17521164.0,
      "step": 1300
    },
    {
      "epoch": 0.7978075517661388,
      "grad_norm": 0.5103631615638733,
      "learning_rate": 0.00017502582714884997,
      "loss": 0.7435,
      "mean_token_accuracy": 0.7995276898145676,
      "num_tokens": 17657818.0,
      "step": 1310
    },
    {
      "epoch": 0.8038976857490865,
      "grad_norm": 0.5531247854232788,
      "learning_rate": 0.00017461530392966665,
      "loss": 0.7986,
      "mean_token_accuracy": 0.7892467245459557,
      "num_tokens": 17784361.0,
      "step": 1320
    },
    {
      "epoch": 0.8099878197320342,
      "grad_norm": 0.4574586749076843,
      "learning_rate": 0.00017420198725804517,
      "loss": 0.6889,
      "mean_token_accuracy": 0.8135112956166267,
      "num_tokens": 17929664.0,
      "step": 1330
    },
    {
      "epoch": 0.8160779537149817,
      "grad_norm": 0.4734383225440979,
      "learning_rate": 0.00017378589500251498,
      "loss": 0.7308,
      "mean_token_accuracy": 0.8029947131872177,
      "num_tokens": 18071182.0,
      "step": 1340
    },
    {
      "epoch": 0.8221680876979294,
      "grad_norm": 0.5192279815673828,
      "learning_rate": 0.00017336704515159986,
      "loss": 0.7444,
      "mean_token_accuracy": 0.8012512847781181,
      "num_tokens": 18211136.0,
      "step": 1350
    },
    {
      "epoch": 0.8282582216808769,
      "grad_norm": 0.5378620624542236,
      "learning_rate": 0.00017294545581303996,
      "loss": 0.7459,
      "mean_token_accuracy": 0.7981989249587059,
      "num_tokens": 18340645.0,
      "step": 1360
    },
    {
      "epoch": 0.8343483556638246,
      "grad_norm": 0.4879571497440338,
      "learning_rate": 0.00017252114521300918,
      "loss": 0.7877,
      "mean_token_accuracy": 0.7891893342137337,
      "num_tokens": 18465733.0,
      "step": 1370
    },
    {
      "epoch": 0.8404384896467723,
      "grad_norm": 0.5297388434410095,
      "learning_rate": 0.00017209413169532717,
      "loss": 0.7586,
      "mean_token_accuracy": 0.797142505645752,
      "num_tokens": 18598979.0,
      "step": 1380
    },
    {
      "epoch": 0.8465286236297198,
      "grad_norm": 0.5308396220207214,
      "learning_rate": 0.00017166443372066618,
      "loss": 0.7387,
      "mean_token_accuracy": 0.80123979896307,
      "num_tokens": 18735919.0,
      "step": 1390
    },
    {
      "epoch": 0.8526187576126675,
      "grad_norm": 0.49988579750061035,
      "learning_rate": 0.0001712320698657532,
      "loss": 0.7425,
      "mean_token_accuracy": 0.7996803268790245,
      "num_tokens": 18870877.0,
      "step": 1400
    },
    {
      "epoch": 0.8587088915956151,
      "grad_norm": 0.5971361994743347,
      "learning_rate": 0.0001707970588225665,
      "loss": 0.7691,
      "mean_token_accuracy": 0.7922965154051781,
      "num_tokens": 19000943.0,
      "step": 1410
    },
    {
      "epoch": 0.8647990255785627,
      "grad_norm": 0.5141698718070984,
      "learning_rate": 0.00017035941939752802,
      "loss": 0.7203,
      "mean_token_accuracy": 0.8036229625344277,
      "num_tokens": 19135039.0,
      "step": 1420
    },
    {
      "epoch": 0.8708891595615104,
      "grad_norm": 0.4647749066352844,
      "learning_rate": 0.0001699191705106898,
      "loss": 0.7136,
      "mean_token_accuracy": 0.8064323276281357,
      "num_tokens": 19274069.0,
      "step": 1430
    },
    {
      "epoch": 0.876979293544458,
      "grad_norm": 0.5511934161186218,
      "learning_rate": 0.00016947633119491633,
      "loss": 0.7455,
      "mean_token_accuracy": 0.7985599264502525,
      "num_tokens": 19409679.0,
      "step": 1440
    },
    {
      "epoch": 0.8830694275274056,
      "grad_norm": 0.4936945140361786,
      "learning_rate": 0.00016903092059506182,
      "loss": 0.7087,
      "mean_token_accuracy": 0.806523185968399,
      "num_tokens": 19547419.0,
      "step": 1450
    },
    {
      "epoch": 0.8891595615103532,
      "grad_norm": 0.5227787494659424,
      "learning_rate": 0.00016858295796714213,
      "loss": 0.7739,
      "mean_token_accuracy": 0.7941467314958572,
      "num_tokens": 19674455.0,
      "step": 1460
    },
    {
      "epoch": 0.8952496954933008,
      "grad_norm": 0.5046219825744629,
      "learning_rate": 0.00016813246267750282,
      "loss": 0.7361,
      "mean_token_accuracy": 0.8008369222283364,
      "num_tokens": 19809861.0,
      "step": 1470
    },
    {
      "epoch": 0.9013398294762485,
      "grad_norm": 0.4827081263065338,
      "learning_rate": 0.00016767945420198142,
      "loss": 0.7464,
      "mean_token_accuracy": 0.7986427888274192,
      "num_tokens": 19940696.0,
      "step": 1480
    },
    {
      "epoch": 0.9074299634591961,
      "grad_norm": 0.4970889687538147,
      "learning_rate": 0.00016722395212506567,
      "loss": 0.7528,
      "mean_token_accuracy": 0.7965970665216446,
      "num_tokens": 20070686.0,
      "step": 1490
    },
    {
      "epoch": 0.9135200974421437,
      "grad_norm": 0.44478070735931396,
      "learning_rate": 0.00016676597613904693,
      "loss": 0.7185,
      "mean_token_accuracy": 0.8081388726830483,
      "num_tokens": 20210260.0,
      "step": 1500
    },
    {
      "epoch": 0.9196102314250914,
      "grad_norm": 0.506136417388916,
      "learning_rate": 0.00016630554604316866,
      "loss": 0.7395,
      "mean_token_accuracy": 0.8003876298666001,
      "num_tokens": 20346235.0,
      "step": 1510
    },
    {
      "epoch": 0.925700365408039,
      "grad_norm": 0.500946044921875,
      "learning_rate": 0.00016584268174277053,
      "loss": 0.6889,
      "mean_token_accuracy": 0.8124501362442971,
      "num_tokens": 20481248.0,
      "step": 1520
    },
    {
      "epoch": 0.9317904993909866,
      "grad_norm": 0.48528990149497986,
      "learning_rate": 0.00016537740324842795,
      "loss": 0.7227,
      "mean_token_accuracy": 0.8041250064969063,
      "num_tokens": 20613531.0,
      "step": 1530
    },
    {
      "epoch": 0.9378806333739342,
      "grad_norm": 0.5070951581001282,
      "learning_rate": 0.00016490973067508674,
      "loss": 0.7091,
      "mean_token_accuracy": 0.8082544595003128,
      "num_tokens": 20750784.0,
      "step": 1540
    },
    {
      "epoch": 0.9439707673568819,
      "grad_norm": 0.5583120584487915,
      "learning_rate": 0.0001644396842411939,
      "loss": 0.7405,
      "mean_token_accuracy": 0.7992320343852043,
      "num_tokens": 20883646.0,
      "step": 1550
    },
    {
      "epoch": 0.9500609013398295,
      "grad_norm": 0.5099635124206543,
      "learning_rate": 0.00016396728426782312,
      "loss": 0.7103,
      "mean_token_accuracy": 0.8091216519474983,
      "num_tokens": 21025143.0,
      "step": 1560
    },
    {
      "epoch": 0.9561510353227771,
      "grad_norm": 0.5777808427810669,
      "learning_rate": 0.00016349255117779652,
      "loss": 0.7245,
      "mean_token_accuracy": 0.8023119494318962,
      "num_tokens": 21160014.0,
      "step": 1570
    },
    {
      "epoch": 0.9622411693057247,
      "grad_norm": 0.5206162333488464,
      "learning_rate": 0.0001630155054948016,
      "loss": 0.7185,
      "mean_token_accuracy": 0.8069521963596344,
      "num_tokens": 21299094.0,
      "step": 1580
    },
    {
      "epoch": 0.9683313032886723,
      "grad_norm": 0.5763202905654907,
      "learning_rate": 0.00016253616784250415,
      "loss": 0.7677,
      "mean_token_accuracy": 0.7927820891141891,
      "num_tokens": 21429252.0,
      "step": 1590
    },
    {
      "epoch": 0.97442143727162,
      "grad_norm": 0.5068426728248596,
      "learning_rate": 0.00016205455894365627,
      "loss": 0.7673,
      "mean_token_accuracy": 0.794715291261673,
      "num_tokens": 21556200.0,
      "step": 1600
    },
    {
      "epoch": 0.9805115712545676,
      "grad_norm": 0.46094459295272827,
      "learning_rate": 0.0001615706996192009,
      "loss": 0.771,
      "mean_token_accuracy": 0.7921045809984207,
      "num_tokens": 21681524.0,
      "step": 1610
    },
    {
      "epoch": 0.9866017052375152,
      "grad_norm": 0.5063546299934387,
      "learning_rate": 0.00016108461078737148,
      "loss": 0.7383,
      "mean_token_accuracy": 0.800596435368061,
      "num_tokens": 21814109.0,
      "step": 1620
    },
    {
      "epoch": 0.9926918392204629,
      "grad_norm": 0.5418652296066284,
      "learning_rate": 0.0001605963134627876,
      "loss": 0.7431,
      "mean_token_accuracy": 0.7994748756289483,
      "num_tokens": 21947346.0,
      "step": 1630
    },
    {
      "epoch": 0.9987819732034104,
      "grad_norm": 0.6195595264434814,
      "learning_rate": 0.0001601058287555465,
      "loss": 0.7294,
      "mean_token_accuracy": 0.8030684441328049,
      "num_tokens": 22081340.0,
      "step": 1640
    },
    {
      "epoch": 1.004872107186358,
      "grad_norm": 0.5930359363555908,
      "learning_rate": 0.00015961317787031054,
      "loss": 0.7387,
      "mean_token_accuracy": 0.8013696864247322,
      "num_tokens": 22206441.0,
      "step": 1650
    },
    {
      "epoch": 1.0109622411693058,
      "grad_norm": 0.4926474094390869,
      "learning_rate": 0.00015911838210539038,
      "loss": 0.6743,
      "mean_token_accuracy": 0.8141208037734031,
      "num_tokens": 22344898.0,
      "step": 1660
    },
    {
      "epoch": 1.0170523751522533,
      "grad_norm": 0.5331000685691833,
      "learning_rate": 0.0001586214628518242,
      "loss": 0.7033,
      "mean_token_accuracy": 0.807385990023613,
      "num_tokens": 22483135.0,
      "step": 1670
    },
    {
      "epoch": 1.0231425091352009,
      "grad_norm": 0.5267267227172852,
      "learning_rate": 0.0001581224415924531,
      "loss": 0.6717,
      "mean_token_accuracy": 0.8178876608610153,
      "num_tokens": 22617934.0,
      "step": 1680
    },
    {
      "epoch": 1.0292326431181487,
      "grad_norm": 0.5864041447639465,
      "learning_rate": 0.00015762133990099205,
      "loss": 0.7421,
      "mean_token_accuracy": 0.7981289237737655,
      "num_tokens": 22745190.0,
      "step": 1690
    },
    {
      "epoch": 1.0353227771010962,
      "grad_norm": 0.45681944489479065,
      "learning_rate": 0.00015711817944109738,
      "loss": 0.6646,
      "mean_token_accuracy": 0.8146520599722862,
      "num_tokens": 22887536.0,
      "step": 1700
    },
    {
      "epoch": 1.0414129110840438,
      "grad_norm": 0.5522484183311462,
      "learning_rate": 0.00015661298196543042,
      "loss": 0.6889,
      "mean_token_accuracy": 0.8100781336426734,
      "num_tokens": 23017586.0,
      "step": 1710
    },
    {
      "epoch": 1.0475030450669915,
      "grad_norm": 0.5221629738807678,
      "learning_rate": 0.00015610576931471658,
      "loss": 0.6939,
      "mean_token_accuracy": 0.8114214852452278,
      "num_tokens": 23151737.0,
      "step": 1720
    },
    {
      "epoch": 1.053593179049939,
      "grad_norm": 0.5104020833969116,
      "learning_rate": 0.00015559656341680164,
      "loss": 0.716,
      "mean_token_accuracy": 0.8063826873898506,
      "num_tokens": 23280778.0,
      "step": 1730
    },
    {
      "epoch": 1.0596833130328867,
      "grad_norm": 0.5163984298706055,
      "learning_rate": 0.00015508538628570352,
      "loss": 0.7188,
      "mean_token_accuracy": 0.802527217566967,
      "num_tokens": 23410327.0,
      "step": 1740
    },
    {
      "epoch": 1.0657734470158344,
      "grad_norm": 0.5188373327255249,
      "learning_rate": 0.00015457226002066058,
      "loss": 0.6791,
      "mean_token_accuracy": 0.8127639785408973,
      "num_tokens": 23548616.0,
      "step": 1750
    },
    {
      "epoch": 1.071863580998782,
      "grad_norm": 0.5983869433403015,
      "learning_rate": 0.00015405720680517618,
      "loss": 0.6869,
      "mean_token_accuracy": 0.8110290810465812,
      "num_tokens": 23682446.0,
      "step": 1760
    },
    {
      "epoch": 1.0779537149817295,
      "grad_norm": 0.5919123291969299,
      "learning_rate": 0.00015354024890605985,
      "loss": 0.7419,
      "mean_token_accuracy": 0.7984233900904656,
      "num_tokens": 23806352.0,
      "step": 1770
    },
    {
      "epoch": 1.0840438489646773,
      "grad_norm": 0.4900698661804199,
      "learning_rate": 0.0001530214086724644,
      "loss": 0.6781,
      "mean_token_accuracy": 0.8152358055114746,
      "num_tokens": 23942964.0,
      "step": 1780
    },
    {
      "epoch": 1.0901339829476249,
      "grad_norm": 0.5409672856330872,
      "learning_rate": 0.00015250070853491986,
      "loss": 0.7157,
      "mean_token_accuracy": 0.803682966530323,
      "num_tokens": 24070937.0,
      "step": 1790
    },
    {
      "epoch": 1.0962241169305724,
      "grad_norm": 0.5581572651863098,
      "learning_rate": 0.0001519781710043638,
      "loss": 0.7261,
      "mean_token_accuracy": 0.8027503877878189,
      "num_tokens": 24200686.0,
      "step": 1800
    },
    {
      "epoch": 1.1023142509135202,
      "grad_norm": 0.503963053226471,
      "learning_rate": 0.0001514538186711679,
      "loss": 0.7125,
      "mean_token_accuracy": 0.8042754918336869,
      "num_tokens": 24329983.0,
      "step": 1810
    },
    {
      "epoch": 1.1084043848964678,
      "grad_norm": 0.6159723997116089,
      "learning_rate": 0.00015092767420416168,
      "loss": 0.6873,
      "mean_token_accuracy": 0.8115814313292503,
      "num_tokens": 24465292.0,
      "step": 1820
    },
    {
      "epoch": 1.1144945188794153,
      "grad_norm": 0.518172562122345,
      "learning_rate": 0.00015039976034965214,
      "loss": 0.6805,
      "mean_token_accuracy": 0.8113815248012543,
      "num_tokens": 24599980.0,
      "step": 1830
    },
    {
      "epoch": 1.1205846528623629,
      "grad_norm": 0.5381601452827454,
      "learning_rate": 0.0001498700999304407,
      "loss": 0.6542,
      "mean_token_accuracy": 0.8188014090061188,
      "num_tokens": 24746703.0,
      "step": 1840
    },
    {
      "epoch": 1.1266747868453106,
      "grad_norm": 0.5001223683357239,
      "learning_rate": 0.00014933871584483615,
      "loss": 0.7255,
      "mean_token_accuracy": 0.8022593036293983,
      "num_tokens": 24877604.0,
      "step": 1850
    },
    {
      "epoch": 1.1327649208282582,
      "grad_norm": 0.5812251567840576,
      "learning_rate": 0.00014880563106566512,
      "loss": 0.6638,
      "mean_token_accuracy": 0.8161928996443748,
      "num_tokens": 25023049.0,
      "step": 1860
    },
    {
      "epoch": 1.1388550548112057,
      "grad_norm": 0.5384249091148376,
      "learning_rate": 0.0001482708686392786,
      "loss": 0.6623,
      "mean_token_accuracy": 0.8167443484067917,
      "num_tokens": 25162124.0,
      "step": 1870
    },
    {
      "epoch": 1.1449451887941535,
      "grad_norm": 0.5310192108154297,
      "learning_rate": 0.00014773445168455576,
      "loss": 0.7074,
      "mean_token_accuracy": 0.8042578861117363,
      "num_tokens": 25293569.0,
      "step": 1880
    },
    {
      "epoch": 1.151035322777101,
      "grad_norm": 0.6224446296691895,
      "learning_rate": 0.00014719640339190443,
      "loss": 0.7094,
      "mean_token_accuracy": 0.803679920732975,
      "num_tokens": 25422953.0,
      "step": 1890
    },
    {
      "epoch": 1.1571254567600486,
      "grad_norm": 0.5978189706802368,
      "learning_rate": 0.00014665674702225853,
      "loss": 0.6926,
      "mean_token_accuracy": 0.8080565810203553,
      "num_tokens": 25559091.0,
      "step": 1900
    },
    {
      "epoch": 1.1632155907429964,
      "grad_norm": 0.6134657263755798,
      "learning_rate": 0.00014611550590607245,
      "loss": 0.6716,
      "mean_token_accuracy": 0.8152063637971878,
      "num_tokens": 25698134.0,
      "step": 1910
    },
    {
      "epoch": 1.169305724725944,
      "grad_norm": 0.5075950026512146,
      "learning_rate": 0.00014557270344231246,
      "loss": 0.6613,
      "mean_token_accuracy": 0.8169043198227882,
      "num_tokens": 25835159.0,
      "step": 1920
    },
    {
      "epoch": 1.1753958587088915,
      "grad_norm": 0.5035059452056885,
      "learning_rate": 0.00014502836309744508,
      "loss": 0.6903,
      "mean_token_accuracy": 0.8096718549728393,
      "num_tokens": 25970600.0,
      "step": 1930
    },
    {
      "epoch": 1.1814859926918393,
      "grad_norm": 0.583890438079834,
      "learning_rate": 0.00014448250840442254,
      "loss": 0.6662,
      "mean_token_accuracy": 0.8157578155398368,
      "num_tokens": 26106658.0,
      "step": 1940
    },
    {
      "epoch": 1.1875761266747868,
      "grad_norm": 0.5089572668075562,
      "learning_rate": 0.00014393516296166552,
      "loss": 0.7085,
      "mean_token_accuracy": 0.8082539036870002,
      "num_tokens": 26238847.0,
      "step": 1950
    },
    {
      "epoch": 1.1936662606577344,
      "grad_norm": 0.4495029151439667,
      "learning_rate": 0.00014338635043204288,
      "loss": 0.7085,
      "mean_token_accuracy": 0.8075269401073456,
      "num_tokens": 26366417.0,
      "step": 1960
    },
    {
      "epoch": 1.1997563946406822,
      "grad_norm": 0.6390108466148376,
      "learning_rate": 0.00014283609454184855,
      "loss": 0.6935,
      "mean_token_accuracy": 0.8099950149655342,
      "num_tokens": 26498101.0,
      "step": 1970
    },
    {
      "epoch": 1.2058465286236297,
      "grad_norm": 0.5687986016273499,
      "learning_rate": 0.00014228441907977607,
      "loss": 0.7027,
      "mean_token_accuracy": 0.8083449766039849,
      "num_tokens": 26628513.0,
      "step": 1980
    },
    {
      "epoch": 1.2119366626065773,
      "grad_norm": 0.487954318523407,
      "learning_rate": 0.00014173134789588994,
      "loss": 0.6731,
      "mean_token_accuracy": 0.8129799589514732,
      "num_tokens": 26761671.0,
      "step": 1990
    },
    {
      "epoch": 1.218026796589525,
      "grad_norm": 0.5641826391220093,
      "learning_rate": 0.00014117690490059447,
      "loss": 0.6949,
      "mean_token_accuracy": 0.8118783175945282,
      "num_tokens": 26894870.0,
      "step": 2000
    },
    {
      "epoch": 1.2241169305724726,
      "grad_norm": 0.5209829211235046,
      "learning_rate": 0.00014062111406360034,
      "loss": 0.6742,
      "mean_token_accuracy": 0.816123254597187,
      "num_tokens": 27027902.0,
      "step": 2010
    },
    {
      "epoch": 1.2302070645554202,
      "grad_norm": 0.5218231678009033,
      "learning_rate": 0.00014006399941288812,
      "loss": 0.703,
      "mean_token_accuracy": 0.805295330286026,
      "num_tokens": 27157882.0,
      "step": 2020
    },
    {
      "epoch": 1.236297198538368,
      "grad_norm": 0.48154470324516296,
      "learning_rate": 0.00013950558503366957,
      "loss": 0.6844,
      "mean_token_accuracy": 0.811684039235115,
      "num_tokens": 27290994.0,
      "step": 2030
    },
    {
      "epoch": 1.2423873325213155,
      "grad_norm": 0.5417695045471191,
      "learning_rate": 0.00013894589506734643,
      "loss": 0.7253,
      "mean_token_accuracy": 0.8018206775188446,
      "num_tokens": 27420715.0,
      "step": 2040
    },
    {
      "epoch": 1.248477466504263,
      "grad_norm": 0.5282937288284302,
      "learning_rate": 0.00013838495371046671,
      "loss": 0.682,
      "mean_token_accuracy": 0.8128980100154877,
      "num_tokens": 27552040.0,
      "step": 2050
    },
    {
      "epoch": 1.2545676004872108,
      "grad_norm": 0.5213696360588074,
      "learning_rate": 0.0001378227852136785,
      "loss": 0.6728,
      "mean_token_accuracy": 0.8128269612789154,
      "num_tokens": 27686922.0,
      "step": 2060
    },
    {
      "epoch": 1.2606577344701584,
      "grad_norm": 0.4823834300041199,
      "learning_rate": 0.00013725941388068174,
      "loss": 0.6626,
      "mean_token_accuracy": 0.8177949145436287,
      "num_tokens": 27825036.0,
      "step": 2070
    },
    {
      "epoch": 1.266747868453106,
      "grad_norm": 0.6199477314949036,
      "learning_rate": 0.0001366948640671775,
      "loss": 0.686,
      "mean_token_accuracy": 0.8107037082314491,
      "num_tokens": 27961614.0,
      "step": 2080
    },
    {
      "epoch": 1.2728380024360537,
      "grad_norm": 0.4916837513446808,
      "learning_rate": 0.00013612916017981488,
      "loss": 0.6738,
      "mean_token_accuracy": 0.8149923622608185,
      "num_tokens": 28099524.0,
      "step": 2090
    },
    {
      "epoch": 1.2789281364190013,
      "grad_norm": 0.6001724600791931,
      "learning_rate": 0.00013556232667513607,
      "loss": 0.6637,
      "mean_token_accuracy": 0.8173324480652809,
      "num_tokens": 28237055.0,
      "step": 2100
    },
    {
      "epoch": 1.2850182704019488,
      "grad_norm": 0.5887413620948792,
      "learning_rate": 0.00013499438805851882,
      "loss": 0.6744,
      "mean_token_accuracy": 0.8149967223405838,
      "num_tokens": 28370538.0,
      "step": 2110
    },
    {
      "epoch": 1.2911084043848966,
      "grad_norm": 0.6208155751228333,
      "learning_rate": 0.00013442536888311733,
      "loss": 0.6973,
      "mean_token_accuracy": 0.8103871151804924,
      "num_tokens": 28499232.0,
      "step": 2120
    },
    {
      "epoch": 1.2971985383678442,
      "grad_norm": 0.5026904344558716,
      "learning_rate": 0.0001338552937488003,
      "loss": 0.6739,
      "mean_token_accuracy": 0.8153023451566697,
      "num_tokens": 28633993.0,
      "step": 2130
    },
    {
      "epoch": 1.3032886723507917,
      "grad_norm": 0.5218458771705627,
      "learning_rate": 0.00013328418730108795,
      "loss": 0.6619,
      "mean_token_accuracy": 0.8166303977370262,
      "num_tokens": 28774139.0,
      "step": 2140
    },
    {
      "epoch": 1.3093788063337393,
      "grad_norm": 0.519872784614563,
      "learning_rate": 0.00013271207423008622,
      "loss": 0.6804,
      "mean_token_accuracy": 0.8150519266724586,
      "num_tokens": 28910109.0,
      "step": 2150
    },
    {
      "epoch": 1.315468940316687,
      "grad_norm": 0.5219667553901672,
      "learning_rate": 0.00013213897926941942,
      "loss": 0.6682,
      "mean_token_accuracy": 0.8166522830724716,
      "num_tokens": 29045967.0,
      "step": 2160
    },
    {
      "epoch": 1.3215590742996346,
      "grad_norm": 0.5744656920433044,
      "learning_rate": 0.000131564927195161,
      "loss": 0.6772,
      "mean_token_accuracy": 0.8149690836668014,
      "num_tokens": 29180769.0,
      "step": 2170
    },
    {
      "epoch": 1.3276492082825821,
      "grad_norm": 0.5673508048057556,
      "learning_rate": 0.00013098994282476236,
      "loss": 0.6841,
      "mean_token_accuracy": 0.812624742090702,
      "num_tokens": 29313512.0,
      "step": 2180
    },
    {
      "epoch": 1.3337393422655297,
      "grad_norm": 0.5187074542045593,
      "learning_rate": 0.00013041405101598,
      "loss": 0.6281,
      "mean_token_accuracy": 0.8221091449260711,
      "num_tokens": 29454589.0,
      "step": 2190
    },
    {
      "epoch": 1.3398294762484775,
      "grad_norm": 0.5621201992034912,
      "learning_rate": 0.00012983727666580086,
      "loss": 0.6755,
      "mean_token_accuracy": 0.8157430678606034,
      "num_tokens": 29589968.0,
      "step": 2200
    },
    {
      "epoch": 1.345919610231425,
      "grad_norm": 0.579699695110321,
      "learning_rate": 0.00012925964470936598,
      "loss": 0.6859,
      "mean_token_accuracy": 0.8122102931141854,
      "num_tokens": 29720188.0,
      "step": 2210
    },
    {
      "epoch": 1.3520097442143726,
      "grad_norm": 0.6406823992729187,
      "learning_rate": 0.00012868118011889236,
      "loss": 0.684,
      "mean_token_accuracy": 0.8107294023036957,
      "num_tokens": 29848418.0,
      "step": 2220
    },
    {
      "epoch": 1.3580998781973204,
      "grad_norm": 0.4707708954811096,
      "learning_rate": 0.00012810190790259367,
      "loss": 0.6607,
      "mean_token_accuracy": 0.8182852879166603,
      "num_tokens": 29988202.0,
      "step": 2230
    },
    {
      "epoch": 1.364190012180268,
      "grad_norm": 0.6458183526992798,
      "learning_rate": 0.00012752185310359874,
      "loss": 0.6935,
      "mean_token_accuracy": 0.8089477211236954,
      "num_tokens": 30119777.0,
      "step": 2240
    },
    {
      "epoch": 1.3702801461632155,
      "grad_norm": 0.4278848469257355,
      "learning_rate": 0.00012694104079886918,
      "loss": 0.6565,
      "mean_token_accuracy": 0.8185079246759415,
      "num_tokens": 30256776.0,
      "step": 2250
    },
    {
      "epoch": 1.3763702801461632,
      "grad_norm": 0.5647698044776917,
      "learning_rate": 0.00012635949609811505,
      "loss": 0.6636,
      "mean_token_accuracy": 0.8155051723122597,
      "num_tokens": 30395629.0,
      "step": 2260
    },
    {
      "epoch": 1.3824604141291108,
      "grad_norm": 0.43498411774635315,
      "learning_rate": 0.00012577724414270937,
      "loss": 0.689,
      "mean_token_accuracy": 0.8125654354691505,
      "num_tokens": 30532805.0,
      "step": 2270
    },
    {
      "epoch": 1.3885505481120584,
      "grad_norm": 0.5296844244003296,
      "learning_rate": 0.00012519431010460136,
      "loss": 0.6854,
      "mean_token_accuracy": 0.8122918352484703,
      "num_tokens": 30664642.0,
      "step": 2280
    },
    {
      "epoch": 1.3946406820950061,
      "grad_norm": 0.44080430269241333,
      "learning_rate": 0.000124610719185228,
      "loss": 0.6405,
      "mean_token_accuracy": 0.8192834481596947,
      "num_tokens": 30805370.0,
      "step": 2290
    },
    {
      "epoch": 1.4007308160779537,
      "grad_norm": 0.5946847796440125,
      "learning_rate": 0.00012402649661442453,
      "loss": 0.7025,
      "mean_token_accuracy": 0.8085126876831055,
      "num_tokens": 30936385.0,
      "step": 2300
    },
    {
      "epoch": 1.4068209500609012,
      "grad_norm": 0.6572047472000122,
      "learning_rate": 0.0001234416676493339,
      "loss": 0.709,
      "mean_token_accuracy": 0.8046677514910698,
      "num_tokens": 31067615.0,
      "step": 2310
    },
    {
      "epoch": 1.412911084043849,
      "grad_norm": 0.4797047972679138,
      "learning_rate": 0.0001228562575733147,
      "loss": 0.6675,
      "mean_token_accuracy": 0.8157136350870132,
      "num_tokens": 31200044.0,
      "step": 2320
    },
    {
      "epoch": 1.4190012180267966,
      "grad_norm": 0.5451430082321167,
      "learning_rate": 0.0001222702916948481,
      "loss": 0.6746,
      "mean_token_accuracy": 0.8092615008354187,
      "num_tokens": 31334451.0,
      "step": 2330
    },
    {
      "epoch": 1.4250913520097441,
      "grad_norm": 0.5049906969070435,
      "learning_rate": 0.00012168379534644371,
      "loss": 0.6515,
      "mean_token_accuracy": 0.8203717589378356,
      "num_tokens": 31472218.0,
      "step": 2340
    },
    {
      "epoch": 1.431181485992692,
      "grad_norm": 0.6531693935394287,
      "learning_rate": 0.00012109679388354462,
      "loss": 0.6778,
      "mean_token_accuracy": 0.8134923160076142,
      "num_tokens": 31605853.0,
      "step": 2350
    },
    {
      "epoch": 1.4372716199756395,
      "grad_norm": 0.5340039730072021,
      "learning_rate": 0.00012050931268343089,
      "loss": 0.6628,
      "mean_token_accuracy": 0.8176047816872597,
      "num_tokens": 31741034.0,
      "step": 2360
    },
    {
      "epoch": 1.443361753958587,
      "grad_norm": 0.4518280625343323,
      "learning_rate": 0.00011992137714412266,
      "loss": 0.6407,
      "mean_token_accuracy": 0.8207336485385894,
      "num_tokens": 31878661.0,
      "step": 2370
    },
    {
      "epoch": 1.4494518879415348,
      "grad_norm": 0.5232827067375183,
      "learning_rate": 0.00011933301268328212,
      "loss": 0.6742,
      "mean_token_accuracy": 0.8158077761530876,
      "num_tokens": 32016524.0,
      "step": 2380
    },
    {
      "epoch": 1.4555420219244823,
      "grad_norm": 0.5181542634963989,
      "learning_rate": 0.00011874424473711457,
      "loss": 0.699,
      "mean_token_accuracy": 0.8078866004943848,
      "num_tokens": 32146820.0,
      "step": 2390
    },
    {
      "epoch": 1.46163215590743,
      "grad_norm": 0.5801041126251221,
      "learning_rate": 0.00011815509875926883,
      "loss": 0.6572,
      "mean_token_accuracy": 0.8183338135480881,
      "num_tokens": 32285928.0,
      "step": 2400
    },
    {
      "epoch": 1.4677222898903777,
      "grad_norm": 0.5347133874893188,
      "learning_rate": 0.00011756560021973679,
      "loss": 0.6738,
      "mean_token_accuracy": 0.8143690213561058,
      "num_tokens": 32416470.0,
      "step": 2410
    },
    {
      "epoch": 1.4738124238733252,
      "grad_norm": 0.4945615231990814,
      "learning_rate": 0.0001169757746037524,
      "loss": 0.6505,
      "mean_token_accuracy": 0.8196728631854058,
      "num_tokens": 32553798.0,
      "step": 2420
    },
    {
      "epoch": 1.4799025578562728,
      "grad_norm": 0.5072743892669678,
      "learning_rate": 0.00011638564741068965,
      "loss": 0.625,
      "mean_token_accuracy": 0.826240348815918,
      "num_tokens": 32692511.0,
      "step": 2430
    },
    {
      "epoch": 1.4859926918392206,
      "grad_norm": 0.5887538194656372,
      "learning_rate": 0.00011579524415296043,
      "loss": 0.6904,
      "mean_token_accuracy": 0.8112018033862114,
      "num_tokens": 32818836.0,
      "step": 2440
    },
    {
      "epoch": 1.4920828258221681,
      "grad_norm": 0.5464449524879456,
      "learning_rate": 0.00011520459035491142,
      "loss": 0.6553,
      "mean_token_accuracy": 0.8198345899581909,
      "num_tokens": 32957967.0,
      "step": 2450
    },
    {
      "epoch": 1.4981729598051157,
      "grad_norm": 0.5787419676780701,
      "learning_rate": 0.00011461371155172071,
      "loss": 0.663,
      "mean_token_accuracy": 0.8155046373605728,
      "num_tokens": 33094241.0,
      "step": 2460
    },
    {
      "epoch": 1.5042630937880634,
      "grad_norm": 0.5159268975257874,
      "learning_rate": 0.00011402263328829384,
      "loss": 0.6792,
      "mean_token_accuracy": 0.8127613604068756,
      "num_tokens": 33225474.0,
      "step": 2470
    },
    {
      "epoch": 1.510353227771011,
      "grad_norm": 0.5665333867073059,
      "learning_rate": 0.00011343138111815939,
      "loss": 0.6265,
      "mean_token_accuracy": 0.8276977241039276,
      "num_tokens": 33368246.0,
      "step": 2480
    },
    {
      "epoch": 1.5164433617539586,
      "grad_norm": 0.6272276639938354,
      "learning_rate": 0.00011283998060236421,
      "loss": 0.6734,
      "mean_token_accuracy": 0.816029068827629,
      "num_tokens": 33503967.0,
      "step": 2490
    },
    {
      "epoch": 1.5225334957369063,
      "grad_norm": 0.5275886654853821,
      "learning_rate": 0.0001122484573083686,
      "loss": 0.6457,
      "mean_token_accuracy": 0.8222623988986015,
      "num_tokens": 33641826.0,
      "step": 2500
    },
    {
      "epoch": 1.5286236297198539,
      "grad_norm": 0.5526687502861023,
      "learning_rate": 0.00011165683680894072,
      "loss": 0.6795,
      "mean_token_accuracy": 0.8127825185656548,
      "num_tokens": 33774185.0,
      "step": 2510
    },
    {
      "epoch": 1.5347137637028014,
      "grad_norm": 0.6226133704185486,
      "learning_rate": 0.00011106514468105111,
      "loss": 0.6684,
      "mean_token_accuracy": 0.815614765882492,
      "num_tokens": 33907116.0,
      "step": 2520
    },
    {
      "epoch": 1.5408038976857492,
      "grad_norm": 0.612832248210907,
      "learning_rate": 0.000110473406504767,
      "loss": 0.6287,
      "mean_token_accuracy": 0.8220825806260109,
      "num_tokens": 34048267.0,
      "step": 2530
    },
    {
      "epoch": 1.5468940316686965,
      "grad_norm": 0.6066681742668152,
      "learning_rate": 0.00010988164786214639,
      "loss": 0.6851,
      "mean_token_accuracy": 0.8115911707282066,
      "num_tokens": 34177555.0,
      "step": 2540
    },
    {
      "epoch": 1.5529841656516443,
      "grad_norm": 0.6376360058784485,
      "learning_rate": 0.00010928989433613204,
      "loss": 0.6921,
      "mean_token_accuracy": 0.8096534594893455,
      "num_tokens": 34308932.0,
      "step": 2550
    },
    {
      "epoch": 1.559074299634592,
      "grad_norm": 0.6083400249481201,
      "learning_rate": 0.00010869817150944546,
      "loss": 0.6575,
      "mean_token_accuracy": 0.8187816679477692,
      "num_tokens": 34443994.0,
      "step": 2560
    },
    {
      "epoch": 1.5651644336175394,
      "grad_norm": 0.6098156571388245,
      "learning_rate": 0.00010810650496348116,
      "loss": 0.6092,
      "mean_token_accuracy": 0.8285523638129234,
      "num_tokens": 34588403.0,
      "step": 2570
    },
    {
      "epoch": 1.5712545676004872,
      "grad_norm": 0.47795701026916504,
      "learning_rate": 0.00010751492027720027,
      "loss": 0.6423,
      "mean_token_accuracy": 0.8211737647652626,
      "num_tokens": 34730426.0,
      "step": 2580
    },
    {
      "epoch": 1.577344701583435,
      "grad_norm": 0.560787558555603,
      "learning_rate": 0.00010692344302602515,
      "loss": 0.6707,
      "mean_token_accuracy": 0.8134441033005715,
      "num_tokens": 34861708.0,
      "step": 2590
    },
    {
      "epoch": 1.5834348355663823,
      "grad_norm": 0.5722246766090393,
      "learning_rate": 0.00010633209878073343,
      "loss": 0.6533,
      "mean_token_accuracy": 0.8185199156403542,
      "num_tokens": 34997377.0,
      "step": 2600
    },
    {
      "epoch": 1.58952496954933,
      "grad_norm": 0.4941788613796234,
      "learning_rate": 0.00010574091310635263,
      "loss": 0.6487,
      "mean_token_accuracy": 0.8205527886748314,
      "num_tokens": 35133685.0,
      "step": 2610
    },
    {
      "epoch": 1.5956151035322779,
      "grad_norm": 0.575986921787262,
      "learning_rate": 0.00010514991156105493,
      "loss": 0.6615,
      "mean_token_accuracy": 0.8179458349943161,
      "num_tokens": 35270993.0,
      "step": 2620
    },
    {
      "epoch": 1.6017052375152252,
      "grad_norm": 0.5677866339683533,
      "learning_rate": 0.00010455911969505228,
      "loss": 0.6572,
      "mean_token_accuracy": 0.8155815675854683,
      "num_tokens": 35402062.0,
      "step": 2630
    },
    {
      "epoch": 1.607795371498173,
      "grad_norm": 0.6232825517654419,
      "learning_rate": 0.00010396856304949162,
      "loss": 0.6477,
      "mean_token_accuracy": 0.8209305629134178,
      "num_tokens": 35537394.0,
      "step": 2640
    },
    {
      "epoch": 1.6138855054811205,
      "grad_norm": 0.6252410411834717,
      "learning_rate": 0.00010337826715535102,
      "loss": 0.6819,
      "mean_token_accuracy": 0.8137489795684815,
      "num_tokens": 35669332.0,
      "step": 2650
    },
    {
      "epoch": 1.619975639464068,
      "grad_norm": 0.5850580334663391,
      "learning_rate": 0.0001027882575323356,
      "loss": 0.6831,
      "mean_token_accuracy": 0.8099577218294144,
      "num_tokens": 35799095.0,
      "step": 2660
    },
    {
      "epoch": 1.6260657734470159,
      "grad_norm": 0.5118699073791504,
      "learning_rate": 0.00010219855968777442,
      "loss": 0.681,
      "mean_token_accuracy": 0.8123177006840706,
      "num_tokens": 35928313.0,
      "step": 2670
    },
    {
      "epoch": 1.6321559074299634,
      "grad_norm": 0.5392698645591736,
      "learning_rate": 0.00010160919911551774,
      "loss": 0.6536,
      "mean_token_accuracy": 0.8185337752103805,
      "num_tokens": 36062033.0,
      "step": 2680
    },
    {
      "epoch": 1.638246041412911,
      "grad_norm": 0.5542203783988953,
      "learning_rate": 0.00010102020129483481,
      "loss": 0.6859,
      "mean_token_accuracy": 0.8107540607452393,
      "num_tokens": 36190194.0,
      "step": 2690
    },
    {
      "epoch": 1.6443361753958587,
      "grad_norm": 0.5962918996810913,
      "learning_rate": 0.0001004315916893124,
      "loss": 0.64,
      "mean_token_accuracy": 0.8226593688130379,
      "num_tokens": 36322437.0,
      "step": 2700
    },
    {
      "epoch": 1.6504263093788063,
      "grad_norm": 0.6391364932060242,
      "learning_rate": 9.984339574575394e-05,
      "loss": 0.6457,
      "mean_token_accuracy": 0.8250340327620507,
      "num_tokens": 36463231.0,
      "step": 2710
    },
    {
      "epoch": 1.6565164433617539,
      "grad_norm": 0.5798075795173645,
      "learning_rate": 9.92556388930794e-05,
      "loss": 0.6901,
      "mean_token_accuracy": 0.8104871213436127,
      "num_tokens": 36588963.0,
      "step": 2720
    },
    {
      "epoch": 1.6626065773447016,
      "grad_norm": 0.5375143885612488,
      "learning_rate": 9.866834654122597e-05,
      "loss": 0.6723,
      "mean_token_accuracy": 0.8132491707801819,
      "num_tokens": 36724295.0,
      "step": 2730
    },
    {
      "epoch": 1.6686967113276492,
      "grad_norm": 0.5556331276893616,
      "learning_rate": 9.808154408004942e-05,
      "loss": 0.6316,
      "mean_token_accuracy": 0.8221101492643357,
      "num_tokens": 36855978.0,
      "step": 2740
    },
    {
      "epoch": 1.6747868453105967,
      "grad_norm": 0.5330142974853516,
      "learning_rate": 9.749525687822674e-05,
      "loss": 0.6269,
      "mean_token_accuracy": 0.8239532545208931,
      "num_tokens": 36994164.0,
      "step": 2750
    },
    {
      "epoch": 1.6808769792935445,
      "grad_norm": 0.568084716796875,
      "learning_rate": 9.6909510282159e-05,
      "loss": 0.6568,
      "mean_token_accuracy": 0.8158794924616813,
      "num_tokens": 37130680.0,
      "step": 2760
    },
    {
      "epoch": 1.686967113276492,
      "grad_norm": 0.5072943568229675,
      "learning_rate": 9.632432961487585e-05,
      "loss": 0.6838,
      "mean_token_accuracy": 0.8121756613254547,
      "num_tokens": 37261462.0,
      "step": 2770
    },
    {
      "epoch": 1.6930572472594396,
      "grad_norm": 0.5469337701797485,
      "learning_rate": 9.573974017494069e-05,
      "loss": 0.6447,
      "mean_token_accuracy": 0.8220986798405647,
      "num_tokens": 37395606.0,
      "step": 2780
    },
    {
      "epoch": 1.6991473812423874,
      "grad_norm": 0.57918381690979,
      "learning_rate": 9.515576723535689e-05,
      "loss": 0.6217,
      "mean_token_accuracy": 0.822702020406723,
      "num_tokens": 37533585.0,
      "step": 2790
    },
    {
      "epoch": 1.705237515225335,
      "grad_norm": 0.6425563097000122,
      "learning_rate": 9.45724360424753e-05,
      "loss": 0.6435,
      "mean_token_accuracy": 0.8198476612567902,
      "num_tokens": 37672877.0,
      "step": 2800
    },
    {
      "epoch": 1.7113276492082825,
      "grad_norm": 0.5059729218482971,
      "learning_rate": 9.398977181490274e-05,
      "loss": 0.6579,
      "mean_token_accuracy": 0.8166012555360794,
      "num_tokens": 37809109.0,
      "step": 2810
    },
    {
      "epoch": 1.7174177831912303,
      "grad_norm": 0.5450888276100159,
      "learning_rate": 9.340779974241167e-05,
      "loss": 0.6175,
      "mean_token_accuracy": 0.8274259582161904,
      "num_tokens": 37950597.0,
      "step": 2820
    },
    {
      "epoch": 1.7235079171741778,
      "grad_norm": 0.6464765667915344,
      "learning_rate": 9.282654498485139e-05,
      "loss": 0.6636,
      "mean_token_accuracy": 0.8163545817136765,
      "num_tokens": 38086904.0,
      "step": 2830
    },
    {
      "epoch": 1.7295980511571254,
      "grad_norm": 0.6118177175521851,
      "learning_rate": 9.22460326710601e-05,
      "loss": 0.6696,
      "mean_token_accuracy": 0.8133967757225037,
      "num_tokens": 38219759.0,
      "step": 2840
    },
    {
      "epoch": 1.7356881851400732,
      "grad_norm": 0.5518969893455505,
      "learning_rate": 9.16662878977786e-05,
      "loss": 0.6659,
      "mean_token_accuracy": 0.8180875137448311,
      "num_tokens": 38349770.0,
      "step": 2850
    },
    {
      "epoch": 1.7417783191230207,
      "grad_norm": 0.6465517282485962,
      "learning_rate": 9.108733572856549e-05,
      "loss": 0.6581,
      "mean_token_accuracy": 0.8170399129390716,
      "num_tokens": 38482303.0,
      "step": 2860
    },
    {
      "epoch": 1.7478684531059683,
      "grad_norm": 0.5193557143211365,
      "learning_rate": 9.050920119271335e-05,
      "loss": 0.6543,
      "mean_token_accuracy": 0.8178304255008697,
      "num_tokens": 38615426.0,
      "step": 2870
    },
    {
      "epoch": 1.753958587088916,
      "grad_norm": 0.611529529094696,
      "learning_rate": 8.993190928416682e-05,
      "loss": 0.6248,
      "mean_token_accuracy": 0.8259203046560287,
      "num_tokens": 38755859.0,
      "step": 2880
    },
    {
      "epoch": 1.7600487210718636,
      "grad_norm": 0.5405944585800171,
      "learning_rate": 8.935548496044198e-05,
      "loss": 0.6232,
      "mean_token_accuracy": 0.8281204700469971,
      "num_tokens": 38893007.0,
      "step": 2890
    },
    {
      "epoch": 1.7661388550548112,
      "grad_norm": 0.6433010697364807,
      "learning_rate": 8.877995314154748e-05,
      "loss": 0.6751,
      "mean_token_accuracy": 0.8155393078923225,
      "num_tokens": 39020285.0,
      "step": 2900
    },
    {
      "epoch": 1.772228989037759,
      "grad_norm": 0.47974956035614014,
      "learning_rate": 8.820533870890717e-05,
      "loss": 0.6527,
      "mean_token_accuracy": 0.8197720810770989,
      "num_tokens": 39151426.0,
      "step": 2910
    },
    {
      "epoch": 1.7783191230207065,
      "grad_norm": 0.5529680848121643,
      "learning_rate": 8.763166650428436e-05,
      "loss": 0.6262,
      "mean_token_accuracy": 0.8256829127669334,
      "num_tokens": 39294242.0,
      "step": 2920
    },
    {
      "epoch": 1.784409257003654,
      "grad_norm": 0.6060122847557068,
      "learning_rate": 8.705896132870797e-05,
      "loss": 0.6563,
      "mean_token_accuracy": 0.8192467406392098,
      "num_tokens": 39425879.0,
      "step": 2930
    },
    {
      "epoch": 1.7904993909866018,
      "grad_norm": 0.6099355220794678,
      "learning_rate": 8.648724794140017e-05,
      "loss": 0.6664,
      "mean_token_accuracy": 0.8186777010560036,
      "num_tokens": 39559787.0,
      "step": 2940
    },
    {
      "epoch": 1.7965895249695494,
      "grad_norm": 0.5908733010292053,
      "learning_rate": 8.591655105870615e-05,
      "loss": 0.6712,
      "mean_token_accuracy": 0.8136340633034707,
      "num_tokens": 39689823.0,
      "step": 2950
    },
    {
      "epoch": 1.802679658952497,
      "grad_norm": 0.5845519304275513,
      "learning_rate": 8.534689535302553e-05,
      "loss": 0.6608,
      "mean_token_accuracy": 0.8170475289225578,
      "num_tokens": 39820725.0,
      "step": 2960
    },
    {
      "epoch": 1.8087697929354447,
      "grad_norm": 0.6311175227165222,
      "learning_rate": 8.47783054517457e-05,
      "loss": 0.6491,
      "mean_token_accuracy": 0.8193596869707107,
      "num_tokens": 39947874.0,
      "step": 2970
    },
    {
      "epoch": 1.814859926918392,
      "grad_norm": 0.5293188691139221,
      "learning_rate": 8.421080593617706e-05,
      "loss": 0.6105,
      "mean_token_accuracy": 0.83141258507967,
      "num_tokens": 40091297.0,
      "step": 2980
    },
    {
      "epoch": 1.8209500609013398,
      "grad_norm": 0.5252617597579956,
      "learning_rate": 8.364442134049049e-05,
      "loss": 0.6356,
      "mean_token_accuracy": 0.8237936720252037,
      "num_tokens": 40229207.0,
      "step": 2990
    },
    {
      "epoch": 1.8270401948842876,
      "grad_norm": 0.6039798855781555,
      "learning_rate": 8.30791761506565e-05,
      "loss": 0.6456,
      "mean_token_accuracy": 0.8206364914774895,
      "num_tokens": 40364863.0,
      "step": 3000
    },
    {
      "epoch": 1.833130328867235,
      "grad_norm": 0.5508609414100647,
      "learning_rate": 8.251509480338684e-05,
      "loss": 0.6229,
      "mean_token_accuracy": 0.8255992740392685,
      "num_tokens": 40504123.0,
      "step": 3010
    },
    {
      "epoch": 1.8392204628501827,
      "grad_norm": 0.5637634992599487,
      "learning_rate": 8.195220168507789e-05,
      "loss": 0.6026,
      "mean_token_accuracy": 0.8290412962436676,
      "num_tokens": 40646821.0,
      "step": 3020
    },
    {
      "epoch": 1.8453105968331305,
      "grad_norm": 0.5463610291481018,
      "learning_rate": 8.139052113075645e-05,
      "loss": 0.6278,
      "mean_token_accuracy": 0.8244929850101471,
      "num_tokens": 40778989.0,
      "step": 3030
    },
    {
      "epoch": 1.8514007308160778,
      "grad_norm": 0.5360645055770874,
      "learning_rate": 8.083007742302776e-05,
      "loss": 0.6336,
      "mean_token_accuracy": 0.8228462666273118,
      "num_tokens": 40917560.0,
      "step": 3040
    },
    {
      "epoch": 1.8574908647990256,
      "grad_norm": 0.5185632705688477,
      "learning_rate": 8.02708947910255e-05,
      "loss": 0.5991,
      "mean_token_accuracy": 0.830042028427124,
      "num_tokens": 41059707.0,
      "step": 3050
    },
    {
      "epoch": 1.8635809987819734,
      "grad_norm": 0.6445353627204895,
      "learning_rate": 7.971299740936456e-05,
      "loss": 0.6555,
      "mean_token_accuracy": 0.8169184163212776,
      "num_tokens": 41192515.0,
      "step": 3060
    },
    {
      "epoch": 1.8696711327649207,
      "grad_norm": 0.5360421538352966,
      "learning_rate": 7.915640939709576e-05,
      "loss": 0.6234,
      "mean_token_accuracy": 0.8257398083806038,
      "num_tokens": 41330047.0,
      "step": 3070
    },
    {
      "epoch": 1.8757612667478685,
      "grad_norm": 0.58651202917099,
      "learning_rate": 7.860115481666333e-05,
      "loss": 0.6564,
      "mean_token_accuracy": 0.8205534905195236,
      "num_tokens": 41460379.0,
      "step": 3080
    },
    {
      "epoch": 1.881851400730816,
      "grad_norm": 0.6842640042304993,
      "learning_rate": 7.804725767286427e-05,
      "loss": 0.6935,
      "mean_token_accuracy": 0.8097458809614182,
      "num_tokens": 41581210.0,
      "step": 3090
    },
    {
      "epoch": 1.8879415347137636,
      "grad_norm": 0.5175514817237854,
      "learning_rate": 7.749474191181096e-05,
      "loss": 0.6393,
      "mean_token_accuracy": 0.8219558611512184,
      "num_tokens": 41714792.0,
      "step": 3100
    },
    {
      "epoch": 1.8940316686967114,
      "grad_norm": 0.5963588356971741,
      "learning_rate": 7.694363141989575e-05,
      "loss": 0.658,
      "mean_token_accuracy": 0.8182344615459443,
      "num_tokens": 41846600.0,
      "step": 3110
    },
    {
      "epoch": 1.900121802679659,
      "grad_norm": 0.6149535775184631,
      "learning_rate": 7.639395002275827e-05,
      "loss": 0.6499,
      "mean_token_accuracy": 0.8208124756813049,
      "num_tokens": 41977627.0,
      "step": 3120
    },
    {
      "epoch": 1.9062119366626065,
      "grad_norm": 0.5739808678627014,
      "learning_rate": 7.584572148425544e-05,
      "loss": 0.6703,
      "mean_token_accuracy": 0.8125967502593994,
      "num_tokens": 42104510.0,
      "step": 3130
    },
    {
      "epoch": 1.9123020706455542,
      "grad_norm": 0.5982648730278015,
      "learning_rate": 7.529896950543416e-05,
      "loss": 0.6513,
      "mean_token_accuracy": 0.8201186507940292,
      "num_tokens": 42236168.0,
      "step": 3140
    },
    {
      "epoch": 1.9183922046285018,
      "grad_norm": 0.5896486043930054,
      "learning_rate": 7.475371772350658e-05,
      "loss": 0.6133,
      "mean_token_accuracy": 0.8260134413838387,
      "num_tokens": 42375086.0,
      "step": 3150
    },
    {
      "epoch": 1.9244823386114494,
      "grad_norm": 0.6223361492156982,
      "learning_rate": 7.420998971082833e-05,
      "loss": 0.6638,
      "mean_token_accuracy": 0.8162963137030601,
      "num_tokens": 42506457.0,
      "step": 3160
    },
    {
      "epoch": 1.9305724725943971,
      "grad_norm": 0.709854245185852,
      "learning_rate": 7.366780897387924e-05,
      "loss": 0.6324,
      "mean_token_accuracy": 0.8247174829244613,
      "num_tokens": 42640886.0,
      "step": 3170
    },
    {
      "epoch": 1.9366626065773447,
      "grad_norm": 0.6794169545173645,
      "learning_rate": 7.312719895224736e-05,
      "loss": 0.6164,
      "mean_token_accuracy": 0.82676922082901,
      "num_tokens": 42781318.0,
      "step": 3180
    },
    {
      "epoch": 1.9427527405602922,
      "grad_norm": 0.49305981397628784,
      "learning_rate": 7.258818301761532e-05,
      "loss": 0.6216,
      "mean_token_accuracy": 0.8258268669247627,
      "num_tokens": 42919381.0,
      "step": 3190
    },
    {
      "epoch": 1.94884287454324,
      "grad_norm": 0.5072576999664307,
      "learning_rate": 7.205078447275031e-05,
      "loss": 0.6407,
      "mean_token_accuracy": 0.819316141307354,
      "num_tokens": 43056494.0,
      "step": 3200
    },
    {
      "epoch": 1.9549330085261876,
      "grad_norm": 0.6188381314277649,
      "learning_rate": 7.151502655049623e-05,
      "loss": 0.6022,
      "mean_token_accuracy": 0.8328602254390717,
      "num_tokens": 43197795.0,
      "step": 3210
    },
    {
      "epoch": 1.9610231425091351,
      "grad_norm": 0.5626131296157837,
      "learning_rate": 7.098093241276962e-05,
      "loss": 0.6245,
      "mean_token_accuracy": 0.8258091285824776,
      "num_tokens": 43340325.0,
      "step": 3220
    },
    {
      "epoch": 1.967113276492083,
      "grad_norm": 0.5364338755607605,
      "learning_rate": 7.044852514955816e-05,
      "loss": 0.6454,
      "mean_token_accuracy": 0.8199462234973908,
      "num_tokens": 43472226.0,
      "step": 3230
    },
    {
      "epoch": 1.9732034104750305,
      "grad_norm": 0.5460382699966431,
      "learning_rate": 6.991782777792244e-05,
      "loss": 0.6214,
      "mean_token_accuracy": 0.8251617640256882,
      "num_tokens": 43609559.0,
      "step": 3240
    },
    {
      "epoch": 1.979293544457978,
      "grad_norm": 0.6013203263282776,
      "learning_rate": 6.938886324100097e-05,
      "loss": 0.6422,
      "mean_token_accuracy": 0.8197862133383751,
      "num_tokens": 43743060.0,
      "step": 3250
    },
    {
      "epoch": 1.9853836784409258,
      "grad_norm": 0.6688512563705444,
      "learning_rate": 6.88616544070182e-05,
      "loss": 0.6447,
      "mean_token_accuracy": 0.820338460803032,
      "num_tokens": 43876874.0,
      "step": 3260
    },
    {
      "epoch": 1.9914738124238733,
      "grad_norm": 0.6146946549415588,
      "learning_rate": 6.8336224068296e-05,
      "loss": 0.6015,
      "mean_token_accuracy": 0.8318811848759651,
      "num_tokens": 44021963.0,
      "step": 3270
    },
    {
      "epoch": 1.997563946406821,
      "grad_norm": 0.5972597599029541,
      "learning_rate": 6.781259494026821e-05,
      "loss": 0.6094,
      "mean_token_accuracy": 0.8282003849744797,
      "num_tokens": 44159207.0,
      "step": 3280
    }
  ],
  "logging_steps": 10,
  "max_steps": 4926,
  "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": 5.398768566923166e+18,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}