{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 3670,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0002725306171115161,
      "grad_norm": 58.5752067565918,
      "learning_rate": 0.0,
      "loss": 4.1988,
      "mean_token_accuracy": 0.27959318878129125,
      "num_tokens": 180497.0,
      "step": 1
    },
    {
      "epoch": 0.0027253061711151614,
      "grad_norm": 55.061241149902344,
      "learning_rate": 2.0361990950226245e-06,
      "loss": 4.1578,
      "mean_token_accuracy": 0.2840940863307979,
      "num_tokens": 1772191.0,
      "step": 10
    },
    {
      "epoch": 0.005450612342230323,
      "grad_norm": 31.145395278930664,
      "learning_rate": 4.298642533936651e-06,
      "loss": 3.805,
      "mean_token_accuracy": 0.30883588222786784,
      "num_tokens": 3528128.0,
      "step": 20
    },
    {
      "epoch": 0.008175918513345483,
      "grad_norm": 12.37236213684082,
      "learning_rate": 6.5610859728506795e-06,
      "loss": 3.0369,
      "mean_token_accuracy": 0.3894659894518554,
      "num_tokens": 5321820.0,
      "step": 30
    },
    {
      "epoch": 0.010901224684460645,
      "grad_norm": 4.672379016876221,
      "learning_rate": 8.823529411764707e-06,
      "loss": 2.3771,
      "mean_token_accuracy": 0.48662197683006525,
      "num_tokens": 7036353.0,
      "step": 40
    },
    {
      "epoch": 0.013626530855575806,
      "grad_norm": 2.293363094329834,
      "learning_rate": 1.1085972850678733e-05,
      "loss": 1.9359,
      "mean_token_accuracy": 0.5632315624505282,
      "num_tokens": 8794502.0,
      "step": 50
    },
    {
      "epoch": 0.016351837026690966,
      "grad_norm": 1.8514742851257324,
      "learning_rate": 1.3348416289592761e-05,
      "loss": 1.687,
      "mean_token_accuracy": 0.6138357298448682,
      "num_tokens": 10519144.0,
      "step": 60
    },
    {
      "epoch": 0.019077143197806127,
      "grad_norm": 1.5001717805862427,
      "learning_rate": 1.5610859728506788e-05,
      "loss": 1.511,
      "mean_token_accuracy": 0.6497381268069148,
      "num_tokens": 12272638.0,
      "step": 70
    },
    {
      "epoch": 0.02180244936892129,
      "grad_norm": 1.2374789714813232,
      "learning_rate": 1.7873303167420814e-05,
      "loss": 1.4223,
      "mean_token_accuracy": 0.667938649840653,
      "num_tokens": 14037595.0,
      "step": 80
    },
    {
      "epoch": 0.02452775554003645,
      "grad_norm": 1.232344388961792,
      "learning_rate": 2.0135746606334844e-05,
      "loss": 1.3629,
      "mean_token_accuracy": 0.6789748859591782,
      "num_tokens": 15813413.0,
      "step": 90
    },
    {
      "epoch": 0.027253061711151612,
      "grad_norm": 1.2523256540298462,
      "learning_rate": 2.239819004524887e-05,
      "loss": 1.3328,
      "mean_token_accuracy": 0.6849433833733201,
      "num_tokens": 17606667.0,
      "step": 100
    },
    {
      "epoch": 0.029978367882266772,
      "grad_norm": 1.1663717031478882,
      "learning_rate": 2.4660633484162897e-05,
      "loss": 1.2645,
      "mean_token_accuracy": 0.7004281104542315,
      "num_tokens": 19387650.0,
      "step": 110
    },
    {
      "epoch": 0.03270367405338193,
      "grad_norm": 1.148568868637085,
      "learning_rate": 2.6923076923076923e-05,
      "loss": 1.2482,
      "mean_token_accuracy": 0.7018253466114401,
      "num_tokens": 21111629.0,
      "step": 120
    },
    {
      "epoch": 0.03542898022449709,
      "grad_norm": 1.0586553812026978,
      "learning_rate": 2.9185520361990953e-05,
      "loss": 1.2162,
      "mean_token_accuracy": 0.7089157540351152,
      "num_tokens": 22855847.0,
      "step": 130
    },
    {
      "epoch": 0.038154286395612254,
      "grad_norm": 1.1304233074188232,
      "learning_rate": 3.1447963800904976e-05,
      "loss": 1.1918,
      "mean_token_accuracy": 0.7152549306862056,
      "num_tokens": 24647294.0,
      "step": 140
    },
    {
      "epoch": 0.04087959256672742,
      "grad_norm": 1.1450897455215454,
      "learning_rate": 3.371040723981901e-05,
      "loss": 1.1655,
      "mean_token_accuracy": 0.7181231830269098,
      "num_tokens": 26336128.0,
      "step": 150
    },
    {
      "epoch": 0.04360489873784258,
      "grad_norm": 1.0469363927841187,
      "learning_rate": 3.5972850678733036e-05,
      "loss": 1.2058,
      "mean_token_accuracy": 0.7123475090600551,
      "num_tokens": 28061528.0,
      "step": 160
    },
    {
      "epoch": 0.04633020490895774,
      "grad_norm": 1.0489130020141602,
      "learning_rate": 3.8235294117647055e-05,
      "loss": 1.178,
      "mean_token_accuracy": 0.7180261522531509,
      "num_tokens": 29834651.0,
      "step": 170
    },
    {
      "epoch": 0.0490555110800729,
      "grad_norm": 1.1838144063949585,
      "learning_rate": 4.049773755656109e-05,
      "loss": 1.1964,
      "mean_token_accuracy": 0.7144321620464325,
      "num_tokens": 31533806.0,
      "step": 180
    },
    {
      "epoch": 0.05178081725118806,
      "grad_norm": 1.0851892232894897,
      "learning_rate": 4.2760180995475115e-05,
      "loss": 1.1408,
      "mean_token_accuracy": 0.72315530218184,
      "num_tokens": 33328644.0,
      "step": 190
    },
    {
      "epoch": 0.054506123422303224,
      "grad_norm": 1.1279138326644897,
      "learning_rate": 4.502262443438914e-05,
      "loss": 1.1406,
      "mean_token_accuracy": 0.7242222094908357,
      "num_tokens": 35085202.0,
      "step": 200
    },
    {
      "epoch": 0.057231429593418384,
      "grad_norm": 1.07979416847229,
      "learning_rate": 4.728506787330317e-05,
      "loss": 1.1566,
      "mean_token_accuracy": 0.7219386071898043,
      "num_tokens": 36899685.0,
      "step": 210
    },
    {
      "epoch": 0.059956735764533545,
      "grad_norm": 1.1669384241104126,
      "learning_rate": 4.95475113122172e-05,
      "loss": 1.0724,
      "mean_token_accuracy": 0.7370353294536471,
      "num_tokens": 38663867.0,
      "step": 220
    },
    {
      "epoch": 0.0626820419356487,
      "grad_norm": 1.0999536514282227,
      "learning_rate": 4.994381233319287e-05,
      "loss": 1.0878,
      "mean_token_accuracy": 0.733780630491674,
      "num_tokens": 40368688.0,
      "step": 230
    },
    {
      "epoch": 0.06540734810676387,
      "grad_norm": 1.029019832611084,
      "learning_rate": 4.9873577749683945e-05,
      "loss": 1.0928,
      "mean_token_accuracy": 0.7319904958829284,
      "num_tokens": 42091556.0,
      "step": 240
    },
    {
      "epoch": 0.06813265427787903,
      "grad_norm": 1.0037535429000854,
      "learning_rate": 4.9803343166175026e-05,
      "loss": 1.0902,
      "mean_token_accuracy": 0.7336283173412085,
      "num_tokens": 43832865.0,
      "step": 250
    },
    {
      "epoch": 0.07085796044899419,
      "grad_norm": 1.0020638704299927,
      "learning_rate": 4.9733108582666106e-05,
      "loss": 1.0952,
      "mean_token_accuracy": 0.7330724668689073,
      "num_tokens": 45556421.0,
      "step": 260
    },
    {
      "epoch": 0.07358326662010935,
      "grad_norm": 1.0238029956817627,
      "learning_rate": 4.9662873999157186e-05,
      "loss": 1.1225,
      "mean_token_accuracy": 0.7294999814592302,
      "num_tokens": 47284838.0,
      "step": 270
    },
    {
      "epoch": 0.07630857279122451,
      "grad_norm": 1.0471445322036743,
      "learning_rate": 4.9592639415648266e-05,
      "loss": 1.0805,
      "mean_token_accuracy": 0.734851144067943,
      "num_tokens": 49025592.0,
      "step": 280
    },
    {
      "epoch": 0.07903387896233968,
      "grad_norm": 1.1734944581985474,
      "learning_rate": 4.9522404832139346e-05,
      "loss": 1.1459,
      "mean_token_accuracy": 0.7227287207730114,
      "num_tokens": 50771353.0,
      "step": 290
    },
    {
      "epoch": 0.08175918513345484,
      "grad_norm": 1.0051774978637695,
      "learning_rate": 4.945217024863043e-05,
      "loss": 1.0595,
      "mean_token_accuracy": 0.7396277584135532,
      "num_tokens": 52521467.0,
      "step": 300
    },
    {
      "epoch": 0.08448449130457,
      "grad_norm": 1.1502430438995361,
      "learning_rate": 4.938193566512151e-05,
      "loss": 1.0856,
      "mean_token_accuracy": 0.7348609266802668,
      "num_tokens": 54338858.0,
      "step": 310
    },
    {
      "epoch": 0.08720979747568516,
      "grad_norm": 1.079554796218872,
      "learning_rate": 4.931170108161259e-05,
      "loss": 1.0876,
      "mean_token_accuracy": 0.7352117918431759,
      "num_tokens": 56014081.0,
      "step": 320
    },
    {
      "epoch": 0.08993510364680032,
      "grad_norm": 1.0328309535980225,
      "learning_rate": 4.924146649810367e-05,
      "loss": 1.0622,
      "mean_token_accuracy": 0.7386171711608768,
      "num_tokens": 57778337.0,
      "step": 330
    },
    {
      "epoch": 0.09266040981791548,
      "grad_norm": 0.9490402936935425,
      "learning_rate": 4.9171231914594754e-05,
      "loss": 1.0549,
      "mean_token_accuracy": 0.74089086484164,
      "num_tokens": 59528977.0,
      "step": 340
    },
    {
      "epoch": 0.09538571598903064,
      "grad_norm": 0.9298028349876404,
      "learning_rate": 4.910099733108583e-05,
      "loss": 1.0704,
      "mean_token_accuracy": 0.7380094730295241,
      "num_tokens": 61295985.0,
      "step": 350
    },
    {
      "epoch": 0.0981110221601458,
      "grad_norm": 1.0823702812194824,
      "learning_rate": 4.903076274757691e-05,
      "loss": 1.0535,
      "mean_token_accuracy": 0.7405891207046806,
      "num_tokens": 63026897.0,
      "step": 360
    },
    {
      "epoch": 0.10083632833126097,
      "grad_norm": 0.9523594379425049,
      "learning_rate": 4.896052816406799e-05,
      "loss": 1.0444,
      "mean_token_accuracy": 0.7437008136883378,
      "num_tokens": 64770711.0,
      "step": 370
    },
    {
      "epoch": 0.10356163450237613,
      "grad_norm": 1.1027802228927612,
      "learning_rate": 4.889029358055907e-05,
      "loss": 1.0446,
      "mean_token_accuracy": 0.7440237908624112,
      "num_tokens": 66551008.0,
      "step": 380
    },
    {
      "epoch": 0.10628694067349129,
      "grad_norm": 0.9772301912307739,
      "learning_rate": 4.882005899705015e-05,
      "loss": 1.0598,
      "mean_token_accuracy": 0.7391570524312556,
      "num_tokens": 68320690.0,
      "step": 390
    },
    {
      "epoch": 0.10901224684460645,
      "grad_norm": 0.8989792466163635,
      "learning_rate": 4.874982441354123e-05,
      "loss": 1.0231,
      "mean_token_accuracy": 0.7470484189689159,
      "num_tokens": 70095284.0,
      "step": 400
    },
    {
      "epoch": 0.11173755301572161,
      "grad_norm": 0.9571840167045593,
      "learning_rate": 4.8679589830032316e-05,
      "loss": 1.0416,
      "mean_token_accuracy": 0.7444285966455937,
      "num_tokens": 71803511.0,
      "step": 410
    },
    {
      "epoch": 0.11446285918683677,
      "grad_norm": 0.9428116679191589,
      "learning_rate": 4.860935524652339e-05,
      "loss": 0.9991,
      "mean_token_accuracy": 0.7514733706600965,
      "num_tokens": 73528297.0,
      "step": 420
    },
    {
      "epoch": 0.11718816535795193,
      "grad_norm": 0.9464392066001892,
      "learning_rate": 4.853912066301447e-05,
      "loss": 1.0022,
      "mean_token_accuracy": 0.7517835930921137,
      "num_tokens": 75274289.0,
      "step": 430
    },
    {
      "epoch": 0.11991347152906709,
      "grad_norm": 1.0934405326843262,
      "learning_rate": 4.846888607950555e-05,
      "loss": 1.0466,
      "mean_token_accuracy": 0.7423655292019248,
      "num_tokens": 77077403.0,
      "step": 440
    },
    {
      "epoch": 0.12263877770018225,
      "grad_norm": 0.9798712134361267,
      "learning_rate": 4.8398651495996636e-05,
      "loss": 1.0416,
      "mean_token_accuracy": 0.7423399617895484,
      "num_tokens": 78859882.0,
      "step": 450
    },
    {
      "epoch": 0.1253640838712974,
      "grad_norm": 0.9301265478134155,
      "learning_rate": 4.832841691248771e-05,
      "loss": 1.0564,
      "mean_token_accuracy": 0.7402447827160359,
      "num_tokens": 80590976.0,
      "step": 460
    },
    {
      "epoch": 0.12808939004241257,
      "grad_norm": 1.028621792793274,
      "learning_rate": 4.825818232897879e-05,
      "loss": 1.0135,
      "mean_token_accuracy": 0.7484846420586109,
      "num_tokens": 82275126.0,
      "step": 470
    },
    {
      "epoch": 0.13081469621352773,
      "grad_norm": 0.9343350529670715,
      "learning_rate": 4.818794774546987e-05,
      "loss": 1.0102,
      "mean_token_accuracy": 0.7489786158315838,
      "num_tokens": 84096910.0,
      "step": 480
    },
    {
      "epoch": 0.1335400023846429,
      "grad_norm": 0.9360990524291992,
      "learning_rate": 4.811771316196095e-05,
      "loss": 1.0402,
      "mean_token_accuracy": 0.7446358595974744,
      "num_tokens": 85876326.0,
      "step": 490
    },
    {
      "epoch": 0.13626530855575805,
      "grad_norm": 0.9260503053665161,
      "learning_rate": 4.804747857845203e-05,
      "loss": 1.0412,
      "mean_token_accuracy": 0.7441607388667763,
      "num_tokens": 87607478.0,
      "step": 500
    },
    {
      "epoch": 0.1389906147268732,
      "grad_norm": 0.9006381034851074,
      "learning_rate": 4.797724399494311e-05,
      "loss": 1.0014,
      "mean_token_accuracy": 0.7509601839818061,
      "num_tokens": 89350066.0,
      "step": 510
    },
    {
      "epoch": 0.14171592089798837,
      "grad_norm": 0.9362842440605164,
      "learning_rate": 4.79070094114342e-05,
      "loss": 1.0138,
      "mean_token_accuracy": 0.7497716393321753,
      "num_tokens": 91067910.0,
      "step": 520
    },
    {
      "epoch": 0.14444122706910353,
      "grad_norm": 0.9452476501464844,
      "learning_rate": 4.783677482792527e-05,
      "loss": 1.0052,
      "mean_token_accuracy": 0.7499888489954174,
      "num_tokens": 92797017.0,
      "step": 530
    },
    {
      "epoch": 0.1471665332402187,
      "grad_norm": 0.9116583466529846,
      "learning_rate": 4.776654024441635e-05,
      "loss": 1.0263,
      "mean_token_accuracy": 0.7451268086209893,
      "num_tokens": 94606329.0,
      "step": 540
    },
    {
      "epoch": 0.14989183941133385,
      "grad_norm": 0.9305568933486938,
      "learning_rate": 4.769630566090743e-05,
      "loss": 1.0039,
      "mean_token_accuracy": 0.7499175570905209,
      "num_tokens": 96331087.0,
      "step": 550
    },
    {
      "epoch": 0.15261714558244902,
      "grad_norm": 0.9302889704704285,
      "learning_rate": 4.762607107739852e-05,
      "loss": 1.0239,
      "mean_token_accuracy": 0.7466340310871601,
      "num_tokens": 98138711.0,
      "step": 560
    },
    {
      "epoch": 0.15534245175356418,
      "grad_norm": 0.9294123649597168,
      "learning_rate": 4.755583649388959e-05,
      "loss": 1.0285,
      "mean_token_accuracy": 0.7461859748698771,
      "num_tokens": 99885005.0,
      "step": 570
    },
    {
      "epoch": 0.15806775792467936,
      "grad_norm": 0.8810741901397705,
      "learning_rate": 4.748560191038067e-05,
      "loss": 0.9361,
      "mean_token_accuracy": 0.7639201901853084,
      "num_tokens": 101636075.0,
      "step": 580
    },
    {
      "epoch": 0.16079306409579452,
      "grad_norm": 0.987924337387085,
      "learning_rate": 4.741536732687175e-05,
      "loss": 0.9949,
      "mean_token_accuracy": 0.7529382834210991,
      "num_tokens": 103349118.0,
      "step": 590
    },
    {
      "epoch": 0.16351837026690969,
      "grad_norm": 0.9978275895118713,
      "learning_rate": 4.734513274336283e-05,
      "loss": 1.0056,
      "mean_token_accuracy": 0.7498491807840765,
      "num_tokens": 105033010.0,
      "step": 600
    },
    {
      "epoch": 0.16624367643802485,
      "grad_norm": 0.9176990985870361,
      "learning_rate": 4.727489815985391e-05,
      "loss": 0.9731,
      "mean_token_accuracy": 0.7562306389212609,
      "num_tokens": 106723283.0,
      "step": 610
    },
    {
      "epoch": 0.16896898260914,
      "grad_norm": 0.8888489007949829,
      "learning_rate": 4.720466357634499e-05,
      "loss": 0.9907,
      "mean_token_accuracy": 0.7524073977023363,
      "num_tokens": 108436878.0,
      "step": 620
    },
    {
      "epoch": 0.17169428878025517,
      "grad_norm": 0.9666376709938049,
      "learning_rate": 4.713442899283608e-05,
      "loss": 0.9752,
      "mean_token_accuracy": 0.7562545623630286,
      "num_tokens": 110203157.0,
      "step": 630
    },
    {
      "epoch": 0.17441959495137033,
      "grad_norm": 1.034400463104248,
      "learning_rate": 4.706419440932715e-05,
      "loss": 1.0029,
      "mean_token_accuracy": 0.7515896405093372,
      "num_tokens": 111949130.0,
      "step": 640
    },
    {
      "epoch": 0.1771449011224855,
      "grad_norm": 1.044718623161316,
      "learning_rate": 4.6993959825818233e-05,
      "loss": 1.0037,
      "mean_token_accuracy": 0.7487363073974848,
      "num_tokens": 113652926.0,
      "step": 650
    },
    {
      "epoch": 0.17987020729360065,
      "grad_norm": 0.906299352645874,
      "learning_rate": 4.6923725242309314e-05,
      "loss": 0.9903,
      "mean_token_accuracy": 0.7514964304864407,
      "num_tokens": 115334647.0,
      "step": 660
    },
    {
      "epoch": 0.1825955134647158,
      "grad_norm": 0.8646354675292969,
      "learning_rate": 4.68534906588004e-05,
      "loss": 0.9501,
      "mean_token_accuracy": 0.7617973797954619,
      "num_tokens": 117032381.0,
      "step": 670
    },
    {
      "epoch": 0.18532081963583097,
      "grad_norm": 0.8861644864082336,
      "learning_rate": 4.6783256075291474e-05,
      "loss": 0.9859,
      "mean_token_accuracy": 0.7530626423656941,
      "num_tokens": 118801553.0,
      "step": 680
    },
    {
      "epoch": 0.18804612580694613,
      "grad_norm": 0.9357079863548279,
      "learning_rate": 4.6713021491782554e-05,
      "loss": 0.9989,
      "mean_token_accuracy": 0.7504322250373662,
      "num_tokens": 120530271.0,
      "step": 690
    },
    {
      "epoch": 0.1907714319780613,
      "grad_norm": 0.8540539741516113,
      "learning_rate": 4.6642786908273634e-05,
      "loss": 0.9904,
      "mean_token_accuracy": 0.7538057889789342,
      "num_tokens": 122314411.0,
      "step": 700
    },
    {
      "epoch": 0.19349673814917645,
      "grad_norm": 0.8792332410812378,
      "learning_rate": 4.6572552324764715e-05,
      "loss": 0.9839,
      "mean_token_accuracy": 0.7542797921225428,
      "num_tokens": 124054113.0,
      "step": 710
    },
    {
      "epoch": 0.1962220443202916,
      "grad_norm": 0.9250332713127136,
      "learning_rate": 4.6502317741255795e-05,
      "loss": 0.9988,
      "mean_token_accuracy": 0.7507336236536503,
      "num_tokens": 125786705.0,
      "step": 720
    },
    {
      "epoch": 0.19894735049140677,
      "grad_norm": 0.9268140196800232,
      "learning_rate": 4.6432083157746875e-05,
      "loss": 0.9998,
      "mean_token_accuracy": 0.7502107596024871,
      "num_tokens": 127510112.0,
      "step": 730
    },
    {
      "epoch": 0.20167265666252193,
      "grad_norm": 0.9151081442832947,
      "learning_rate": 4.636184857423796e-05,
      "loss": 0.9637,
      "mean_token_accuracy": 0.7585853973403573,
      "num_tokens": 129321733.0,
      "step": 740
    },
    {
      "epoch": 0.2043979628336371,
      "grad_norm": 0.8551493287086487,
      "learning_rate": 4.6291613990729035e-05,
      "loss": 1.0006,
      "mean_token_accuracy": 0.7512839620932936,
      "num_tokens": 131068939.0,
      "step": 750
    },
    {
      "epoch": 0.20712326900475225,
      "grad_norm": 1.0179383754730225,
      "learning_rate": 4.6221379407220116e-05,
      "loss": 1.0366,
      "mean_token_accuracy": 0.7441357178613544,
      "num_tokens": 132800192.0,
      "step": 760
    },
    {
      "epoch": 0.2098485751758674,
      "grad_norm": 0.8894006013870239,
      "learning_rate": 4.6151144823711196e-05,
      "loss": 0.9706,
      "mean_token_accuracy": 0.7571743190288543,
      "num_tokens": 134501880.0,
      "step": 770
    },
    {
      "epoch": 0.21257388134698257,
      "grad_norm": 0.9344801306724548,
      "learning_rate": 4.608091024020228e-05,
      "loss": 0.9495,
      "mean_token_accuracy": 0.7609304752200842,
      "num_tokens": 136227230.0,
      "step": 780
    },
    {
      "epoch": 0.21529918751809773,
      "grad_norm": 0.9688650965690613,
      "learning_rate": 4.6010675656693356e-05,
      "loss": 0.9634,
      "mean_token_accuracy": 0.7590550154447555,
      "num_tokens": 137938433.0,
      "step": 790
    },
    {
      "epoch": 0.2180244936892129,
      "grad_norm": 0.8858566880226135,
      "learning_rate": 4.5940441073184436e-05,
      "loss": 0.9736,
      "mean_token_accuracy": 0.7575385818257928,
      "num_tokens": 139667163.0,
      "step": 800
    },
    {
      "epoch": 0.22074979986032806,
      "grad_norm": 0.7869423031806946,
      "learning_rate": 4.5870206489675517e-05,
      "loss": 0.9494,
      "mean_token_accuracy": 0.7619166353717446,
      "num_tokens": 141481099.0,
      "step": 810
    },
    {
      "epoch": 0.22347510603144322,
      "grad_norm": 0.889818549156189,
      "learning_rate": 4.57999719061666e-05,
      "loss": 1.0118,
      "mean_token_accuracy": 0.7489442000165581,
      "num_tokens": 143204243.0,
      "step": 820
    },
    {
      "epoch": 0.22620041220255838,
      "grad_norm": 0.8601938486099243,
      "learning_rate": 4.572973732265768e-05,
      "loss": 0.9731,
      "mean_token_accuracy": 0.7579358568415046,
      "num_tokens": 144995581.0,
      "step": 830
    },
    {
      "epoch": 0.22892571837367354,
      "grad_norm": 0.868138313293457,
      "learning_rate": 4.565950273914876e-05,
      "loss": 0.9795,
      "mean_token_accuracy": 0.7540082458406687,
      "num_tokens": 146711076.0,
      "step": 840
    },
    {
      "epoch": 0.2316510245447887,
      "grad_norm": 0.8250023722648621,
      "learning_rate": 4.5589268155639844e-05,
      "loss": 0.971,
      "mean_token_accuracy": 0.7573945553973317,
      "num_tokens": 148463902.0,
      "step": 850
    },
    {
      "epoch": 0.23437633071590386,
      "grad_norm": 0.8885928392410278,
      "learning_rate": 4.551903357213092e-05,
      "loss": 0.9964,
      "mean_token_accuracy": 0.7519168628379702,
      "num_tokens": 150212801.0,
      "step": 860
    },
    {
      "epoch": 0.23710163688701902,
      "grad_norm": 0.8809767961502075,
      "learning_rate": 4.5448798988622e-05,
      "loss": 0.945,
      "mean_token_accuracy": 0.7604282542131842,
      "num_tokens": 151950016.0,
      "step": 870
    },
    {
      "epoch": 0.23982694305813418,
      "grad_norm": 1.013855218887329,
      "learning_rate": 4.537856440511308e-05,
      "loss": 0.9567,
      "mean_token_accuracy": 0.7587636118754745,
      "num_tokens": 153686341.0,
      "step": 880
    },
    {
      "epoch": 0.24255224922924934,
      "grad_norm": 0.8054437637329102,
      "learning_rate": 4.5308329821604165e-05,
      "loss": 0.9852,
      "mean_token_accuracy": 0.753642223123461,
      "num_tokens": 155442220.0,
      "step": 890
    },
    {
      "epoch": 0.2452775554003645,
      "grad_norm": 0.9395073056221008,
      "learning_rate": 4.523809523809524e-05,
      "loss": 0.9552,
      "mean_token_accuracy": 0.7594983844086528,
      "num_tokens": 157222278.0,
      "step": 900
    },
    {
      "epoch": 0.24800286157147966,
      "grad_norm": 0.8815314173698425,
      "learning_rate": 4.516786065458632e-05,
      "loss": 0.9484,
      "mean_token_accuracy": 0.7606731666252017,
      "num_tokens": 158989873.0,
      "step": 910
    },
    {
      "epoch": 0.2507281677425948,
      "grad_norm": 0.9090284109115601,
      "learning_rate": 4.50976260710774e-05,
      "loss": 0.9742,
      "mean_token_accuracy": 0.7581370803527534,
      "num_tokens": 160740690.0,
      "step": 920
    },
    {
      "epoch": 0.25345347391371,
      "grad_norm": 0.907217800617218,
      "learning_rate": 4.502739148756848e-05,
      "loss": 0.9779,
      "mean_token_accuracy": 0.7548290783539414,
      "num_tokens": 162505210.0,
      "step": 930
    },
    {
      "epoch": 0.25617878008482514,
      "grad_norm": 0.9016557335853577,
      "learning_rate": 4.495715690405956e-05,
      "loss": 0.9842,
      "mean_token_accuracy": 0.7541791129857301,
      "num_tokens": 164184119.0,
      "step": 940
    },
    {
      "epoch": 0.2589040862559403,
      "grad_norm": 0.8286603093147278,
      "learning_rate": 4.488692232055064e-05,
      "loss": 0.9433,
      "mean_token_accuracy": 0.7619124032557011,
      "num_tokens": 165971146.0,
      "step": 950
    },
    {
      "epoch": 0.26162939242705546,
      "grad_norm": 0.9046083092689514,
      "learning_rate": 4.4816687737041726e-05,
      "loss": 0.9781,
      "mean_token_accuracy": 0.7572765513323247,
      "num_tokens": 167701701.0,
      "step": 960
    },
    {
      "epoch": 0.2643546985981706,
      "grad_norm": 0.936540961265564,
      "learning_rate": 4.47464531535328e-05,
      "loss": 0.969,
      "mean_token_accuracy": 0.7577900117263198,
      "num_tokens": 169443326.0,
      "step": 970
    },
    {
      "epoch": 0.2670800047692858,
      "grad_norm": 0.8902041912078857,
      "learning_rate": 4.467621857002388e-05,
      "loss": 0.9706,
      "mean_token_accuracy": 0.7567241456359625,
      "num_tokens": 171199385.0,
      "step": 980
    },
    {
      "epoch": 0.26980531094040094,
      "grad_norm": 0.9205445647239685,
      "learning_rate": 4.460598398651496e-05,
      "loss": 0.9383,
      "mean_token_accuracy": 0.7634873200207949,
      "num_tokens": 172929493.0,
      "step": 990
    },
    {
      "epoch": 0.2725306171115161,
      "grad_norm": 0.9309903979301453,
      "learning_rate": 4.453574940300605e-05,
      "loss": 0.9442,
      "mean_token_accuracy": 0.7614085223525763,
      "num_tokens": 174672601.0,
      "step": 1000
    },
    {
      "epoch": 0.27525592328263127,
      "grad_norm": 0.8428636193275452,
      "learning_rate": 4.446551481949712e-05,
      "loss": 0.9607,
      "mean_token_accuracy": 0.7597405662760138,
      "num_tokens": 176441962.0,
      "step": 1010
    },
    {
      "epoch": 0.2779812294537464,
      "grad_norm": 0.9295870661735535,
      "learning_rate": 4.43952802359882e-05,
      "loss": 0.9468,
      "mean_token_accuracy": 0.7614248601719737,
      "num_tokens": 178197361.0,
      "step": 1020
    },
    {
      "epoch": 0.2807065356248616,
      "grad_norm": 0.9550367593765259,
      "learning_rate": 4.432504565247929e-05,
      "loss": 0.9766,
      "mean_token_accuracy": 0.7538504358381033,
      "num_tokens": 179887580.0,
      "step": 1030
    },
    {
      "epoch": 0.28343184179597675,
      "grad_norm": 0.9089730978012085,
      "learning_rate": 4.425481106897036e-05,
      "loss": 0.9382,
      "mean_token_accuracy": 0.7631622815504671,
      "num_tokens": 181655865.0,
      "step": 1040
    },
    {
      "epoch": 0.2861571479670919,
      "grad_norm": 1.0070922374725342,
      "learning_rate": 4.418457648546144e-05,
      "loss": 0.962,
      "mean_token_accuracy": 0.7598225042223931,
      "num_tokens": 183445880.0,
      "step": 1050
    },
    {
      "epoch": 0.28888245413820707,
      "grad_norm": 0.8544902205467224,
      "learning_rate": 4.411434190195252e-05,
      "loss": 0.9296,
      "mean_token_accuracy": 0.7648898587562144,
      "num_tokens": 185175706.0,
      "step": 1060
    },
    {
      "epoch": 0.29160776030932223,
      "grad_norm": 0.9063568115234375,
      "learning_rate": 4.404410731844361e-05,
      "loss": 0.942,
      "mean_token_accuracy": 0.7613558363169431,
      "num_tokens": 186873396.0,
      "step": 1070
    },
    {
      "epoch": 0.2943330664804374,
      "grad_norm": 0.8668667078018188,
      "learning_rate": 4.397387273493468e-05,
      "loss": 0.9765,
      "mean_token_accuracy": 0.7555979456752538,
      "num_tokens": 188591288.0,
      "step": 1080
    },
    {
      "epoch": 0.29705837265155255,
      "grad_norm": 0.9103459119796753,
      "learning_rate": 4.390363815142576e-05,
      "loss": 0.9691,
      "mean_token_accuracy": 0.7566804554313421,
      "num_tokens": 190375182.0,
      "step": 1090
    },
    {
      "epoch": 0.2997836788226677,
      "grad_norm": 0.9022314548492432,
      "learning_rate": 4.383340356791684e-05,
      "loss": 0.955,
      "mean_token_accuracy": 0.7589525409042835,
      "num_tokens": 192104609.0,
      "step": 1100
    },
    {
      "epoch": 0.30250898499378287,
      "grad_norm": 0.8525939583778381,
      "learning_rate": 4.376316898440793e-05,
      "loss": 0.8952,
      "mean_token_accuracy": 0.7715249681845308,
      "num_tokens": 193819768.0,
      "step": 1110
    },
    {
      "epoch": 0.30523429116489803,
      "grad_norm": 0.8655187487602234,
      "learning_rate": 4.3692934400899e-05,
      "loss": 0.957,
      "mean_token_accuracy": 0.7608988376334309,
      "num_tokens": 195541705.0,
      "step": 1120
    },
    {
      "epoch": 0.3079595973360132,
      "grad_norm": 0.8646089434623718,
      "learning_rate": 4.362269981739008e-05,
      "loss": 0.944,
      "mean_token_accuracy": 0.7616777900606394,
      "num_tokens": 197250976.0,
      "step": 1130
    },
    {
      "epoch": 0.31068490350712835,
      "grad_norm": 0.8167153000831604,
      "learning_rate": 4.355246523388117e-05,
      "loss": 0.9149,
      "mean_token_accuracy": 0.7681540355086327,
      "num_tokens": 198964087.0,
      "step": 1140
    },
    {
      "epoch": 0.31341020967824357,
      "grad_norm": 0.9367976188659668,
      "learning_rate": 4.348223065037224e-05,
      "loss": 0.9599,
      "mean_token_accuracy": 0.7586290875449777,
      "num_tokens": 200654341.0,
      "step": 1150
    },
    {
      "epoch": 0.31613551584935873,
      "grad_norm": 0.8891621828079224,
      "learning_rate": 4.3411996066863323e-05,
      "loss": 0.9865,
      "mean_token_accuracy": 0.7538609990850091,
      "num_tokens": 202410389.0,
      "step": 1160
    },
    {
      "epoch": 0.3188608220204739,
      "grad_norm": 0.8583462834358215,
      "learning_rate": 4.3341761483354404e-05,
      "loss": 0.9421,
      "mean_token_accuracy": 0.7621322822757065,
      "num_tokens": 204176073.0,
      "step": 1170
    },
    {
      "epoch": 0.32158612819158905,
      "grad_norm": 0.9226254224777222,
      "learning_rate": 4.327152689984549e-05,
      "loss": 0.9369,
      "mean_token_accuracy": 0.7635338991880417,
      "num_tokens": 206017214.0,
      "step": 1180
    },
    {
      "epoch": 0.3243114343627042,
      "grad_norm": 0.958926796913147,
      "learning_rate": 4.3201292316336564e-05,
      "loss": 0.948,
      "mean_token_accuracy": 0.7608529534190893,
      "num_tokens": 207832984.0,
      "step": 1190
    },
    {
      "epoch": 0.32703674053381937,
      "grad_norm": 0.9389734268188477,
      "learning_rate": 4.3131057732827644e-05,
      "loss": 0.9579,
      "mean_token_accuracy": 0.7594182942993939,
      "num_tokens": 209533427.0,
      "step": 1200
    },
    {
      "epoch": 0.32976204670493453,
      "grad_norm": 0.8880856037139893,
      "learning_rate": 4.3060823149318724e-05,
      "loss": 0.9352,
      "mean_token_accuracy": 0.7635649677366019,
      "num_tokens": 211344332.0,
      "step": 1210
    },
    {
      "epoch": 0.3324873528760497,
      "grad_norm": 0.9591731429100037,
      "learning_rate": 4.299058856580981e-05,
      "loss": 0.9633,
      "mean_token_accuracy": 0.7583153389394284,
      "num_tokens": 213088746.0,
      "step": 1220
    },
    {
      "epoch": 0.33521265904716485,
      "grad_norm": 0.9024151563644409,
      "learning_rate": 4.2920353982300885e-05,
      "loss": 0.9647,
      "mean_token_accuracy": 0.7571177691221237,
      "num_tokens": 214889528.0,
      "step": 1230
    },
    {
      "epoch": 0.33793796521828,
      "grad_norm": 0.9884682893753052,
      "learning_rate": 4.2850119398791965e-05,
      "loss": 0.9719,
      "mean_token_accuracy": 0.7572076172567904,
      "num_tokens": 216603081.0,
      "step": 1240
    },
    {
      "epoch": 0.3406632713893952,
      "grad_norm": 0.8637387752532959,
      "learning_rate": 4.277988481528305e-05,
      "loss": 0.9211,
      "mean_token_accuracy": 0.7659906836226582,
      "num_tokens": 218292845.0,
      "step": 1250
    },
    {
      "epoch": 0.34338857756051033,
      "grad_norm": 0.9273796081542969,
      "learning_rate": 4.2709650231774125e-05,
      "loss": 0.9454,
      "mean_token_accuracy": 0.7626127757132053,
      "num_tokens": 220029696.0,
      "step": 1260
    },
    {
      "epoch": 0.3461138837316255,
      "grad_norm": 0.8521084189414978,
      "learning_rate": 4.2639415648265206e-05,
      "loss": 1.0002,
      "mean_token_accuracy": 0.7520326241850853,
      "num_tokens": 221810553.0,
      "step": 1270
    },
    {
      "epoch": 0.34883918990274065,
      "grad_norm": 0.9349818229675293,
      "learning_rate": 4.2569181064756286e-05,
      "loss": 0.9412,
      "mean_token_accuracy": 0.7652541293762625,
      "num_tokens": 223552801.0,
      "step": 1280
    },
    {
      "epoch": 0.3515644960738558,
      "grad_norm": 0.9496633410453796,
      "learning_rate": 4.249894648124737e-05,
      "loss": 0.9486,
      "mean_token_accuracy": 0.7618991539813578,
      "num_tokens": 225254929.0,
      "step": 1290
    },
    {
      "epoch": 0.354289802244971,
      "grad_norm": 0.8988118767738342,
      "learning_rate": 4.2428711897738446e-05,
      "loss": 0.9679,
      "mean_token_accuracy": 0.7575280690565706,
      "num_tokens": 227017304.0,
      "step": 1300
    },
    {
      "epoch": 0.35701510841608614,
      "grad_norm": 0.9880586862564087,
      "learning_rate": 4.2358477314229526e-05,
      "loss": 0.9199,
      "mean_token_accuracy": 0.7675149342045188,
      "num_tokens": 228785771.0,
      "step": 1310
    },
    {
      "epoch": 0.3597404145872013,
      "grad_norm": 0.9548985958099365,
      "learning_rate": 4.2288242730720607e-05,
      "loss": 0.9276,
      "mean_token_accuracy": 0.7658209035173058,
      "num_tokens": 230489667.0,
      "step": 1320
    },
    {
      "epoch": 0.36246572075831646,
      "grad_norm": 0.8222749829292297,
      "learning_rate": 4.2218008147211694e-05,
      "loss": 0.9464,
      "mean_token_accuracy": 0.7618509028106928,
      "num_tokens": 232269581.0,
      "step": 1330
    },
    {
      "epoch": 0.3651910269294316,
      "grad_norm": 0.8697149753570557,
      "learning_rate": 4.214777356370277e-05,
      "loss": 0.9507,
      "mean_token_accuracy": 0.7589831293560565,
      "num_tokens": 234028126.0,
      "step": 1340
    },
    {
      "epoch": 0.3679163331005468,
      "grad_norm": 1.0205236673355103,
      "learning_rate": 4.207753898019385e-05,
      "loss": 0.9703,
      "mean_token_accuracy": 0.7572768048383296,
      "num_tokens": 235817129.0,
      "step": 1350
    },
    {
      "epoch": 0.37064163927166194,
      "grad_norm": 0.8218963742256165,
      "learning_rate": 4.2007304396684934e-05,
      "loss": 0.9508,
      "mean_token_accuracy": 0.7612483019009233,
      "num_tokens": 237588984.0,
      "step": 1360
    },
    {
      "epoch": 0.3733669454427771,
      "grad_norm": 0.9388371109962463,
      "learning_rate": 4.193706981317601e-05,
      "loss": 0.9144,
      "mean_token_accuracy": 0.7678116019815207,
      "num_tokens": 239305200.0,
      "step": 1370
    },
    {
      "epoch": 0.37609225161389226,
      "grad_norm": 0.8573911190032959,
      "learning_rate": 4.186683522966709e-05,
      "loss": 0.9368,
      "mean_token_accuracy": 0.7639053143560887,
      "num_tokens": 241078759.0,
      "step": 1380
    },
    {
      "epoch": 0.3788175577850074,
      "grad_norm": 0.918056845664978,
      "learning_rate": 4.179660064615817e-05,
      "loss": 0.9596,
      "mean_token_accuracy": 0.7594995486550034,
      "num_tokens": 242876841.0,
      "step": 1390
    },
    {
      "epoch": 0.3815428639561226,
      "grad_norm": 0.8583497405052185,
      "learning_rate": 4.1726366062649255e-05,
      "loss": 0.9252,
      "mean_token_accuracy": 0.7642927357926965,
      "num_tokens": 244578724.0,
      "step": 1400
    },
    {
      "epoch": 0.38426817012723774,
      "grad_norm": 0.8453887104988098,
      "learning_rate": 4.165613147914033e-05,
      "loss": 0.9525,
      "mean_token_accuracy": 0.7590734004974365,
      "num_tokens": 246363009.0,
      "step": 1410
    },
    {
      "epoch": 0.3869934762983529,
      "grad_norm": 0.8691115975379944,
      "learning_rate": 4.158589689563141e-05,
      "loss": 0.9764,
      "mean_token_accuracy": 0.7560055317357183,
      "num_tokens": 248120613.0,
      "step": 1420
    },
    {
      "epoch": 0.38971878246946806,
      "grad_norm": 0.8741028904914856,
      "learning_rate": 4.151566231212249e-05,
      "loss": 0.9514,
      "mean_token_accuracy": 0.7612073054537177,
      "num_tokens": 249941867.0,
      "step": 1430
    },
    {
      "epoch": 0.3924440886405832,
      "grad_norm": 0.8339214324951172,
      "learning_rate": 4.1445427728613576e-05,
      "loss": 0.9189,
      "mean_token_accuracy": 0.7668958367779851,
      "num_tokens": 251695734.0,
      "step": 1440
    },
    {
      "epoch": 0.3951693948116984,
      "grad_norm": 0.8296242356300354,
      "learning_rate": 4.137519314510465e-05,
      "loss": 0.9627,
      "mean_token_accuracy": 0.7588108706288039,
      "num_tokens": 253503211.0,
      "step": 1450
    },
    {
      "epoch": 0.39789470098281354,
      "grad_norm": 1.0061241388320923,
      "learning_rate": 4.130495856159573e-05,
      "loss": 0.9399,
      "mean_token_accuracy": 0.7630201263353229,
      "num_tokens": 255313734.0,
      "step": 1460
    },
    {
      "epoch": 0.4006200071539287,
      "grad_norm": 0.795073390007019,
      "learning_rate": 4.1234723978086816e-05,
      "loss": 0.8906,
      "mean_token_accuracy": 0.7732970826327801,
      "num_tokens": 257016513.0,
      "step": 1470
    },
    {
      "epoch": 0.40334531332504386,
      "grad_norm": 0.9418059587478638,
      "learning_rate": 4.1164489394577896e-05,
      "loss": 0.9642,
      "mean_token_accuracy": 0.757817355170846,
      "num_tokens": 258760362.0,
      "step": 1480
    },
    {
      "epoch": 0.406070619496159,
      "grad_norm": 0.8528169989585876,
      "learning_rate": 4.109425481106897e-05,
      "loss": 0.924,
      "mean_token_accuracy": 0.7663666036911309,
      "num_tokens": 260546971.0,
      "step": 1490
    },
    {
      "epoch": 0.4087959256672742,
      "grad_norm": 0.8477678298950195,
      "learning_rate": 4.102402022756005e-05,
      "loss": 0.9325,
      "mean_token_accuracy": 0.7635815197601914,
      "num_tokens": 262307672.0,
      "step": 1500
    },
    {
      "epoch": 0.41152123183838935,
      "grad_norm": 0.9021455645561218,
      "learning_rate": 4.095378564405114e-05,
      "loss": 0.9451,
      "mean_token_accuracy": 0.7621158785186708,
      "num_tokens": 264023889.0,
      "step": 1510
    },
    {
      "epoch": 0.4142465380095045,
      "grad_norm": 0.8655399084091187,
      "learning_rate": 4.088355106054221e-05,
      "loss": 0.9435,
      "mean_token_accuracy": 0.76206961190328,
      "num_tokens": 265798519.0,
      "step": 1520
    },
    {
      "epoch": 0.41697184418061967,
      "grad_norm": 0.8364386558532715,
      "learning_rate": 4.081331647703329e-05,
      "loss": 0.9599,
      "mean_token_accuracy": 0.7590883946046233,
      "num_tokens": 267546100.0,
      "step": 1530
    },
    {
      "epoch": 0.4196971503517348,
      "grad_norm": 0.8579459190368652,
      "learning_rate": 4.074308189352437e-05,
      "loss": 0.9233,
      "mean_token_accuracy": 0.7652724193409085,
      "num_tokens": 269297641.0,
      "step": 1540
    },
    {
      "epoch": 0.42242245652285,
      "grad_norm": 0.7599934339523315,
      "learning_rate": 4.067284731001546e-05,
      "loss": 0.9606,
      "mean_token_accuracy": 0.7579821115359664,
      "num_tokens": 271053006.0,
      "step": 1550
    },
    {
      "epoch": 0.42514776269396515,
      "grad_norm": 0.934144139289856,
      "learning_rate": 4.060261272650653e-05,
      "loss": 0.9439,
      "mean_token_accuracy": 0.7628989959135651,
      "num_tokens": 272829487.0,
      "step": 1560
    },
    {
      "epoch": 0.4278730688650803,
      "grad_norm": 0.8444865345954895,
      "learning_rate": 4.053237814299761e-05,
      "loss": 0.937,
      "mean_token_accuracy": 0.7626888936385512,
      "num_tokens": 274621990.0,
      "step": 1570
    },
    {
      "epoch": 0.43059837503619547,
      "grad_norm": 0.854570209980011,
      "learning_rate": 4.04621435594887e-05,
      "loss": 0.9555,
      "mean_token_accuracy": 0.7589856326580048,
      "num_tokens": 276398989.0,
      "step": 1580
    },
    {
      "epoch": 0.43332368120731063,
      "grad_norm": 0.916355550289154,
      "learning_rate": 4.039190897597978e-05,
      "loss": 0.9217,
      "mean_token_accuracy": 0.7648066570982337,
      "num_tokens": 278192884.0,
      "step": 1590
    },
    {
      "epoch": 0.4360489873784258,
      "grad_norm": 0.9150972962379456,
      "learning_rate": 4.032167439247085e-05,
      "loss": 0.9223,
      "mean_token_accuracy": 0.7662323093041777,
      "num_tokens": 279937931.0,
      "step": 1600
    },
    {
      "epoch": 0.43877429354954095,
      "grad_norm": 0.7945618629455566,
      "learning_rate": 4.025143980896193e-05,
      "loss": 0.9226,
      "mean_token_accuracy": 0.7670198388397693,
      "num_tokens": 281698143.0,
      "step": 1610
    },
    {
      "epoch": 0.4414995997206561,
      "grad_norm": 0.826994776725769,
      "learning_rate": 4.018120522545302e-05,
      "loss": 0.9348,
      "mean_token_accuracy": 0.7642012946307659,
      "num_tokens": 283412565.0,
      "step": 1620
    },
    {
      "epoch": 0.44422490589177127,
      "grad_norm": 0.9012675881385803,
      "learning_rate": 4.011097064194409e-05,
      "loss": 0.9413,
      "mean_token_accuracy": 0.7626177223399282,
      "num_tokens": 285184244.0,
      "step": 1630
    },
    {
      "epoch": 0.44695021206288643,
      "grad_norm": 0.7998985648155212,
      "learning_rate": 4.004073605843517e-05,
      "loss": 0.8893,
      "mean_token_accuracy": 0.7731544019654393,
      "num_tokens": 286963226.0,
      "step": 1640
    },
    {
      "epoch": 0.4496755182340016,
      "grad_norm": 0.894074559211731,
      "learning_rate": 3.997050147492625e-05,
      "loss": 0.9101,
      "mean_token_accuracy": 0.7684361206367611,
      "num_tokens": 288742524.0,
      "step": 1650
    },
    {
      "epoch": 0.45240082440511675,
      "grad_norm": 0.9063016176223755,
      "learning_rate": 3.990026689141734e-05,
      "loss": 0.9408,
      "mean_token_accuracy": 0.7627444183453918,
      "num_tokens": 290486439.0,
      "step": 1660
    },
    {
      "epoch": 0.4551261305762319,
      "grad_norm": 0.9253380298614502,
      "learning_rate": 3.9830032307908413e-05,
      "loss": 0.9489,
      "mean_token_accuracy": 0.761582453828305,
      "num_tokens": 292244616.0,
      "step": 1670
    },
    {
      "epoch": 0.4578514367473471,
      "grad_norm": 0.908446192741394,
      "learning_rate": 3.9759797724399494e-05,
      "loss": 0.9435,
      "mean_token_accuracy": 0.7627298279665411,
      "num_tokens": 294028863.0,
      "step": 1680
    },
    {
      "epoch": 0.46057674291846223,
      "grad_norm": 0.8264852166175842,
      "learning_rate": 3.968956314089058e-05,
      "loss": 0.9158,
      "mean_token_accuracy": 0.765919309295714,
      "num_tokens": 295842660.0,
      "step": 1690
    },
    {
      "epoch": 0.4633020490895774,
      "grad_norm": 0.8423106074333191,
      "learning_rate": 3.961932855738166e-05,
      "loss": 0.9214,
      "mean_token_accuracy": 0.767298517934978,
      "num_tokens": 297606076.0,
      "step": 1700
    },
    {
      "epoch": 0.46602735526069256,
      "grad_norm": 0.8728176355361938,
      "learning_rate": 3.9549093973872734e-05,
      "loss": 0.9501,
      "mean_token_accuracy": 0.7602741612121463,
      "num_tokens": 299367632.0,
      "step": 1710
    },
    {
      "epoch": 0.4687526614318077,
      "grad_norm": 0.7749947309494019,
      "learning_rate": 3.9478859390363814e-05,
      "loss": 0.9076,
      "mean_token_accuracy": 0.770167650654912,
      "num_tokens": 301097990.0,
      "step": 1720
    },
    {
      "epoch": 0.4714779676029229,
      "grad_norm": 0.9254289865493774,
      "learning_rate": 3.94086248068549e-05,
      "loss": 0.9142,
      "mean_token_accuracy": 0.7686513620428741,
      "num_tokens": 302847308.0,
      "step": 1730
    },
    {
      "epoch": 0.47420327377403804,
      "grad_norm": 0.870252788066864,
      "learning_rate": 3.9338390223345975e-05,
      "loss": 0.9356,
      "mean_token_accuracy": 0.7628345469944179,
      "num_tokens": 304582625.0,
      "step": 1740
    },
    {
      "epoch": 0.4769285799451532,
      "grad_norm": 0.845609188079834,
      "learning_rate": 3.9268155639837055e-05,
      "loss": 0.9179,
      "mean_token_accuracy": 0.7681567031890154,
      "num_tokens": 306358206.0,
      "step": 1750
    },
    {
      "epoch": 0.47965388611626836,
      "grad_norm": 0.8974276185035706,
      "learning_rate": 3.9197921056328135e-05,
      "loss": 0.9172,
      "mean_token_accuracy": 0.7672481311485171,
      "num_tokens": 308026562.0,
      "step": 1760
    },
    {
      "epoch": 0.4823791922873835,
      "grad_norm": 0.9142021536827087,
      "learning_rate": 3.912768647281922e-05,
      "loss": 0.9492,
      "mean_token_accuracy": 0.7608077014796436,
      "num_tokens": 309804543.0,
      "step": 1770
    },
    {
      "epoch": 0.4851044984584987,
      "grad_norm": 0.8676534295082092,
      "learning_rate": 3.9057451889310296e-05,
      "loss": 0.9406,
      "mean_token_accuracy": 0.7620649378746748,
      "num_tokens": 311549771.0,
      "step": 1780
    },
    {
      "epoch": 0.48782980462961384,
      "grad_norm": 0.9465579986572266,
      "learning_rate": 3.8987217305801376e-05,
      "loss": 0.9559,
      "mean_token_accuracy": 0.7597391474992037,
      "num_tokens": 313326889.0,
      "step": 1790
    },
    {
      "epoch": 0.490555110800729,
      "grad_norm": 1.1392822265625,
      "learning_rate": 3.891698272229246e-05,
      "loss": 0.9123,
      "mean_token_accuracy": 0.7693880930542946,
      "num_tokens": 315064715.0,
      "step": 1800
    },
    {
      "epoch": 0.49328041697184416,
      "grad_norm": 0.8468901515007019,
      "learning_rate": 3.884674813878354e-05,
      "loss": 0.9293,
      "mean_token_accuracy": 0.7640517603605985,
      "num_tokens": 316809509.0,
      "step": 1810
    },
    {
      "epoch": 0.4960057231429593,
      "grad_norm": 0.831238865852356,
      "learning_rate": 3.8776513555274616e-05,
      "loss": 0.9097,
      "mean_token_accuracy": 0.7691411901265383,
      "num_tokens": 318597481.0,
      "step": 1820
    },
    {
      "epoch": 0.4987310293140745,
      "grad_norm": 0.8473038077354431,
      "learning_rate": 3.8706278971765697e-05,
      "loss": 0.9099,
      "mean_token_accuracy": 0.7685648757964373,
      "num_tokens": 320382863.0,
      "step": 1830
    },
    {
      "epoch": 0.5014563354851896,
      "grad_norm": 0.8610597848892212,
      "learning_rate": 3.8636044388256784e-05,
      "loss": 0.9255,
      "mean_token_accuracy": 0.7672168118879199,
      "num_tokens": 322136431.0,
      "step": 1840
    },
    {
      "epoch": 0.5041816416563049,
      "grad_norm": 0.8556321859359741,
      "learning_rate": 3.856580980474786e-05,
      "loss": 0.9108,
      "mean_token_accuracy": 0.7693611716851592,
      "num_tokens": 323892089.0,
      "step": 1850
    },
    {
      "epoch": 0.50690694782742,
      "grad_norm": 0.8000465631484985,
      "learning_rate": 3.849557522123894e-05,
      "loss": 0.9198,
      "mean_token_accuracy": 0.7665763789787888,
      "num_tokens": 325693457.0,
      "step": 1860
    },
    {
      "epoch": 0.5096322539985352,
      "grad_norm": 0.9535455107688904,
      "learning_rate": 3.842534063773002e-05,
      "loss": 0.8953,
      "mean_token_accuracy": 0.7704567154869437,
      "num_tokens": 327397884.0,
      "step": 1870
    },
    {
      "epoch": 0.5123575601696503,
      "grad_norm": 0.9103294610977173,
      "learning_rate": 3.8355106054221104e-05,
      "loss": 0.9415,
      "mean_token_accuracy": 0.7628723161295057,
      "num_tokens": 329077502.0,
      "step": 1880
    },
    {
      "epoch": 0.5150828663407655,
      "grad_norm": 0.8798259496688843,
      "learning_rate": 3.828487147071218e-05,
      "loss": 0.9294,
      "mean_token_accuracy": 0.7647034784778952,
      "num_tokens": 330847695.0,
      "step": 1890
    },
    {
      "epoch": 0.5178081725118806,
      "grad_norm": 0.8437985181808472,
      "learning_rate": 3.821463688720326e-05,
      "loss": 0.9127,
      "mean_token_accuracy": 0.7662377031520009,
      "num_tokens": 332578366.0,
      "step": 1900
    },
    {
      "epoch": 0.5205334786829958,
      "grad_norm": 0.877572774887085,
      "learning_rate": 3.8144402303694345e-05,
      "loss": 0.9395,
      "mean_token_accuracy": 0.7628075636923313,
      "num_tokens": 334293589.0,
      "step": 1910
    },
    {
      "epoch": 0.5232587848541109,
      "grad_norm": 0.7866714596748352,
      "learning_rate": 3.8074167720185425e-05,
      "loss": 0.9268,
      "mean_token_accuracy": 0.7632251624017954,
      "num_tokens": 336016972.0,
      "step": 1920
    },
    {
      "epoch": 0.5259840910252261,
      "grad_norm": 0.8031445741653442,
      "learning_rate": 3.80039331366765e-05,
      "loss": 0.9032,
      "mean_token_accuracy": 0.7699568081647158,
      "num_tokens": 337805977.0,
      "step": 1930
    },
    {
      "epoch": 0.5287093971963412,
      "grad_norm": 0.7914207577705383,
      "learning_rate": 3.793369855316758e-05,
      "loss": 0.8779,
      "mean_token_accuracy": 0.7737740287557244,
      "num_tokens": 339548221.0,
      "step": 1940
    },
    {
      "epoch": 0.5314347033674565,
      "grad_norm": 0.8687313795089722,
      "learning_rate": 3.7863463969658666e-05,
      "loss": 0.9549,
      "mean_token_accuracy": 0.7590207350440323,
      "num_tokens": 341268344.0,
      "step": 1950
    },
    {
      "epoch": 0.5341600095385716,
      "grad_norm": 0.9336804747581482,
      "learning_rate": 3.779322938614974e-05,
      "loss": 0.9221,
      "mean_token_accuracy": 0.765609790198505,
      "num_tokens": 343017522.0,
      "step": 1960
    },
    {
      "epoch": 0.5368853157096868,
      "grad_norm": 0.9002737998962402,
      "learning_rate": 3.772299480264082e-05,
      "loss": 0.9181,
      "mean_token_accuracy": 0.7670193919911981,
      "num_tokens": 344699192.0,
      "step": 1970
    },
    {
      "epoch": 0.5396106218808019,
      "grad_norm": 0.8401851654052734,
      "learning_rate": 3.7652760219131906e-05,
      "loss": 0.9043,
      "mean_token_accuracy": 0.7700289068743587,
      "num_tokens": 346458527.0,
      "step": 1980
    },
    {
      "epoch": 0.5423359280519171,
      "grad_norm": 0.8544691801071167,
      "learning_rate": 3.7582525635622986e-05,
      "loss": 0.9203,
      "mean_token_accuracy": 0.7662555683404207,
      "num_tokens": 348204585.0,
      "step": 1990
    },
    {
      "epoch": 0.5450612342230322,
      "grad_norm": 0.8688130378723145,
      "learning_rate": 3.751229105211406e-05,
      "loss": 0.9467,
      "mean_token_accuracy": 0.7613058963790535,
      "num_tokens": 349928259.0,
      "step": 2000
    },
    {
      "epoch": 0.5477865403941474,
      "grad_norm": 0.856677234172821,
      "learning_rate": 3.744205646860514e-05,
      "loss": 0.9124,
      "mean_token_accuracy": 0.767791903577745,
      "num_tokens": 351568018.0,
      "step": 2010
    },
    {
      "epoch": 0.5505118465652625,
      "grad_norm": 0.8962033987045288,
      "learning_rate": 3.737182188509623e-05,
      "loss": 0.8993,
      "mean_token_accuracy": 0.7711729943752289,
      "num_tokens": 353325716.0,
      "step": 2020
    },
    {
      "epoch": 0.5532371527363777,
      "grad_norm": 0.8988960385322571,
      "learning_rate": 3.730158730158731e-05,
      "loss": 0.9338,
      "mean_token_accuracy": 0.7638613487593829,
      "num_tokens": 355099655.0,
      "step": 2030
    },
    {
      "epoch": 0.5559624589074929,
      "grad_norm": 0.7775430083274841,
      "learning_rate": 3.723135271807838e-05,
      "loss": 0.9109,
      "mean_token_accuracy": 0.7683175005018711,
      "num_tokens": 356892448.0,
      "step": 2040
    },
    {
      "epoch": 0.5586877650786081,
      "grad_norm": 0.8810161352157593,
      "learning_rate": 3.716111813456946e-05,
      "loss": 0.9376,
      "mean_token_accuracy": 0.7628290761262179,
      "num_tokens": 358599855.0,
      "step": 2050
    },
    {
      "epoch": 0.5614130712497232,
      "grad_norm": 0.8723873496055603,
      "learning_rate": 3.709088355106055e-05,
      "loss": 0.9433,
      "mean_token_accuracy": 0.7625115603208542,
      "num_tokens": 360366528.0,
      "step": 2060
    },
    {
      "epoch": 0.5641383774208384,
      "grad_norm": 0.9130221605300903,
      "learning_rate": 3.702064896755162e-05,
      "loss": 0.915,
      "mean_token_accuracy": 0.7676266288384795,
      "num_tokens": 362123903.0,
      "step": 2070
    },
    {
      "epoch": 0.5668636835919535,
      "grad_norm": 0.7740192413330078,
      "learning_rate": 3.69504143840427e-05,
      "loss": 0.8998,
      "mean_token_accuracy": 0.7719444134272635,
      "num_tokens": 363907270.0,
      "step": 2080
    },
    {
      "epoch": 0.5695889897630687,
      "grad_norm": 0.931222677230835,
      "learning_rate": 3.688017980053379e-05,
      "loss": 0.8834,
      "mean_token_accuracy": 0.772805467247963,
      "num_tokens": 365602883.0,
      "step": 2090
    },
    {
      "epoch": 0.5723142959341838,
      "grad_norm": 0.9044320583343506,
      "learning_rate": 3.680994521702487e-05,
      "loss": 0.9081,
      "mean_token_accuracy": 0.7687753444537521,
      "num_tokens": 367308921.0,
      "step": 2100
    },
    {
      "epoch": 0.575039602105299,
      "grad_norm": 0.8750004768371582,
      "learning_rate": 3.673971063351594e-05,
      "loss": 0.909,
      "mean_token_accuracy": 0.7677365321666002,
      "num_tokens": 369068618.0,
      "step": 2110
    },
    {
      "epoch": 0.5777649082764141,
      "grad_norm": 0.836656391620636,
      "learning_rate": 3.666947605000702e-05,
      "loss": 0.9237,
      "mean_token_accuracy": 0.7663044361397624,
      "num_tokens": 370792272.0,
      "step": 2120
    },
    {
      "epoch": 0.5804902144475294,
      "grad_norm": 0.9042032361030579,
      "learning_rate": 3.659924146649811e-05,
      "loss": 0.9282,
      "mean_token_accuracy": 0.7646197730675339,
      "num_tokens": 372506541.0,
      "step": 2130
    },
    {
      "epoch": 0.5832155206186445,
      "grad_norm": 0.8443713784217834,
      "learning_rate": 3.652900688298919e-05,
      "loss": 0.8791,
      "mean_token_accuracy": 0.7750100614503026,
      "num_tokens": 374308542.0,
      "step": 2140
    },
    {
      "epoch": 0.5859408267897597,
      "grad_norm": 0.840281069278717,
      "learning_rate": 3.645877229948026e-05,
      "loss": 0.8876,
      "mean_token_accuracy": 0.7726960835978389,
      "num_tokens": 376016266.0,
      "step": 2150
    },
    {
      "epoch": 0.5886661329608748,
      "grad_norm": 0.80794757604599,
      "learning_rate": 3.638853771597134e-05,
      "loss": 0.9308,
      "mean_token_accuracy": 0.7654405352659523,
      "num_tokens": 377732713.0,
      "step": 2160
    },
    {
      "epoch": 0.59139143913199,
      "grad_norm": 0.9291797280311584,
      "learning_rate": 3.631830313246243e-05,
      "loss": 0.9172,
      "mean_token_accuracy": 0.76725367102772,
      "num_tokens": 379457927.0,
      "step": 2170
    },
    {
      "epoch": 0.5941167453031051,
      "grad_norm": 0.7284071445465088,
      "learning_rate": 3.6248068548953503e-05,
      "loss": 0.9533,
      "mean_token_accuracy": 0.7596265442669392,
      "num_tokens": 381311531.0,
      "step": 2180
    },
    {
      "epoch": 0.5968420514742203,
      "grad_norm": 0.8543251156806946,
      "learning_rate": 3.6177833965444584e-05,
      "loss": 0.9449,
      "mean_token_accuracy": 0.7630881817080081,
      "num_tokens": 383076036.0,
      "step": 2190
    },
    {
      "epoch": 0.5995673576453354,
      "grad_norm": 0.8819230198860168,
      "learning_rate": 3.610759938193567e-05,
      "loss": 0.9396,
      "mean_token_accuracy": 0.7628208780661225,
      "num_tokens": 384869981.0,
      "step": 2200
    },
    {
      "epoch": 0.6022926638164506,
      "grad_norm": 0.9766581058502197,
      "learning_rate": 3.603736479842675e-05,
      "loss": 0.9203,
      "mean_token_accuracy": 0.7657232660800218,
      "num_tokens": 386654361.0,
      "step": 2210
    },
    {
      "epoch": 0.6050179699875657,
      "grad_norm": 0.8702138662338257,
      "learning_rate": 3.5967130214917824e-05,
      "loss": 0.8969,
      "mean_token_accuracy": 0.7711658507585526,
      "num_tokens": 388405388.0,
      "step": 2220
    },
    {
      "epoch": 0.607743276158681,
      "grad_norm": 0.8681634068489075,
      "learning_rate": 3.5896895631408904e-05,
      "loss": 0.928,
      "mean_token_accuracy": 0.7651922093704343,
      "num_tokens": 390160411.0,
      "step": 2230
    },
    {
      "epoch": 0.6104685823297961,
      "grad_norm": 0.8256725668907166,
      "learning_rate": 3.582666104789999e-05,
      "loss": 0.9334,
      "mean_token_accuracy": 0.7633168527856469,
      "num_tokens": 391963393.0,
      "step": 2240
    },
    {
      "epoch": 0.6131938885009113,
      "grad_norm": 0.8524363040924072,
      "learning_rate": 3.575642646439107e-05,
      "loss": 0.9321,
      "mean_token_accuracy": 0.7641467567533254,
      "num_tokens": 393719494.0,
      "step": 2250
    },
    {
      "epoch": 0.6159191946720264,
      "grad_norm": 0.8173193335533142,
      "learning_rate": 3.5686191880882145e-05,
      "loss": 0.93,
      "mean_token_accuracy": 0.7653383906930685,
      "num_tokens": 395419867.0,
      "step": 2260
    },
    {
      "epoch": 0.6186445008431416,
      "grad_norm": 0.7731173634529114,
      "learning_rate": 3.5615957297373225e-05,
      "loss": 0.9061,
      "mean_token_accuracy": 0.7694511445239186,
      "num_tokens": 397164636.0,
      "step": 2270
    },
    {
      "epoch": 0.6213698070142567,
      "grad_norm": 0.9051828980445862,
      "learning_rate": 3.554572271386431e-05,
      "loss": 0.9409,
      "mean_token_accuracy": 0.761752724274993,
      "num_tokens": 398897880.0,
      "step": 2280
    },
    {
      "epoch": 0.6240951131853719,
      "grad_norm": 0.9036754965782166,
      "learning_rate": 3.5475488130355386e-05,
      "loss": 0.9232,
      "mean_token_accuracy": 0.7656918304041028,
      "num_tokens": 400668616.0,
      "step": 2290
    },
    {
      "epoch": 0.6268204193564871,
      "grad_norm": 0.8142061233520508,
      "learning_rate": 3.5405253546846466e-05,
      "loss": 0.915,
      "mean_token_accuracy": 0.7672401297837496,
      "num_tokens": 402379710.0,
      "step": 2300
    },
    {
      "epoch": 0.6295457255276022,
      "grad_norm": 0.8594102263450623,
      "learning_rate": 3.533501896333755e-05,
      "loss": 0.9119,
      "mean_token_accuracy": 0.7676404159516096,
      "num_tokens": 404084345.0,
      "step": 2310
    },
    {
      "epoch": 0.6322710316987175,
      "grad_norm": 0.8828473091125488,
      "learning_rate": 3.526478437982863e-05,
      "loss": 0.8836,
      "mean_token_accuracy": 0.7746170441620052,
      "num_tokens": 405770321.0,
      "step": 2320
    },
    {
      "epoch": 0.6349963378698326,
      "grad_norm": 0.8342154026031494,
      "learning_rate": 3.5194549796319706e-05,
      "loss": 0.9064,
      "mean_token_accuracy": 0.7688289485871792,
      "num_tokens": 407492020.0,
      "step": 2330
    },
    {
      "epoch": 0.6377216440409478,
      "grad_norm": 0.8321424722671509,
      "learning_rate": 3.5124315212810787e-05,
      "loss": 0.9023,
      "mean_token_accuracy": 0.7712408562190831,
      "num_tokens": 409257115.0,
      "step": 2340
    },
    {
      "epoch": 0.6404469502120629,
      "grad_norm": 0.8814114332199097,
      "learning_rate": 3.5054080629301874e-05,
      "loss": 0.9279,
      "mean_token_accuracy": 0.7646240370348096,
      "num_tokens": 410986071.0,
      "step": 2350
    },
    {
      "epoch": 0.6431722563831781,
      "grad_norm": 0.8669275045394897,
      "learning_rate": 3.4983846045792954e-05,
      "loss": 0.9115,
      "mean_token_accuracy": 0.7685579599812626,
      "num_tokens": 412763897.0,
      "step": 2360
    },
    {
      "epoch": 0.6458975625542932,
      "grad_norm": 0.8246127367019653,
      "learning_rate": 3.491361146228403e-05,
      "loss": 0.8796,
      "mean_token_accuracy": 0.7756225952878595,
      "num_tokens": 414503168.0,
      "step": 2370
    },
    {
      "epoch": 0.6486228687254084,
      "grad_norm": 0.811538577079773,
      "learning_rate": 3.484337687877511e-05,
      "loss": 0.9046,
      "mean_token_accuracy": 0.7706884844228625,
      "num_tokens": 416227145.0,
      "step": 2380
    },
    {
      "epoch": 0.6513481748965235,
      "grad_norm": 0.8389571309089661,
      "learning_rate": 3.4773142295266194e-05,
      "loss": 0.9023,
      "mean_token_accuracy": 0.7694288929924369,
      "num_tokens": 417973040.0,
      "step": 2390
    },
    {
      "epoch": 0.6540734810676387,
      "grad_norm": 0.9339456558227539,
      "learning_rate": 3.470290771175727e-05,
      "loss": 0.9104,
      "mean_token_accuracy": 0.7684263320639729,
      "num_tokens": 419736780.0,
      "step": 2400
    },
    {
      "epoch": 0.6567987872387538,
      "grad_norm": 0.7914407253265381,
      "learning_rate": 3.463267312824835e-05,
      "loss": 0.9007,
      "mean_token_accuracy": 0.7712036456912756,
      "num_tokens": 421516023.0,
      "step": 2410
    },
    {
      "epoch": 0.6595240934098691,
      "grad_norm": 0.8769999146461487,
      "learning_rate": 3.4562438544739435e-05,
      "loss": 0.9433,
      "mean_token_accuracy": 0.7634194398298859,
      "num_tokens": 423324296.0,
      "step": 2420
    },
    {
      "epoch": 0.6622493995809842,
      "grad_norm": 0.9201934337615967,
      "learning_rate": 3.4492203961230515e-05,
      "loss": 0.9272,
      "mean_token_accuracy": 0.7628090158104897,
      "num_tokens": 425114987.0,
      "step": 2430
    },
    {
      "epoch": 0.6649747057520994,
      "grad_norm": 0.833123505115509,
      "learning_rate": 3.442196937772159e-05,
      "loss": 0.8962,
      "mean_token_accuracy": 0.7704461555927992,
      "num_tokens": 426875751.0,
      "step": 2440
    },
    {
      "epoch": 0.6677000119232145,
      "grad_norm": 0.84730464220047,
      "learning_rate": 3.435173479421267e-05,
      "loss": 0.9062,
      "mean_token_accuracy": 0.7695771548897028,
      "num_tokens": 428715433.0,
      "step": 2450
    },
    {
      "epoch": 0.6704253180943297,
      "grad_norm": 0.9411203265190125,
      "learning_rate": 3.4281500210703756e-05,
      "loss": 0.9435,
      "mean_token_accuracy": 0.761421543173492,
      "num_tokens": 430443301.0,
      "step": 2460
    },
    {
      "epoch": 0.6731506242654448,
      "grad_norm": 0.8246670365333557,
      "learning_rate": 3.4211265627194836e-05,
      "loss": 0.8701,
      "mean_token_accuracy": 0.7766519106924534,
      "num_tokens": 432149898.0,
      "step": 2470
    },
    {
      "epoch": 0.67587593043656,
      "grad_norm": 0.8873474597930908,
      "learning_rate": 3.414103104368591e-05,
      "loss": 0.8906,
      "mean_token_accuracy": 0.7729644425213337,
      "num_tokens": 433904775.0,
      "step": 2480
    },
    {
      "epoch": 0.6786012366076751,
      "grad_norm": 0.8719823956489563,
      "learning_rate": 3.407079646017699e-05,
      "loss": 0.8915,
      "mean_token_accuracy": 0.7718378273770213,
      "num_tokens": 435648944.0,
      "step": 2490
    },
    {
      "epoch": 0.6813265427787903,
      "grad_norm": 0.8626745343208313,
      "learning_rate": 3.4000561876668076e-05,
      "loss": 0.9286,
      "mean_token_accuracy": 0.7654551040381193,
      "num_tokens": 437390122.0,
      "step": 2500
    },
    {
      "epoch": 0.6840518489499054,
      "grad_norm": 0.8651098608970642,
      "learning_rate": 3.393032729315915e-05,
      "loss": 0.9176,
      "mean_token_accuracy": 0.767458095960319,
      "num_tokens": 439126311.0,
      "step": 2510
    },
    {
      "epoch": 0.6867771551210207,
      "grad_norm": 0.82646644115448,
      "learning_rate": 3.386009270965023e-05,
      "loss": 0.9056,
      "mean_token_accuracy": 0.7701940412633121,
      "num_tokens": 440890933.0,
      "step": 2520
    },
    {
      "epoch": 0.6895024612921358,
      "grad_norm": 0.9855936169624329,
      "learning_rate": 3.378985812614132e-05,
      "loss": 0.8948,
      "mean_token_accuracy": 0.7717040576040745,
      "num_tokens": 442611770.0,
      "step": 2530
    },
    {
      "epoch": 0.692227767463251,
      "grad_norm": 0.8282904624938965,
      "learning_rate": 3.37196235426324e-05,
      "loss": 0.8981,
      "mean_token_accuracy": 0.7711870735511184,
      "num_tokens": 444358166.0,
      "step": 2540
    },
    {
      "epoch": 0.6949530736343661,
      "grad_norm": 0.8497506380081177,
      "learning_rate": 3.364938895912347e-05,
      "loss": 0.9359,
      "mean_token_accuracy": 0.7633687354624271,
      "num_tokens": 446078790.0,
      "step": 2550
    },
    {
      "epoch": 0.6976783798054813,
      "grad_norm": 0.9366584420204163,
      "learning_rate": 3.357915437561455e-05,
      "loss": 0.9024,
      "mean_token_accuracy": 0.769352519325912,
      "num_tokens": 447811186.0,
      "step": 2560
    },
    {
      "epoch": 0.7004036859765964,
      "grad_norm": 0.8402838110923767,
      "learning_rate": 3.350891979210564e-05,
      "loss": 0.9157,
      "mean_token_accuracy": 0.7674468743614853,
      "num_tokens": 449537962.0,
      "step": 2570
    },
    {
      "epoch": 0.7031289921477116,
      "grad_norm": 0.8306329250335693,
      "learning_rate": 3.343868520859672e-05,
      "loss": 0.8868,
      "mean_token_accuracy": 0.7735649706795812,
      "num_tokens": 451259469.0,
      "step": 2580
    },
    {
      "epoch": 0.7058542983188267,
      "grad_norm": 0.8826439380645752,
      "learning_rate": 3.336845062508779e-05,
      "loss": 0.8963,
      "mean_token_accuracy": 0.7722868470475077,
      "num_tokens": 453041146.0,
      "step": 2590
    },
    {
      "epoch": 0.708579604489942,
      "grad_norm": 0.8931313157081604,
      "learning_rate": 3.329821604157887e-05,
      "loss": 0.9309,
      "mean_token_accuracy": 0.7645103015005589,
      "num_tokens": 454753651.0,
      "step": 2600
    },
    {
      "epoch": 0.711304910661057,
      "grad_norm": 0.8628774285316467,
      "learning_rate": 3.322798145806996e-05,
      "loss": 0.8983,
      "mean_token_accuracy": 0.7708307450637222,
      "num_tokens": 456537057.0,
      "step": 2610
    },
    {
      "epoch": 0.7140302168321723,
      "grad_norm": 0.8462346196174622,
      "learning_rate": 3.315774687456103e-05,
      "loss": 0.9424,
      "mean_token_accuracy": 0.7633946686983109,
      "num_tokens": 458271296.0,
      "step": 2620
    },
    {
      "epoch": 0.7167555230032874,
      "grad_norm": 0.7972195744514465,
      "learning_rate": 3.308751229105211e-05,
      "loss": 0.9151,
      "mean_token_accuracy": 0.7673410098999739,
      "num_tokens": 460014932.0,
      "step": 2630
    },
    {
      "epoch": 0.7194808291744026,
      "grad_norm": 0.912936270236969,
      "learning_rate": 3.30172777075432e-05,
      "loss": 0.9266,
      "mean_token_accuracy": 0.7656251262873411,
      "num_tokens": 461818480.0,
      "step": 2640
    },
    {
      "epoch": 0.7222061353455177,
      "grad_norm": 0.9048453569412231,
      "learning_rate": 3.294704312403428e-05,
      "loss": 0.9464,
      "mean_token_accuracy": 0.7605526108294726,
      "num_tokens": 463519993.0,
      "step": 2650
    },
    {
      "epoch": 0.7249314415166329,
      "grad_norm": 0.947562038898468,
      "learning_rate": 3.287680854052535e-05,
      "loss": 0.9135,
      "mean_token_accuracy": 0.7671187071129679,
      "num_tokens": 465333101.0,
      "step": 2660
    },
    {
      "epoch": 0.727656747687748,
      "grad_norm": 0.8569560050964355,
      "learning_rate": 3.280657395701643e-05,
      "loss": 0.9084,
      "mean_token_accuracy": 0.76853600833565,
      "num_tokens": 467047198.0,
      "step": 2670
    },
    {
      "epoch": 0.7303820538588632,
      "grad_norm": 0.8648985624313354,
      "learning_rate": 3.273633937350752e-05,
      "loss": 0.9213,
      "mean_token_accuracy": 0.766769905667752,
      "num_tokens": 468806711.0,
      "step": 2680
    },
    {
      "epoch": 0.7331073600299783,
      "grad_norm": 0.8183128237724304,
      "learning_rate": 3.26661047899986e-05,
      "loss": 0.8789,
      "mean_token_accuracy": 0.7748562408611178,
      "num_tokens": 470600782.0,
      "step": 2690
    },
    {
      "epoch": 0.7358326662010936,
      "grad_norm": 0.8345265984535217,
      "learning_rate": 3.2595870206489674e-05,
      "loss": 0.9198,
      "mean_token_accuracy": 0.767015983723104,
      "num_tokens": 472349330.0,
      "step": 2700
    },
    {
      "epoch": 0.7385579723722087,
      "grad_norm": 0.836322009563446,
      "learning_rate": 3.2525635622980754e-05,
      "loss": 0.907,
      "mean_token_accuracy": 0.7705251289531588,
      "num_tokens": 474106983.0,
      "step": 2710
    },
    {
      "epoch": 0.7412832785433239,
      "grad_norm": 0.8931109309196472,
      "learning_rate": 3.245540103947184e-05,
      "loss": 0.939,
      "mean_token_accuracy": 0.7626953650265932,
      "num_tokens": 475902407.0,
      "step": 2720
    },
    {
      "epoch": 0.744008584714439,
      "grad_norm": 0.8067692518234253,
      "learning_rate": 3.2385166455962914e-05,
      "loss": 0.8629,
      "mean_token_accuracy": 0.7779960019513965,
      "num_tokens": 477654644.0,
      "step": 2730
    },
    {
      "epoch": 0.7467338908855542,
      "grad_norm": 0.8653134107589722,
      "learning_rate": 3.2314931872453994e-05,
      "loss": 0.9292,
      "mean_token_accuracy": 0.764538905583322,
      "num_tokens": 479409078.0,
      "step": 2740
    },
    {
      "epoch": 0.7494591970566693,
      "grad_norm": 0.9657193422317505,
      "learning_rate": 3.224469728894508e-05,
      "loss": 0.9136,
      "mean_token_accuracy": 0.767500625923276,
      "num_tokens": 481154972.0,
      "step": 2750
    },
    {
      "epoch": 0.7521845032277845,
      "grad_norm": 0.9113588929176331,
      "learning_rate": 3.217446270543616e-05,
      "loss": 0.9395,
      "mean_token_accuracy": 0.7628454895690083,
      "num_tokens": 482957554.0,
      "step": 2760
    },
    {
      "epoch": 0.7549098093988996,
      "grad_norm": 0.8665691614151001,
      "learning_rate": 3.2104228121927235e-05,
      "loss": 0.9139,
      "mean_token_accuracy": 0.7677292965352536,
      "num_tokens": 484730289.0,
      "step": 2770
    },
    {
      "epoch": 0.7576351155700148,
      "grad_norm": 0.9618145227432251,
      "learning_rate": 3.2033993538418315e-05,
      "loss": 0.9206,
      "mean_token_accuracy": 0.7647297469899058,
      "num_tokens": 486390658.0,
      "step": 2780
    },
    {
      "epoch": 0.7603604217411299,
      "grad_norm": 0.9165218472480774,
      "learning_rate": 3.19637589549094e-05,
      "loss": 0.8822,
      "mean_token_accuracy": 0.7746361203491687,
      "num_tokens": 488084357.0,
      "step": 2790
    },
    {
      "epoch": 0.7630857279122452,
      "grad_norm": 0.9210852980613708,
      "learning_rate": 3.189352437140048e-05,
      "loss": 0.9589,
      "mean_token_accuracy": 0.7591536954045296,
      "num_tokens": 489867504.0,
      "step": 2800
    },
    {
      "epoch": 0.7658110340833603,
      "grad_norm": 0.8209601044654846,
      "learning_rate": 3.1823289787891556e-05,
      "loss": 0.926,
      "mean_token_accuracy": 0.7646390557289123,
      "num_tokens": 491614120.0,
      "step": 2810
    },
    {
      "epoch": 0.7685363402544755,
      "grad_norm": 0.8096806406974792,
      "learning_rate": 3.1753055204382636e-05,
      "loss": 0.9237,
      "mean_token_accuracy": 0.7666580315679312,
      "num_tokens": 493360108.0,
      "step": 2820
    },
    {
      "epoch": 0.7712616464255906,
      "grad_norm": 0.8601098656654358,
      "learning_rate": 3.168282062087372e-05,
      "loss": 0.9003,
      "mean_token_accuracy": 0.7720298429951071,
      "num_tokens": 495060867.0,
      "step": 2830
    },
    {
      "epoch": 0.7739869525967058,
      "grad_norm": 0.896539568901062,
      "learning_rate": 3.16125860373648e-05,
      "loss": 0.9224,
      "mean_token_accuracy": 0.7646079597994685,
      "num_tokens": 496731129.0,
      "step": 2840
    },
    {
      "epoch": 0.7767122587678209,
      "grad_norm": 0.7847779393196106,
      "learning_rate": 3.1542351453855877e-05,
      "loss": 0.8955,
      "mean_token_accuracy": 0.7723177457228303,
      "num_tokens": 498416929.0,
      "step": 2850
    },
    {
      "epoch": 0.7794375649389361,
      "grad_norm": 0.9052691459655762,
      "learning_rate": 3.1472116870346964e-05,
      "loss": 0.8798,
      "mean_token_accuracy": 0.7750480374321341,
      "num_tokens": 500131030.0,
      "step": 2860
    },
    {
      "epoch": 0.7821628711100512,
      "grad_norm": 0.7930359244346619,
      "learning_rate": 3.1401882286838044e-05,
      "loss": 0.8905,
      "mean_token_accuracy": 0.7722199564799667,
      "num_tokens": 501860254.0,
      "step": 2870
    },
    {
      "epoch": 0.7848881772811664,
      "grad_norm": 0.89003986120224,
      "learning_rate": 3.133164770332912e-05,
      "loss": 0.9411,
      "mean_token_accuracy": 0.7619166776537896,
      "num_tokens": 503599791.0,
      "step": 2880
    },
    {
      "epoch": 0.7876134834522815,
      "grad_norm": 0.8397514224052429,
      "learning_rate": 3.12614131198202e-05,
      "loss": 0.96,
      "mean_token_accuracy": 0.7589967941865325,
      "num_tokens": 505277304.0,
      "step": 2890
    },
    {
      "epoch": 0.7903387896233968,
      "grad_norm": 0.9073812365531921,
      "learning_rate": 3.1191178536311284e-05,
      "loss": 0.9119,
      "mean_token_accuracy": 0.767801726795733,
      "num_tokens": 507010506.0,
      "step": 2900
    },
    {
      "epoch": 0.7930640957945119,
      "grad_norm": 0.8646041750907898,
      "learning_rate": 3.1120943952802364e-05,
      "loss": 0.8999,
      "mean_token_accuracy": 0.7706972565501928,
      "num_tokens": 508717678.0,
      "step": 2910
    },
    {
      "epoch": 0.7957894019656271,
      "grad_norm": 0.8266422748565674,
      "learning_rate": 3.105070936929344e-05,
      "loss": 0.878,
      "mean_token_accuracy": 0.7762510352768004,
      "num_tokens": 510478982.0,
      "step": 2920
    },
    {
      "epoch": 0.7985147081367422,
      "grad_norm": 0.8507207036018372,
      "learning_rate": 3.0980474785784525e-05,
      "loss": 0.8742,
      "mean_token_accuracy": 0.774853196553886,
      "num_tokens": 512223492.0,
      "step": 2930
    },
    {
      "epoch": 0.8012400143078574,
      "grad_norm": 0.8072352409362793,
      "learning_rate": 3.0910240202275605e-05,
      "loss": 0.8996,
      "mean_token_accuracy": 0.770034134015441,
      "num_tokens": 513950826.0,
      "step": 2940
    },
    {
      "epoch": 0.8039653204789725,
      "grad_norm": 0.88204026222229,
      "learning_rate": 3.0840005618766685e-05,
      "loss": 0.9293,
      "mean_token_accuracy": 0.7653471390716732,
      "num_tokens": 515747512.0,
      "step": 2950
    },
    {
      "epoch": 0.8066906266500877,
      "grad_norm": 0.8134139180183411,
      "learning_rate": 3.076977103525776e-05,
      "loss": 0.8995,
      "mean_token_accuracy": 0.7705760426819325,
      "num_tokens": 517564408.0,
      "step": 2960
    },
    {
      "epoch": 0.8094159328212028,
      "grad_norm": 0.810004711151123,
      "learning_rate": 3.0699536451748846e-05,
      "loss": 0.8433,
      "mean_token_accuracy": 0.7816835820674897,
      "num_tokens": 519335113.0,
      "step": 2970
    },
    {
      "epoch": 0.812141238992318,
      "grad_norm": 0.8259313106536865,
      "learning_rate": 3.0629301868239926e-05,
      "loss": 0.8837,
      "mean_token_accuracy": 0.7741564935073256,
      "num_tokens": 521052694.0,
      "step": 2980
    },
    {
      "epoch": 0.8148665451634332,
      "grad_norm": 0.8008886575698853,
      "learning_rate": 3.0559067284731e-05,
      "loss": 0.9214,
      "mean_token_accuracy": 0.7659341244027018,
      "num_tokens": 522838886.0,
      "step": 2990
    },
    {
      "epoch": 0.8175918513345484,
      "grad_norm": 0.7934632897377014,
      "learning_rate": 3.048883270122208e-05,
      "loss": 0.9041,
      "mean_token_accuracy": 0.7696094507351517,
      "num_tokens": 524659688.0,
      "step": 3000
    },
    {
      "epoch": 0.8203171575056635,
      "grad_norm": 0.839688241481781,
      "learning_rate": 3.0418598117713166e-05,
      "loss": 0.8899,
      "mean_token_accuracy": 0.7726705798879265,
      "num_tokens": 526428828.0,
      "step": 3010
    },
    {
      "epoch": 0.8230424636767787,
      "grad_norm": 0.8166800737380981,
      "learning_rate": 3.0348363534204243e-05,
      "loss": 0.9119,
      "mean_token_accuracy": 0.7672878581099212,
      "num_tokens": 528178357.0,
      "step": 3020
    },
    {
      "epoch": 0.8257677698478938,
      "grad_norm": 0.8090632557868958,
      "learning_rate": 3.0278128950695323e-05,
      "loss": 0.8927,
      "mean_token_accuracy": 0.7714125035330653,
      "num_tokens": 529968661.0,
      "step": 3030
    },
    {
      "epoch": 0.828493076019009,
      "grad_norm": 0.9141647219657898,
      "learning_rate": 3.0207894367186407e-05,
      "loss": 0.9022,
      "mean_token_accuracy": 0.7716053297743202,
      "num_tokens": 531735846.0,
      "step": 3040
    },
    {
      "epoch": 0.8312183821901241,
      "grad_norm": 0.8853212594985962,
      "learning_rate": 3.0137659783677484e-05,
      "loss": 0.923,
      "mean_token_accuracy": 0.7664798295125366,
      "num_tokens": 533482605.0,
      "step": 3050
    },
    {
      "epoch": 0.8339436883612393,
      "grad_norm": 0.8621941804885864,
      "learning_rate": 3.0067425200168564e-05,
      "loss": 0.9064,
      "mean_token_accuracy": 0.7699273196980357,
      "num_tokens": 535172117.0,
      "step": 3060
    },
    {
      "epoch": 0.8366689945323545,
      "grad_norm": 0.840312123298645,
      "learning_rate": 2.9997190616659644e-05,
      "loss": 0.8953,
      "mean_token_accuracy": 0.7715860839933157,
      "num_tokens": 536922982.0,
      "step": 3070
    },
    {
      "epoch": 0.8393943007034697,
      "grad_norm": 0.863222062587738,
      "learning_rate": 2.9926956033150728e-05,
      "loss": 0.906,
      "mean_token_accuracy": 0.7695185262709856,
      "num_tokens": 538697561.0,
      "step": 3080
    },
    {
      "epoch": 0.8421196068745849,
      "grad_norm": 0.925609290599823,
      "learning_rate": 2.9856721449641805e-05,
      "loss": 0.9123,
      "mean_token_accuracy": 0.7679164627566933,
      "num_tokens": 540498809.0,
      "step": 3090
    },
    {
      "epoch": 0.8448449130457,
      "grad_norm": 0.8675920367240906,
      "learning_rate": 2.9786486866132885e-05,
      "loss": 0.882,
      "mean_token_accuracy": 0.7736795306205749,
      "num_tokens": 542220573.0,
      "step": 3100
    },
    {
      "epoch": 0.8475702192168152,
      "grad_norm": 0.8793986439704895,
      "learning_rate": 2.971625228262396e-05,
      "loss": 0.9003,
      "mean_token_accuracy": 0.770465992949903,
      "num_tokens": 543995979.0,
      "step": 3110
    },
    {
      "epoch": 0.8502955253879303,
      "grad_norm": 0.8559759259223938,
      "learning_rate": 2.964601769911505e-05,
      "loss": 0.9159,
      "mean_token_accuracy": 0.7657374052330852,
      "num_tokens": 545762046.0,
      "step": 3120
    },
    {
      "epoch": 0.8530208315590455,
      "grad_norm": 0.7674428224563599,
      "learning_rate": 2.9575783115606125e-05,
      "loss": 0.8943,
      "mean_token_accuracy": 0.7716390110552311,
      "num_tokens": 547563120.0,
      "step": 3130
    },
    {
      "epoch": 0.8557461377301606,
      "grad_norm": 0.8808310627937317,
      "learning_rate": 2.9505548532097206e-05,
      "loss": 0.9121,
      "mean_token_accuracy": 0.7677836719900369,
      "num_tokens": 549276605.0,
      "step": 3140
    },
    {
      "epoch": 0.8584714439012758,
      "grad_norm": 0.8529791235923767,
      "learning_rate": 2.943531394858829e-05,
      "loss": 0.901,
      "mean_token_accuracy": 0.7700381241738796,
      "num_tokens": 550991479.0,
      "step": 3150
    },
    {
      "epoch": 0.8611967500723909,
      "grad_norm": 0.8383409976959229,
      "learning_rate": 2.9365079365079366e-05,
      "loss": 0.8921,
      "mean_token_accuracy": 0.7733515216037631,
      "num_tokens": 552777662.0,
      "step": 3160
    },
    {
      "epoch": 0.8639220562435062,
      "grad_norm": 0.9514780640602112,
      "learning_rate": 2.9294844781570446e-05,
      "loss": 0.8801,
      "mean_token_accuracy": 0.7743920419365168,
      "num_tokens": 554539356.0,
      "step": 3170
    },
    {
      "epoch": 0.8666473624146213,
      "grad_norm": 0.8511672019958496,
      "learning_rate": 2.9224610198061526e-05,
      "loss": 0.9443,
      "mean_token_accuracy": 0.7616958959959448,
      "num_tokens": 556212548.0,
      "step": 3180
    },
    {
      "epoch": 0.8693726685857365,
      "grad_norm": 0.8855804800987244,
      "learning_rate": 2.915437561455261e-05,
      "loss": 0.9129,
      "mean_token_accuracy": 0.7673460508696734,
      "num_tokens": 557950461.0,
      "step": 3190
    },
    {
      "epoch": 0.8720979747568516,
      "grad_norm": 0.8707140684127808,
      "learning_rate": 2.9084141031043687e-05,
      "loss": 0.8903,
      "mean_token_accuracy": 0.771921418607235,
      "num_tokens": 559715926.0,
      "step": 3200
    },
    {
      "epoch": 0.8748232809279668,
      "grad_norm": 0.9066351652145386,
      "learning_rate": 2.9013906447534767e-05,
      "loss": 0.9122,
      "mean_token_accuracy": 0.7690841609612107,
      "num_tokens": 561479394.0,
      "step": 3210
    },
    {
      "epoch": 0.8775485870990819,
      "grad_norm": 0.8212366104125977,
      "learning_rate": 2.8943671864025844e-05,
      "loss": 0.9047,
      "mean_token_accuracy": 0.7691477432847023,
      "num_tokens": 563271872.0,
      "step": 3220
    },
    {
      "epoch": 0.8802738932701971,
      "grad_norm": 0.8219311833381653,
      "learning_rate": 2.887343728051693e-05,
      "loss": 0.8994,
      "mean_token_accuracy": 0.7715168165042996,
      "num_tokens": 565002210.0,
      "step": 3230
    },
    {
      "epoch": 0.8829991994413122,
      "grad_norm": 0.898818850517273,
      "learning_rate": 2.8803202697008008e-05,
      "loss": 0.8221,
      "mean_token_accuracy": 0.786996117606759,
      "num_tokens": 566740287.0,
      "step": 3240
    },
    {
      "epoch": 0.8857245056124274,
      "grad_norm": 0.7879354953765869,
      "learning_rate": 2.8732968113499088e-05,
      "loss": 0.9305,
      "mean_token_accuracy": 0.7649352782405913,
      "num_tokens": 568560935.0,
      "step": 3250
    },
    {
      "epoch": 0.8884498117835425,
      "grad_norm": 0.9506646394729614,
      "learning_rate": 2.866273352999017e-05,
      "loss": 0.9164,
      "mean_token_accuracy": 0.7674483682960271,
      "num_tokens": 570290159.0,
      "step": 3260
    },
    {
      "epoch": 0.8911751179546578,
      "grad_norm": 0.8495944142341614,
      "learning_rate": 2.8592498946481248e-05,
      "loss": 0.9218,
      "mean_token_accuracy": 0.76632971893996,
      "num_tokens": 572076704.0,
      "step": 3270
    },
    {
      "epoch": 0.8939004241257729,
      "grad_norm": 0.8082846999168396,
      "learning_rate": 2.852226436297233e-05,
      "loss": 0.9318,
      "mean_token_accuracy": 0.76468366086483,
      "num_tokens": 573841975.0,
      "step": 3280
    },
    {
      "epoch": 0.8966257302968881,
      "grad_norm": 0.809038519859314,
      "learning_rate": 2.845202977946341e-05,
      "loss": 0.9057,
      "mean_token_accuracy": 0.7704976342618466,
      "num_tokens": 575587035.0,
      "step": 3290
    },
    {
      "epoch": 0.8993510364680032,
      "grad_norm": 0.8163307309150696,
      "learning_rate": 2.8381795195954492e-05,
      "loss": 0.9047,
      "mean_token_accuracy": 0.769212381914258,
      "num_tokens": 577285852.0,
      "step": 3300
    },
    {
      "epoch": 0.9020763426391184,
      "grad_norm": 0.8320499658584595,
      "learning_rate": 2.831156061244557e-05,
      "loss": 0.9208,
      "mean_token_accuracy": 0.7671002107672393,
      "num_tokens": 579085053.0,
      "step": 3310
    },
    {
      "epoch": 0.9048016488102335,
      "grad_norm": 0.7883128523826599,
      "learning_rate": 2.824132602893665e-05,
      "loss": 0.8903,
      "mean_token_accuracy": 0.7726314144209028,
      "num_tokens": 580874041.0,
      "step": 3320
    },
    {
      "epoch": 0.9075269549813487,
      "grad_norm": 0.8674151301383972,
      "learning_rate": 2.8171091445427726e-05,
      "loss": 0.8888,
      "mean_token_accuracy": 0.7714107790961862,
      "num_tokens": 582616757.0,
      "step": 3330
    },
    {
      "epoch": 0.9102522611524638,
      "grad_norm": 0.8607017397880554,
      "learning_rate": 2.8100856861918813e-05,
      "loss": 0.8852,
      "mean_token_accuracy": 0.7738766206428409,
      "num_tokens": 584370810.0,
      "step": 3340
    },
    {
      "epoch": 0.912977567323579,
      "grad_norm": 0.850727915763855,
      "learning_rate": 2.803062227840989e-05,
      "loss": 0.9006,
      "mean_token_accuracy": 0.7707684559747576,
      "num_tokens": 586143330.0,
      "step": 3350
    },
    {
      "epoch": 0.9157028734946941,
      "grad_norm": 0.8423539400100708,
      "learning_rate": 2.796038769490097e-05,
      "loss": 0.8657,
      "mean_token_accuracy": 0.777682445757091,
      "num_tokens": 587851884.0,
      "step": 3360
    },
    {
      "epoch": 0.9184281796658094,
      "grad_norm": 0.8592097759246826,
      "learning_rate": 2.7890153111392054e-05,
      "loss": 0.8937,
      "mean_token_accuracy": 0.7706205818802119,
      "num_tokens": 589601267.0,
      "step": 3370
    },
    {
      "epoch": 0.9211534858369245,
      "grad_norm": 0.943073034286499,
      "learning_rate": 2.7819918527883134e-05,
      "loss": 0.9254,
      "mean_token_accuracy": 0.7644982000812888,
      "num_tokens": 591321205.0,
      "step": 3380
    },
    {
      "epoch": 0.9238787920080397,
      "grad_norm": 0.7838475704193115,
      "learning_rate": 2.774968394437421e-05,
      "loss": 0.8704,
      "mean_token_accuracy": 0.7773344025947153,
      "num_tokens": 593110103.0,
      "step": 3390
    },
    {
      "epoch": 0.9266040981791548,
      "grad_norm": 0.8899914622306824,
      "learning_rate": 2.767944936086529e-05,
      "loss": 0.9234,
      "mean_token_accuracy": 0.7661155204288661,
      "num_tokens": 594823039.0,
      "step": 3400
    },
    {
      "epoch": 0.92932940435027,
      "grad_norm": 0.8291741609573364,
      "learning_rate": 2.7609214777356374e-05,
      "loss": 0.9049,
      "mean_token_accuracy": 0.7698826493695379,
      "num_tokens": 596625643.0,
      "step": 3410
    },
    {
      "epoch": 0.9320547105213851,
      "grad_norm": 0.8303592205047607,
      "learning_rate": 2.753898019384745e-05,
      "loss": 0.9194,
      "mean_token_accuracy": 0.7646890787407756,
      "num_tokens": 598409404.0,
      "step": 3420
    },
    {
      "epoch": 0.9347800166925003,
      "grad_norm": 1.093499779701233,
      "learning_rate": 2.746874561033853e-05,
      "loss": 0.9067,
      "mean_token_accuracy": 0.7687176914885641,
      "num_tokens": 600131817.0,
      "step": 3430
    },
    {
      "epoch": 0.9375053228636154,
      "grad_norm": 0.7767653465270996,
      "learning_rate": 2.7398511026829608e-05,
      "loss": 0.9087,
      "mean_token_accuracy": 0.7694620480760932,
      "num_tokens": 601867823.0,
      "step": 3440
    },
    {
      "epoch": 0.9402306290347306,
      "grad_norm": 0.8396127223968506,
      "learning_rate": 2.7328276443320695e-05,
      "loss": 0.8845,
      "mean_token_accuracy": 0.7755044038407505,
      "num_tokens": 603591263.0,
      "step": 3450
    },
    {
      "epoch": 0.9429559352058458,
      "grad_norm": 0.8479565382003784,
      "learning_rate": 2.7258041859811772e-05,
      "loss": 0.9151,
      "mean_token_accuracy": 0.7681015016511082,
      "num_tokens": 605400422.0,
      "step": 3460
    },
    {
      "epoch": 0.945681241376961,
      "grad_norm": 0.8595352172851562,
      "learning_rate": 2.7187807276302852e-05,
      "loss": 0.9071,
      "mean_token_accuracy": 0.7697312761098146,
      "num_tokens": 607180330.0,
      "step": 3470
    },
    {
      "epoch": 0.9484065475480761,
      "grad_norm": 0.7876641154289246,
      "learning_rate": 2.7117572692793936e-05,
      "loss": 0.8734,
      "mean_token_accuracy": 0.7765149438753725,
      "num_tokens": 608937194.0,
      "step": 3480
    },
    {
      "epoch": 0.9511318537191913,
      "grad_norm": 0.8947613835334778,
      "learning_rate": 2.7047338109285016e-05,
      "loss": 0.9133,
      "mean_token_accuracy": 0.7668988397344947,
      "num_tokens": 610676168.0,
      "step": 3490
    },
    {
      "epoch": 0.9538571598903064,
      "grad_norm": 0.8162944912910461,
      "learning_rate": 2.6977103525776093e-05,
      "loss": 0.8681,
      "mean_token_accuracy": 0.7774568162858486,
      "num_tokens": 612491555.0,
      "step": 3500
    },
    {
      "epoch": 0.9565824660614216,
      "grad_norm": 0.900248110294342,
      "learning_rate": 2.6906868942267173e-05,
      "loss": 0.9189,
      "mean_token_accuracy": 0.7660327769815922,
      "num_tokens": 614310820.0,
      "step": 3510
    },
    {
      "epoch": 0.9593077722325367,
      "grad_norm": 0.8359858989715576,
      "learning_rate": 2.6836634358758256e-05,
      "loss": 0.9224,
      "mean_token_accuracy": 0.7663468400016427,
      "num_tokens": 616071087.0,
      "step": 3520
    },
    {
      "epoch": 0.9620330784036519,
      "grad_norm": 0.872882604598999,
      "learning_rate": 2.6766399775249333e-05,
      "loss": 0.8712,
      "mean_token_accuracy": 0.7756553223356605,
      "num_tokens": 617769166.0,
      "step": 3530
    },
    {
      "epoch": 0.964758384574767,
      "grad_norm": 0.8891421556472778,
      "learning_rate": 2.6696165191740413e-05,
      "loss": 0.9005,
      "mean_token_accuracy": 0.7701712142676115,
      "num_tokens": 619538232.0,
      "step": 3540
    },
    {
      "epoch": 0.9674836907458823,
      "grad_norm": 0.7362189888954163,
      "learning_rate": 2.662593060823149e-05,
      "loss": 0.8938,
      "mean_token_accuracy": 0.7728794073686004,
      "num_tokens": 621303314.0,
      "step": 3550
    },
    {
      "epoch": 0.9702089969169974,
      "grad_norm": 0.7591912150382996,
      "learning_rate": 2.6555696024722577e-05,
      "loss": 0.8785,
      "mean_token_accuracy": 0.7754427226260304,
      "num_tokens": 623100776.0,
      "step": 3560
    },
    {
      "epoch": 0.9729343030881126,
      "grad_norm": 0.8404043912887573,
      "learning_rate": 2.6485461441213654e-05,
      "loss": 0.9387,
      "mean_token_accuracy": 0.7616929830983281,
      "num_tokens": 624821413.0,
      "step": 3570
    },
    {
      "epoch": 0.9756596092592277,
      "grad_norm": 0.8890523314476013,
      "learning_rate": 2.6415226857704734e-05,
      "loss": 0.9095,
      "mean_token_accuracy": 0.7678571032360196,
      "num_tokens": 626628545.0,
      "step": 3580
    },
    {
      "epoch": 0.9783849154303429,
      "grad_norm": 0.7919222712516785,
      "learning_rate": 2.6344992274195818e-05,
      "loss": 0.8906,
      "mean_token_accuracy": 0.7729906640946865,
      "num_tokens": 628425036.0,
      "step": 3590
    },
    {
      "epoch": 0.981110221601458,
      "grad_norm": 0.7847779393196106,
      "learning_rate": 2.6274757690686898e-05,
      "loss": 0.8636,
      "mean_token_accuracy": 0.7760783301666379,
      "num_tokens": 630092739.0,
      "step": 3600
    },
    {
      "epoch": 0.9838355277725732,
      "grad_norm": 0.964148998260498,
      "learning_rate": 2.6204523107177975e-05,
      "loss": 0.915,
      "mean_token_accuracy": 0.7690061263740062,
      "num_tokens": 631791605.0,
      "step": 3610
    },
    {
      "epoch": 0.9865608339436883,
      "grad_norm": 0.9110335111618042,
      "learning_rate": 2.6134288523669055e-05,
      "loss": 0.9026,
      "mean_token_accuracy": 0.7687728941440582,
      "num_tokens": 633554055.0,
      "step": 3620
    },
    {
      "epoch": 0.9892861401148035,
      "grad_norm": 0.8210707902908325,
      "learning_rate": 2.606405394016014e-05,
      "loss": 0.8729,
      "mean_token_accuracy": 0.7756300155073405,
      "num_tokens": 635269789.0,
      "step": 3630
    },
    {
      "epoch": 0.9920114462859186,
      "grad_norm": 0.8379694819450378,
      "learning_rate": 2.5993819356651215e-05,
      "loss": 0.8912,
      "mean_token_accuracy": 0.7726977992802858,
      "num_tokens": 637078092.0,
      "step": 3640
    },
    {
      "epoch": 0.9947367524570339,
      "grad_norm": 0.867736279964447,
      "learning_rate": 2.5923584773142296e-05,
      "loss": 0.8709,
      "mean_token_accuracy": 0.7778057091869414,
      "num_tokens": 638813927.0,
      "step": 3650
    },
    {
      "epoch": 0.997462058628149,
      "grad_norm": 0.9498102068901062,
      "learning_rate": 2.5853350189633372e-05,
      "loss": 0.8953,
      "mean_token_accuracy": 0.7722262794151902,
      "num_tokens": 640594007.0,
      "step": 3660
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.4446529746055603,
      "learning_rate": 2.578311560612446e-05,
      "loss": 0.8011,
      "mean_token_accuracy": 0.7779095188883327,
      "num_tokens": 642209396.0,
      "step": 3670
    }
  ],
  "logging_steps": 10,
  "max_steps": 7340,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 500.0,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.1108118326981394e+19,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}