{
  "best_global_step": 6948,
  "best_metric": 5.525067329406738,
  "best_model_checkpoint": "./output/checkpoint-6948",
  "epoch": 4.0,
  "eval_steps": 500,
  "global_step": 6948,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "entropy": 3.606692385673523,
      "epoch": 0.028785261945883708,
      "grad_norm": 3.2999913692474365,
      "learning_rate": 4.9e-07,
      "loss": 13.6598,
      "mean_token_accuracy": 0.16028020828962325,
      "num_tokens": 53993.0,
      "step": 50
    },
    {
      "entropy": 3.618675880432129,
      "epoch": 0.057570523891767415,
      "grad_norm": 3.101252555847168,
      "learning_rate": 9.9e-07,
      "loss": 14.0188,
      "mean_token_accuracy": 0.1508466500043869,
      "num_tokens": 110134.0,
      "step": 100
    },
    {
      "entropy": 3.5215235900878907,
      "epoch": 0.08635578583765112,
      "grad_norm": 3.513662815093994,
      "learning_rate": 1.49e-06,
      "loss": 12.8555,
      "mean_token_accuracy": 0.18527640983462335,
      "num_tokens": 160191.0,
      "step": 150
    },
    {
      "entropy": 3.667909698486328,
      "epoch": 0.11514104778353483,
      "grad_norm": 4.327610492706299,
      "learning_rate": 1.99e-06,
      "loss": 13.5394,
      "mean_token_accuracy": 0.157139780074358,
      "num_tokens": 214993.0,
      "step": 200
    },
    {
      "entropy": 3.768263258934021,
      "epoch": 0.14392630972941853,
      "grad_norm": 4.290107250213623,
      "learning_rate": 1.988450206246317e-06,
      "loss": 12.8912,
      "mean_token_accuracy": 0.17374794125556947,
      "num_tokens": 268184.0,
      "step": 250
    },
    {
      "entropy": 3.990619196891785,
      "epoch": 0.17271157167530224,
      "grad_norm": 4.444278717041016,
      "learning_rate": 1.976664702416028e-06,
      "loss": 12.455,
      "mean_token_accuracy": 0.17780130118131637,
      "num_tokens": 319458.0,
      "step": 300
    },
    {
      "entropy": 4.162646284103394,
      "epoch": 0.20149683362118595,
      "grad_norm": 5.615262508392334,
      "learning_rate": 1.9648791985857395e-06,
      "loss": 12.0893,
      "mean_token_accuracy": 0.18191319867968558,
      "num_tokens": 373337.0,
      "step": 350
    },
    {
      "entropy": 4.532100868225098,
      "epoch": 0.23028209556706966,
      "grad_norm": 10.074016571044922,
      "learning_rate": 1.9530936947554507e-06,
      "loss": 11.9261,
      "mean_token_accuracy": 0.169477596282959,
      "num_tokens": 427526.0,
      "step": 400
    },
    {
      "entropy": 4.923871030807495,
      "epoch": 0.25906735751295334,
      "grad_norm": 16.220163345336914,
      "learning_rate": 1.9413081909251622e-06,
      "loss": 11.0048,
      "mean_token_accuracy": 0.1704501649737358,
      "num_tokens": 480528.0,
      "step": 450
    },
    {
      "entropy": 5.521005854606629,
      "epoch": 0.28785261945883706,
      "grad_norm": 29.904008865356445,
      "learning_rate": 1.9295226870948733e-06,
      "loss": 9.6524,
      "mean_token_accuracy": 0.16450899541378022,
      "num_tokens": 535314.0,
      "step": 500
    },
    {
      "entropy": 6.092623329162597,
      "epoch": 0.31663788140472077,
      "grad_norm": 17.821575164794922,
      "learning_rate": 1.9177371832645845e-06,
      "loss": 8.1054,
      "mean_token_accuracy": 0.17205011785030366,
      "num_tokens": 588410.0,
      "step": 550
    },
    {
      "entropy": 6.385262680053711,
      "epoch": 0.3454231433506045,
      "grad_norm": 5.502202987670898,
      "learning_rate": 1.9059516794342958e-06,
      "loss": 7.4313,
      "mean_token_accuracy": 0.1734227080643177,
      "num_tokens": 641736.0,
      "step": 600
    },
    {
      "entropy": 6.278562617301941,
      "epoch": 0.3742084052964882,
      "grad_norm": 5.4657697677612305,
      "learning_rate": 1.8941661756040071e-06,
      "loss": 6.9266,
      "mean_token_accuracy": 0.18680249139666558,
      "num_tokens": 692200.0,
      "step": 650
    },
    {
      "entropy": 6.553266277313233,
      "epoch": 0.4029936672423719,
      "grad_norm": 4.955812931060791,
      "learning_rate": 1.8823806717737183e-06,
      "loss": 6.9847,
      "mean_token_accuracy": 0.16679802387952805,
      "num_tokens": 745830.0,
      "step": 700
    },
    {
      "entropy": 6.470935583114624,
      "epoch": 0.4317789291882556,
      "grad_norm": 4.198381423950195,
      "learning_rate": 1.8705951679434296e-06,
      "loss": 6.7277,
      "mean_token_accuracy": 0.17847734570503235,
      "num_tokens": 798872.0,
      "step": 750
    },
    {
      "entropy": 6.5620588779449465,
      "epoch": 0.4605641911341393,
      "grad_norm": 3.1793746948242188,
      "learning_rate": 1.8588096641131407e-06,
      "loss": 6.7032,
      "mean_token_accuracy": 0.17336134731769562,
      "num_tokens": 853045.0,
      "step": 800
    },
    {
      "entropy": 6.532204885482788,
      "epoch": 0.48934945308002303,
      "grad_norm": 3.824537515640259,
      "learning_rate": 1.847024160282852e-06,
      "loss": 6.5762,
      "mean_token_accuracy": 0.1805124071240425,
      "num_tokens": 907679.0,
      "step": 850
    },
    {
      "entropy": 6.535988225936889,
      "epoch": 0.5181347150259067,
      "grad_norm": 4.350001811981201,
      "learning_rate": 1.8352386564525632e-06,
      "loss": 6.505,
      "mean_token_accuracy": 0.1842605724930763,
      "num_tokens": 964170.0,
      "step": 900
    },
    {
      "entropy": 6.204533562660218,
      "epoch": 0.5469199769717904,
      "grad_norm": 2.193660020828247,
      "learning_rate": 1.8234531526222745e-06,
      "loss": 6.1211,
      "mean_token_accuracy": 0.21968430042266845,
      "num_tokens": 1015909.0,
      "step": 950
    },
    {
      "entropy": 6.308737449645996,
      "epoch": 0.5757052389176741,
      "grad_norm": 2.325622320175171,
      "learning_rate": 1.8116676487919857e-06,
      "loss": 6.1653,
      "mean_token_accuracy": 0.21636426240205764,
      "num_tokens": 1068859.0,
      "step": 1000
    },
    {
      "entropy": 6.332560749053955,
      "epoch": 0.6044905008635578,
      "grad_norm": 2.0439090728759766,
      "learning_rate": 1.799882144961697e-06,
      "loss": 6.1559,
      "mean_token_accuracy": 0.21859725564718246,
      "num_tokens": 1123202.0,
      "step": 1050
    },
    {
      "entropy": 6.042124252319336,
      "epoch": 0.6332757628094415,
      "grad_norm": 3.621903657913208,
      "learning_rate": 1.7880966411314081e-06,
      "loss": 5.8441,
      "mean_token_accuracy": 0.24906315237283708,
      "num_tokens": 1173403.0,
      "step": 1100
    },
    {
      "entropy": 5.921343173980713,
      "epoch": 0.6620610247553252,
      "grad_norm": 5.658033847808838,
      "learning_rate": 1.7763111373011195e-06,
      "loss": 5.7104,
      "mean_token_accuracy": 0.2625067520141602,
      "num_tokens": 1225026.0,
      "step": 1150
    },
    {
      "entropy": 6.093586492538452,
      "epoch": 0.690846286701209,
      "grad_norm": 2.4292995929718018,
      "learning_rate": 1.7645256334708308e-06,
      "loss": 5.8658,
      "mean_token_accuracy": 0.24842385441064835,
      "num_tokens": 1279013.0,
      "step": 1200
    },
    {
      "entropy": 6.119112596511841,
      "epoch": 0.7196315486470927,
      "grad_norm": 3.369384288787842,
      "learning_rate": 1.752740129640542e-06,
      "loss": 5.8784,
      "mean_token_accuracy": 0.24857850253582,
      "num_tokens": 1332547.0,
      "step": 1250
    },
    {
      "entropy": 6.025163550376892,
      "epoch": 0.7484168105929764,
      "grad_norm": 2.5110116004943848,
      "learning_rate": 1.7409546258102533e-06,
      "loss": 5.7769,
      "mean_token_accuracy": 0.25835376888513567,
      "num_tokens": 1385192.0,
      "step": 1300
    },
    {
      "entropy": 5.877259612083435,
      "epoch": 0.7772020725388601,
      "grad_norm": 2.4179303646087646,
      "learning_rate": 1.7291691219799646e-06,
      "loss": 5.6284,
      "mean_token_accuracy": 0.2756252554059029,
      "num_tokens": 1437071.0,
      "step": 1350
    },
    {
      "entropy": 6.002246947288513,
      "epoch": 0.8059873344847438,
      "grad_norm": 3.494359016418457,
      "learning_rate": 1.717383618149676e-06,
      "loss": 5.747,
      "mean_token_accuracy": 0.26462210685014725,
      "num_tokens": 1490818.0,
      "step": 1400
    },
    {
      "entropy": 5.991955623626709,
      "epoch": 0.8347725964306275,
      "grad_norm": 2.340975761413574,
      "learning_rate": 1.705598114319387e-06,
      "loss": 5.7379,
      "mean_token_accuracy": 0.26444981098175047,
      "num_tokens": 1544997.0,
      "step": 1450
    },
    {
      "entropy": 5.91768889427185,
      "epoch": 0.8635578583765112,
      "grad_norm": 2.2394514083862305,
      "learning_rate": 1.6938126104890984e-06,
      "loss": 5.6564,
      "mean_token_accuracy": 0.2730415526032448,
      "num_tokens": 1598302.0,
      "step": 1500
    },
    {
      "entropy": 5.982716989517212,
      "epoch": 0.8923431203223949,
      "grad_norm": 1.876839518547058,
      "learning_rate": 1.6820271066588098e-06,
      "loss": 5.7215,
      "mean_token_accuracy": 0.26642445534467696,
      "num_tokens": 1655267.0,
      "step": 1550
    },
    {
      "entropy": 5.820467872619629,
      "epoch": 0.9211283822682786,
      "grad_norm": 2.219966173171997,
      "learning_rate": 1.6702416028285209e-06,
      "loss": 5.5555,
      "mean_token_accuracy": 0.2856418335437775,
      "num_tokens": 1709199.0,
      "step": 1600
    },
    {
      "entropy": 5.996349005699158,
      "epoch": 0.9499136442141624,
      "grad_norm": 2.247213840484619,
      "learning_rate": 1.6584560989982322e-06,
      "loss": 5.7283,
      "mean_token_accuracy": 0.2696125540137291,
      "num_tokens": 1765443.0,
      "step": 1650
    },
    {
      "entropy": 5.696683068275451,
      "epoch": 0.9786989061600461,
      "grad_norm": 2.8499979972839355,
      "learning_rate": 1.6466705951679433e-06,
      "loss": 5.4335,
      "mean_token_accuracy": 0.29918427973985673,
      "num_tokens": 1817494.0,
      "step": 1700
    },
    {
      "epoch": 1.0,
      "eval_entropy": 5.993559589034401,
      "eval_loss": 5.737204551696777,
      "eval_mean_token_accuracy": 0.2618687468739699,
      "eval_model_preparation_time": 0.0045,
      "eval_num_tokens": 1856362.0,
      "eval_runtime": 50.5332,
      "eval_samples_per_second": 8.588,
      "eval_steps_per_second": 4.294,
      "step": 1737
    },
    {
      "entropy": 5.746842083930969,
      "epoch": 1.0074841681059297,
      "grad_norm": 2.33052921295166,
      "learning_rate": 1.6348850913376547e-06,
      "loss": 5.4796,
      "mean_token_accuracy": 0.2966849410533905,
      "num_tokens": 1870353.0,
      "step": 1750
    },
    {
      "entropy": 5.859029049873352,
      "epoch": 1.0362694300518134,
      "grad_norm": 1.6248886585235596,
      "learning_rate": 1.6230995875073658e-06,
      "loss": 5.5975,
      "mean_token_accuracy": 0.2838129925727844,
      "num_tokens": 1926205.0,
      "step": 1800
    },
    {
      "entropy": 5.731445336341858,
      "epoch": 1.065054691997697,
      "grad_norm": 1.6941566467285156,
      "learning_rate": 1.6113140836770771e-06,
      "loss": 5.476,
      "mean_token_accuracy": 0.2992346465587616,
      "num_tokens": 1979821.0,
      "step": 1850
    },
    {
      "entropy": 5.6993954515457155,
      "epoch": 1.0938399539435808,
      "grad_norm": 1.1746597290039062,
      "learning_rate": 1.5995285798467883e-06,
      "loss": 5.4608,
      "mean_token_accuracy": 0.3000726142525673,
      "num_tokens": 2034373.0,
      "step": 1900
    },
    {
      "entropy": 5.668873124122619,
      "epoch": 1.1226252158894645,
      "grad_norm": 1.728211760520935,
      "learning_rate": 1.5877430760164996e-06,
      "loss": 5.4347,
      "mean_token_accuracy": 0.3033922725915909,
      "num_tokens": 2087339.0,
      "step": 1950
    },
    {
      "entropy": 5.624621086120605,
      "epoch": 1.1514104778353482,
      "grad_norm": 1.4078539609909058,
      "learning_rate": 1.5759575721862107e-06,
      "loss": 5.3954,
      "mean_token_accuracy": 0.30784171640872954,
      "num_tokens": 2139520.0,
      "step": 2000
    },
    {
      "entropy": 5.7141213130950925,
      "epoch": 1.180195739781232,
      "grad_norm": 2.186459541320801,
      "learning_rate": 1.564172068355922e-06,
      "loss": 5.4847,
      "mean_token_accuracy": 0.29594049394130706,
      "num_tokens": 2193987.0,
      "step": 2050
    },
    {
      "entropy": 5.632415266036987,
      "epoch": 1.2089810017271156,
      "grad_norm": 1.3601349592208862,
      "learning_rate": 1.5523865645256334e-06,
      "loss": 5.4135,
      "mean_token_accuracy": 0.30366597563028336,
      "num_tokens": 2249616.0,
      "step": 2100
    },
    {
      "entropy": 5.510904269218445,
      "epoch": 1.2377662636729994,
      "grad_norm": 2.065760612487793,
      "learning_rate": 1.5406010606953445e-06,
      "loss": 5.2904,
      "mean_token_accuracy": 0.3211754837632179,
      "num_tokens": 2300863.0,
      "step": 2150
    },
    {
      "entropy": 5.703383626937867,
      "epoch": 1.266551525618883,
      "grad_norm": 1.1172698736190796,
      "learning_rate": 1.5288155568650559e-06,
      "loss": 5.4802,
      "mean_token_accuracy": 0.29713701367378237,
      "num_tokens": 2356029.0,
      "step": 2200
    },
    {
      "entropy": 5.565930342674255,
      "epoch": 1.2953367875647668,
      "grad_norm": 1.7528513669967651,
      "learning_rate": 1.5170300530347672e-06,
      "loss": 5.3518,
      "mean_token_accuracy": 0.31301232606172563,
      "num_tokens": 2408957.0,
      "step": 2250
    },
    {
      "entropy": 5.496430187225342,
      "epoch": 1.3241220495106505,
      "grad_norm": 1.892640233039856,
      "learning_rate": 1.5052445492044786e-06,
      "loss": 5.2967,
      "mean_token_accuracy": 0.3181899458169937,
      "num_tokens": 2462569.0,
      "step": 2300
    },
    {
      "entropy": 5.725150098800659,
      "epoch": 1.3529073114565342,
      "grad_norm": 1.774940848350525,
      "learning_rate": 1.4934590453741897e-06,
      "loss": 5.5215,
      "mean_token_accuracy": 0.29055028676986694,
      "num_tokens": 2518544.0,
      "step": 2350
    },
    {
      "entropy": 5.4884827613830565,
      "epoch": 1.381692573402418,
      "grad_norm": 2.2167599201202393,
      "learning_rate": 1.481673541543901e-06,
      "loss": 5.2917,
      "mean_token_accuracy": 0.31803421139717103,
      "num_tokens": 2570863.0,
      "step": 2400
    },
    {
      "entropy": 5.697079472541809,
      "epoch": 1.4104778353483016,
      "grad_norm": 1.6489030122756958,
      "learning_rate": 1.4698880377136124e-06,
      "loss": 5.4982,
      "mean_token_accuracy": 0.2925163987278938,
      "num_tokens": 2626998.0,
      "step": 2450
    },
    {
      "entropy": 5.46209939956665,
      "epoch": 1.4392630972941853,
      "grad_norm": 1.153914451599121,
      "learning_rate": 1.4581025338833235e-06,
      "loss": 5.2736,
      "mean_token_accuracy": 0.3182168474793434,
      "num_tokens": 2681568.0,
      "step": 2500
    },
    {
      "entropy": 5.4405768728256225,
      "epoch": 1.468048359240069,
      "grad_norm": 3.6614978313446045,
      "learning_rate": 1.4463170300530348e-06,
      "loss": 5.2515,
      "mean_token_accuracy": 0.3218736210465431,
      "num_tokens": 2733587.0,
      "step": 2550
    },
    {
      "entropy": 5.528175053596496,
      "epoch": 1.4968336211859528,
      "grad_norm": 1.0849746465682983,
      "learning_rate": 1.434531526222746e-06,
      "loss": 5.3378,
      "mean_token_accuracy": 0.31061659604310987,
      "num_tokens": 2787003.0,
      "step": 2600
    },
    {
      "entropy": 5.46110897064209,
      "epoch": 1.5256188831318365,
      "grad_norm": 1.8315683603286743,
      "learning_rate": 1.4227460223924573e-06,
      "loss": 5.2782,
      "mean_token_accuracy": 0.31781029611825945,
      "num_tokens": 2840263.0,
      "step": 2650
    },
    {
      "entropy": 5.455560960769653,
      "epoch": 1.5544041450777202,
      "grad_norm": 1.1859091520309448,
      "learning_rate": 1.4109605185621684e-06,
      "loss": 5.2735,
      "mean_token_accuracy": 0.3194814011454582,
      "num_tokens": 2894186.0,
      "step": 2700
    },
    {
      "entropy": 5.430496115684509,
      "epoch": 1.583189407023604,
      "grad_norm": 2.3500001430511475,
      "learning_rate": 1.3991750147318797e-06,
      "loss": 5.2464,
      "mean_token_accuracy": 0.32140792965888976,
      "num_tokens": 2948171.0,
      "step": 2750
    },
    {
      "entropy": 5.588023023605347,
      "epoch": 1.6119746689694876,
      "grad_norm": 1.727825403213501,
      "learning_rate": 1.3873895109015909e-06,
      "loss": 5.4028,
      "mean_token_accuracy": 0.3039530631899834,
      "num_tokens": 3002678.0,
      "step": 2800
    },
    {
      "entropy": 5.410525422096253,
      "epoch": 1.6407599309153713,
      "grad_norm": 1.3401474952697754,
      "learning_rate": 1.3756040070713022e-06,
      "loss": 5.2298,
      "mean_token_accuracy": 0.324065263569355,
      "num_tokens": 3055844.0,
      "step": 2850
    },
    {
      "entropy": 5.36959942817688,
      "epoch": 1.669545192861255,
      "grad_norm": 1.1892589330673218,
      "learning_rate": 1.3638185032410133e-06,
      "loss": 5.1956,
      "mean_token_accuracy": 0.32639502108097074,
      "num_tokens": 3108636.0,
      "step": 2900
    },
    {
      "entropy": 5.53826907157898,
      "epoch": 1.6983304548071387,
      "grad_norm": 1.2652360200881958,
      "learning_rate": 1.3520329994107247e-06,
      "loss": 5.3583,
      "mean_token_accuracy": 0.3074926760792732,
      "num_tokens": 3162627.0,
      "step": 2950
    },
    {
      "entropy": 5.417449145317078,
      "epoch": 1.7271157167530224,
      "grad_norm": 1.584312915802002,
      "learning_rate": 1.340247495580436e-06,
      "loss": 5.2388,
      "mean_token_accuracy": 0.32019727885723115,
      "num_tokens": 3216409.0,
      "step": 3000
    },
    {
      "entropy": 5.241390740871429,
      "epoch": 1.7559009786989062,
      "grad_norm": 1.5219439268112183,
      "learning_rate": 1.3284619917501471e-06,
      "loss": 5.0645,
      "mean_token_accuracy": 0.3445430138707161,
      "num_tokens": 3266967.0,
      "step": 3050
    },
    {
      "entropy": 5.405424036979675,
      "epoch": 1.7846862406447899,
      "grad_norm": 2.1165153980255127,
      "learning_rate": 1.3166764879198585e-06,
      "loss": 5.232,
      "mean_token_accuracy": 0.32085000157356264,
      "num_tokens": 3319877.0,
      "step": 3100
    },
    {
      "entropy": 5.123006024360657,
      "epoch": 1.8134715025906736,
      "grad_norm": 1.2189785242080688,
      "learning_rate": 1.3048909840895698e-06,
      "loss": 4.9582,
      "mean_token_accuracy": 0.356108532845974,
      "num_tokens": 3368569.0,
      "step": 3150
    },
    {
      "entropy": 5.417610831260681,
      "epoch": 1.8422567645365573,
      "grad_norm": 1.5157604217529297,
      "learning_rate": 1.2931054802592812e-06,
      "loss": 5.2454,
      "mean_token_accuracy": 0.31976755023002623,
      "num_tokens": 3422449.0,
      "step": 3200
    },
    {
      "entropy": 5.409690895080566,
      "epoch": 1.871042026482441,
      "grad_norm": 1.3088161945343018,
      "learning_rate": 1.2813199764289923e-06,
      "loss": 5.2348,
      "mean_token_accuracy": 0.32325415283441544,
      "num_tokens": 3474399.0,
      "step": 3250
    },
    {
      "entropy": 5.44662567615509,
      "epoch": 1.8998272884283247,
      "grad_norm": 2.178372621536255,
      "learning_rate": 1.2695344725987036e-06,
      "loss": 5.2661,
      "mean_token_accuracy": 0.3182847076654434,
      "num_tokens": 3527726.0,
      "step": 3300
    },
    {
      "entropy": 5.512614865303039,
      "epoch": 1.9286125503742084,
      "grad_norm": 1.3050425052642822,
      "learning_rate": 1.2577489687684147e-06,
      "loss": 5.3416,
      "mean_token_accuracy": 0.3084403133392334,
      "num_tokens": 3581980.0,
      "step": 3350
    },
    {
      "entropy": 5.379772834777832,
      "epoch": 1.9573978123200921,
      "grad_norm": 1.4584404230117798,
      "learning_rate": 1.245963464938126e-06,
      "loss": 5.2087,
      "mean_token_accuracy": 0.32388432770967485,
      "num_tokens": 3635393.0,
      "step": 3400
    },
    {
      "entropy": 5.483665924072266,
      "epoch": 1.9861830742659758,
      "grad_norm": 1.2157734632492065,
      "learning_rate": 1.2341779611078374e-06,
      "loss": 5.3101,
      "mean_token_accuracy": 0.3121953472495079,
      "num_tokens": 3689894.0,
      "step": 3450
    },
    {
      "epoch": 2.0,
      "eval_entropy": 5.711394641805904,
      "eval_loss": 5.55628776550293,
      "eval_mean_token_accuracy": 0.2764948787777105,
      "eval_model_preparation_time": 0.0045,
      "eval_num_tokens": 3712724.0,
      "eval_runtime": 50.187,
      "eval_samples_per_second": 8.648,
      "eval_steps_per_second": 4.324,
      "step": 3474
    },
    {
      "entropy": 5.349283556938172,
      "epoch": 2.0149683362118593,
      "grad_norm": 1.1696771383285522,
      "learning_rate": 1.2223924572775486e-06,
      "loss": 5.1782,
      "mean_token_accuracy": 0.33028870791196824,
      "num_tokens": 3740861.0,
      "step": 3500
    },
    {
      "entropy": 5.4721107006073,
      "epoch": 2.043753598157743,
      "grad_norm": 1.8449370861053467,
      "learning_rate": 1.2106069534472599e-06,
      "loss": 5.2978,
      "mean_token_accuracy": 0.31511022299528124,
      "num_tokens": 3794869.0,
      "step": 3550
    },
    {
      "entropy": 5.404226851463318,
      "epoch": 2.0725388601036268,
      "grad_norm": 3.789496660232544,
      "learning_rate": 1.198821449616971e-06,
      "loss": 5.2371,
      "mean_token_accuracy": 0.32092176616191864,
      "num_tokens": 3848573.0,
      "step": 3600
    },
    {
      "entropy": 5.435445628166199,
      "epoch": 2.1013241220495105,
      "grad_norm": 2.2847959995269775,
      "learning_rate": 1.1870359457866824e-06,
      "loss": 5.2662,
      "mean_token_accuracy": 0.3186633634567261,
      "num_tokens": 3901204.0,
      "step": 3650
    },
    {
      "entropy": 5.4066293334960935,
      "epoch": 2.130109383995394,
      "grad_norm": 1.0950902700424194,
      "learning_rate": 1.1752504419563935e-06,
      "loss": 5.2345,
      "mean_token_accuracy": 0.32156052827835085,
      "num_tokens": 3953753.0,
      "step": 3700
    },
    {
      "entropy": 5.272332944869995,
      "epoch": 2.158894645941278,
      "grad_norm": 2.1477339267730713,
      "learning_rate": 1.1634649381261048e-06,
      "loss": 5.1091,
      "mean_token_accuracy": 0.3380983591079712,
      "num_tokens": 4005481.0,
      "step": 3750
    },
    {
      "entropy": 5.4118804311752315,
      "epoch": 2.1876799078871616,
      "grad_norm": 1.4509484767913818,
      "learning_rate": 1.151679434295816e-06,
      "loss": 5.2448,
      "mean_token_accuracy": 0.3208243528008461,
      "num_tokens": 4058829.0,
      "step": 3800
    },
    {
      "entropy": 5.4763900089263915,
      "epoch": 2.2164651698330453,
      "grad_norm": 1.0856804847717285,
      "learning_rate": 1.1398939304655273e-06,
      "loss": 5.3042,
      "mean_token_accuracy": 0.31338351368904116,
      "num_tokens": 4113326.0,
      "step": 3850
    },
    {
      "entropy": 5.328452725410461,
      "epoch": 2.245250431778929,
      "grad_norm": 3.2843880653381348,
      "learning_rate": 1.1281084266352386e-06,
      "loss": 5.1624,
      "mean_token_accuracy": 0.3305218696594238,
      "num_tokens": 4165454.0,
      "step": 3900
    },
    {
      "entropy": 5.383157343864441,
      "epoch": 2.2740356937248127,
      "grad_norm": 2.207082748413086,
      "learning_rate": 1.1163229228049497e-06,
      "loss": 5.2163,
      "mean_token_accuracy": 0.32331310987472534,
      "num_tokens": 4219250.0,
      "step": 3950
    },
    {
      "entropy": 5.585261764526368,
      "epoch": 2.3028209556706964,
      "grad_norm": 2.7102835178375244,
      "learning_rate": 1.104537418974661e-06,
      "loss": 5.4137,
      "mean_token_accuracy": 0.29959124475717547,
      "num_tokens": 4274711.0,
      "step": 4000
    },
    {
      "entropy": 5.434073266983032,
      "epoch": 2.33160621761658,
      "grad_norm": 1.3775779008865356,
      "learning_rate": 1.0927519151443724e-06,
      "loss": 5.2644,
      "mean_token_accuracy": 0.3175011593103409,
      "num_tokens": 4328616.0,
      "step": 4050
    },
    {
      "entropy": 5.462391858100891,
      "epoch": 2.360391479562464,
      "grad_norm": 1.4101024866104126,
      "learning_rate": 1.0809664113140838e-06,
      "loss": 5.2924,
      "mean_token_accuracy": 0.3137941011786461,
      "num_tokens": 4382416.0,
      "step": 4100
    },
    {
      "entropy": 5.529892563819885,
      "epoch": 2.3891767415083476,
      "grad_norm": 1.2311837673187256,
      "learning_rate": 1.0691809074837949e-06,
      "loss": 5.364,
      "mean_token_accuracy": 0.3046491605043411,
      "num_tokens": 4437848.0,
      "step": 4150
    },
    {
      "entropy": 5.4370484542846675,
      "epoch": 2.4179620034542313,
      "grad_norm": 1.0929864645004272,
      "learning_rate": 1.0573954036535062e-06,
      "loss": 5.2734,
      "mean_token_accuracy": 0.3169013774394989,
      "num_tokens": 4491185.0,
      "step": 4200
    },
    {
      "entropy": 5.395377616882325,
      "epoch": 2.446747265400115,
      "grad_norm": 1.5457273721694946,
      "learning_rate": 1.0456098998232174e-06,
      "loss": 5.2276,
      "mean_token_accuracy": 0.32221508473157884,
      "num_tokens": 4544086.0,
      "step": 4250
    },
    {
      "entropy": 5.443737335205078,
      "epoch": 2.4755325273459987,
      "grad_norm": 1.4844346046447754,
      "learning_rate": 1.0338243959929287e-06,
      "loss": 5.2786,
      "mean_token_accuracy": 0.3157751387357712,
      "num_tokens": 4597677.0,
      "step": 4300
    },
    {
      "entropy": 5.419876251220703,
      "epoch": 2.5043177892918824,
      "grad_norm": 1.2481963634490967,
      "learning_rate": 1.02203889216264e-06,
      "loss": 5.2564,
      "mean_token_accuracy": 0.31889803290367125,
      "num_tokens": 4651343.0,
      "step": 4350
    },
    {
      "entropy": 5.578677978515625,
      "epoch": 2.533103051237766,
      "grad_norm": 2.0005414485931396,
      "learning_rate": 1.0102533883323512e-06,
      "loss": 5.4145,
      "mean_token_accuracy": 0.30037090003490446,
      "num_tokens": 4705985.0,
      "step": 4400
    },
    {
      "entropy": 5.279946126937866,
      "epoch": 2.56188831318365,
      "grad_norm": 1.080521821975708,
      "learning_rate": 9.984678845020625e-07,
      "loss": 5.1226,
      "mean_token_accuracy": 0.3341303279995918,
      "num_tokens": 4757741.0,
      "step": 4450
    },
    {
      "entropy": 5.551463279724121,
      "epoch": 2.5906735751295336,
      "grad_norm": 1.28898024559021,
      "learning_rate": 9.866823806717736e-07,
      "loss": 5.3832,
      "mean_token_accuracy": 0.3028248634934425,
      "num_tokens": 4812808.0,
      "step": 4500
    },
    {
      "entropy": 5.3787487554550175,
      "epoch": 2.6194588370754173,
      "grad_norm": 1.5697983503341675,
      "learning_rate": 9.74896876841485e-07,
      "loss": 5.2141,
      "mean_token_accuracy": 0.3227942296862602,
      "num_tokens": 4866572.0,
      "step": 4550
    },
    {
      "entropy": 5.460358958244324,
      "epoch": 2.648244099021301,
      "grad_norm": 1.3180441856384277,
      "learning_rate": 9.63111373011196e-07,
      "loss": 5.2954,
      "mean_token_accuracy": 0.31269474506378175,
      "num_tokens": 4921312.0,
      "step": 4600
    },
    {
      "entropy": 5.434084935188293,
      "epoch": 2.6770293609671847,
      "grad_norm": 1.2409590482711792,
      "learning_rate": 9.513258691809074e-07,
      "loss": 5.271,
      "mean_token_accuracy": 0.3172155100107193,
      "num_tokens": 4974289.0,
      "step": 4650
    },
    {
      "entropy": 5.406955418586731,
      "epoch": 2.7058146229130684,
      "grad_norm": 1.4782609939575195,
      "learning_rate": 9.395403653506187e-07,
      "loss": 5.2473,
      "mean_token_accuracy": 0.32031788885593415,
      "num_tokens": 5028149.0,
      "step": 4700
    },
    {
      "entropy": 5.206603040695191,
      "epoch": 2.734599884858952,
      "grad_norm": 2.351633071899414,
      "learning_rate": 9.2775486152033e-07,
      "loss": 5.0478,
      "mean_token_accuracy": 0.3428420132398605,
      "num_tokens": 5079349.0,
      "step": 4750
    },
    {
      "entropy": 5.388812799453735,
      "epoch": 2.763385146804836,
      "grad_norm": 7.564618110656738,
      "learning_rate": 9.159693576900412e-07,
      "loss": 5.2281,
      "mean_token_accuracy": 0.3222071170806885,
      "num_tokens": 5132564.0,
      "step": 4800
    },
    {
      "entropy": 5.374106278419495,
      "epoch": 2.7921704087507195,
      "grad_norm": 1.4734679460525513,
      "learning_rate": 9.041838538597525e-07,
      "loss": 5.2161,
      "mean_token_accuracy": 0.3219477406144142,
      "num_tokens": 5185921.0,
      "step": 4850
    },
    {
      "entropy": 5.232998585700988,
      "epoch": 2.8209556706966032,
      "grad_norm": 1.4175471067428589,
      "learning_rate": 8.923983500294637e-07,
      "loss": 5.0769,
      "mean_token_accuracy": 0.3403926733136177,
      "num_tokens": 5237521.0,
      "step": 4900
    },
    {
      "entropy": 5.394891719818116,
      "epoch": 2.849740932642487,
      "grad_norm": 4.951873779296875,
      "learning_rate": 8.806128461991749e-07,
      "loss": 5.2344,
      "mean_token_accuracy": 0.3213117456436157,
      "num_tokens": 5291104.0,
      "step": 4950
    },
    {
      "entropy": 5.413805012702942,
      "epoch": 2.8785261945883707,
      "grad_norm": 1.679518461227417,
      "learning_rate": 8.688273423688863e-07,
      "loss": 5.2597,
      "mean_token_accuracy": 0.3165634173154831,
      "num_tokens": 5345058.0,
      "step": 5000
    },
    {
      "entropy": 5.256177935600281,
      "epoch": 2.9073114565342544,
      "grad_norm": 1.8892916440963745,
      "learning_rate": 8.570418385385975e-07,
      "loss": 5.1004,
      "mean_token_accuracy": 0.3369427987933159,
      "num_tokens": 5395918.0,
      "step": 5050
    },
    {
      "entropy": 5.259814453125,
      "epoch": 2.936096718480138,
      "grad_norm": 1.3802675008773804,
      "learning_rate": 8.452563347083087e-07,
      "loss": 5.1057,
      "mean_token_accuracy": 0.3362414276599884,
      "num_tokens": 5448086.0,
      "step": 5100
    },
    {
      "entropy": 5.416206178665161,
      "epoch": 2.964881980426022,
      "grad_norm": 1.7677236795425415,
      "learning_rate": 8.3347083087802e-07,
      "loss": 5.2562,
      "mean_token_accuracy": 0.31725785195827483,
      "num_tokens": 5501959.0,
      "step": 5150
    },
    {
      "entropy": 5.507337794303894,
      "epoch": 2.9936672423719055,
      "grad_norm": 1.021727442741394,
      "learning_rate": 8.216853270477313e-07,
      "loss": 5.344,
      "mean_token_accuracy": 0.30679062128067014,
      "num_tokens": 5557908.0,
      "step": 5200
    },
    {
      "epoch": 3.0,
      "eval_entropy": 5.682707933786278,
      "eval_loss": 5.53223991394043,
      "eval_mean_token_accuracy": 0.27747743456594404,
      "eval_model_preparation_time": 0.0045,
      "eval_num_tokens": 5569086.0,
      "eval_runtime": 49.9944,
      "eval_samples_per_second": 8.681,
      "eval_steps_per_second": 4.34,
      "step": 5211
    },
    {
      "entropy": 5.209756035804748,
      "epoch": 3.0224525043177892,
      "grad_norm": 1.725786566734314,
      "learning_rate": 8.098998232174425e-07,
      "loss": 5.0541,
      "mean_token_accuracy": 0.34166110813617706,
      "num_tokens": 5608917.0,
      "step": 5250
    },
    {
      "entropy": 5.396296281814575,
      "epoch": 3.051237766263673,
      "grad_norm": 0.7720207571983337,
      "learning_rate": 7.981143193871538e-07,
      "loss": 5.2337,
      "mean_token_accuracy": 0.32116260558366777,
      "num_tokens": 5662712.0,
      "step": 5300
    },
    {
      "entropy": 5.341518473625183,
      "epoch": 3.0800230282095566,
      "grad_norm": 2.2686808109283447,
      "learning_rate": 7.86328815556865e-07,
      "loss": 5.1824,
      "mean_token_accuracy": 0.32726580530405047,
      "num_tokens": 5715921.0,
      "step": 5350
    },
    {
      "entropy": 5.376176896095276,
      "epoch": 3.1088082901554404,
      "grad_norm": 1.2420796155929565,
      "learning_rate": 7.745433117265762e-07,
      "loss": 5.2162,
      "mean_token_accuracy": 0.32142678707838057,
      "num_tokens": 5769436.0,
      "step": 5400
    },
    {
      "entropy": 5.4553061914443965,
      "epoch": 3.137593552101324,
      "grad_norm": 1.2402859926223755,
      "learning_rate": 7.627578078962876e-07,
      "loss": 5.2971,
      "mean_token_accuracy": 0.31396267503499986,
      "num_tokens": 5823126.0,
      "step": 5450
    },
    {
      "entropy": 5.385247969627381,
      "epoch": 3.166378814047208,
      "grad_norm": 1.112062931060791,
      "learning_rate": 7.509723040659988e-07,
      "loss": 5.2324,
      "mean_token_accuracy": 0.3207343602180481,
      "num_tokens": 5875751.0,
      "step": 5500
    },
    {
      "entropy": 5.55422221660614,
      "epoch": 3.1951640759930915,
      "grad_norm": 1.5440446138381958,
      "learning_rate": 7.3918680023571e-07,
      "loss": 5.3902,
      "mean_token_accuracy": 0.3006985321640968,
      "num_tokens": 5932163.0,
      "step": 5550
    },
    {
      "entropy": 5.403217372894287,
      "epoch": 3.223949337938975,
      "grad_norm": 0.8481096625328064,
      "learning_rate": 7.274012964054213e-07,
      "loss": 5.2417,
      "mean_token_accuracy": 0.3210747820138931,
      "num_tokens": 5985889.0,
      "step": 5600
    },
    {
      "entropy": 5.388293180465698,
      "epoch": 3.252734599884859,
      "grad_norm": 0.9305989146232605,
      "learning_rate": 7.156157925751326e-07,
      "loss": 5.2319,
      "mean_token_accuracy": 0.3206030324101448,
      "num_tokens": 6040052.0,
      "step": 5650
    },
    {
      "entropy": 5.401709322929382,
      "epoch": 3.2815198618307426,
      "grad_norm": 0.8080459237098694,
      "learning_rate": 7.038302887448438e-07,
      "loss": 5.2438,
      "mean_token_accuracy": 0.3199671137332916,
      "num_tokens": 6092350.0,
      "step": 5700
    },
    {
      "entropy": 5.4320423412323,
      "epoch": 3.3103051237766263,
      "grad_norm": 1.9186089038848877,
      "learning_rate": 6.920447849145551e-07,
      "loss": 5.2696,
      "mean_token_accuracy": 0.31657984614372253,
      "num_tokens": 6146112.0,
      "step": 5750
    },
    {
      "entropy": 5.276471285820008,
      "epoch": 3.33909038572251,
      "grad_norm": 1.032879114151001,
      "learning_rate": 6.802592810842663e-07,
      "loss": 5.1224,
      "mean_token_accuracy": 0.3347566506266594,
      "num_tokens": 6197916.0,
      "step": 5800
    },
    {
      "entropy": 5.122317051887512,
      "epoch": 3.3678756476683938,
      "grad_norm": 3.156858444213867,
      "learning_rate": 6.684737772539775e-07,
      "loss": 4.9706,
      "mean_token_accuracy": 0.35455317378044127,
      "num_tokens": 6247565.0,
      "step": 5850
    },
    {
      "entropy": 5.346597375869751,
      "epoch": 3.3966609096142775,
      "grad_norm": 1.2619549036026,
      "learning_rate": 6.566882734236889e-07,
      "loss": 5.1902,
      "mean_token_accuracy": 0.3258721518516541,
      "num_tokens": 6300481.0,
      "step": 5900
    },
    {
      "entropy": 5.413151068687439,
      "epoch": 3.425446171560161,
      "grad_norm": 1.801740050315857,
      "learning_rate": 6.449027695934001e-07,
      "loss": 5.2513,
      "mean_token_accuracy": 0.3187857499718666,
      "num_tokens": 6353098.0,
      "step": 5950
    },
    {
      "entropy": 5.464186942577362,
      "epoch": 3.454231433506045,
      "grad_norm": 1.6306997537612915,
      "learning_rate": 6.331172657631113e-07,
      "loss": 5.3043,
      "mean_token_accuracy": 0.31154109388589857,
      "num_tokens": 6407984.0,
      "step": 6000
    },
    {
      "entropy": 5.401795778274536,
      "epoch": 3.4830166954519286,
      "grad_norm": 1.1694583892822266,
      "learning_rate": 6.213317619328226e-07,
      "loss": 5.2427,
      "mean_token_accuracy": 0.31954523265361784,
      "num_tokens": 6461854.0,
      "step": 6050
    },
    {
      "entropy": 5.317689285278321,
      "epoch": 3.5118019573978123,
      "grad_norm": 0.9361855387687683,
      "learning_rate": 6.095462581025339e-07,
      "loss": 5.1588,
      "mean_token_accuracy": 0.330586878657341,
      "num_tokens": 6514882.0,
      "step": 6100
    },
    {
      "entropy": 5.478708257675171,
      "epoch": 3.540587219343696,
      "grad_norm": 1.05711030960083,
      "learning_rate": 5.977607542722451e-07,
      "loss": 5.321,
      "mean_token_accuracy": 0.3104448106884956,
      "num_tokens": 6569455.0,
      "step": 6150
    },
    {
      "entropy": 5.309361801147461,
      "epoch": 3.5693724812895797,
      "grad_norm": 1.3499550819396973,
      "learning_rate": 5.859752504419564e-07,
      "loss": 5.153,
      "mean_token_accuracy": 0.331512533724308,
      "num_tokens": 6621734.0,
      "step": 6200
    },
    {
      "entropy": 5.296572666168213,
      "epoch": 3.5981577432354634,
      "grad_norm": 1.940708875656128,
      "learning_rate": 5.741897466116676e-07,
      "loss": 5.14,
      "mean_token_accuracy": 0.3299832499027252,
      "num_tokens": 6674994.0,
      "step": 6250
    },
    {
      "entropy": 5.544284400939941,
      "epoch": 3.626943005181347,
      "grad_norm": 1.8903827667236328,
      "learning_rate": 5.624042427813788e-07,
      "loss": 5.3885,
      "mean_token_accuracy": 0.3016947290301323,
      "num_tokens": 6730674.0,
      "step": 6300
    },
    {
      "entropy": 5.333053431510925,
      "epoch": 3.655728267127231,
      "grad_norm": 1.1618578433990479,
      "learning_rate": 5.506187389510902e-07,
      "loss": 5.1781,
      "mean_token_accuracy": 0.3275001719594002,
      "num_tokens": 6784235.0,
      "step": 6350
    },
    {
      "entropy": 5.4938449716568,
      "epoch": 3.6845135290731146,
      "grad_norm": 1.384329080581665,
      "learning_rate": 5.388332351208014e-07,
      "loss": 5.3399,
      "mean_token_accuracy": 0.3068840709328651,
      "num_tokens": 6839590.0,
      "step": 6400
    },
    {
      "entropy": 5.277545223236084,
      "epoch": 3.7132987910189983,
      "grad_norm": 1.8918265104293823,
      "learning_rate": 5.270477312905126e-07,
      "loss": 5.1221,
      "mean_token_accuracy": 0.33364981949329375,
      "num_tokens": 6891301.0,
      "step": 6450
    },
    {
      "entropy": 5.40100293636322,
      "epoch": 3.742084052964882,
      "grad_norm": 1.6968809366226196,
      "learning_rate": 5.152622274602239e-07,
      "loss": 5.2471,
      "mean_token_accuracy": 0.31912936180830004,
      "num_tokens": 6945510.0,
      "step": 6500
    },
    {
      "entropy": 5.561220169067383,
      "epoch": 3.7708693149107657,
      "grad_norm": 2.066960573196411,
      "learning_rate": 5.034767236299352e-07,
      "loss": 5.4026,
      "mean_token_accuracy": 0.2984810543060303,
      "num_tokens": 7001870.0,
      "step": 6550
    },
    {
      "entropy": 5.3108087682724,
      "epoch": 3.7996545768566494,
      "grad_norm": 1.6065007448196411,
      "learning_rate": 4.916912197996464e-07,
      "loss": 5.155,
      "mean_token_accuracy": 0.3304683968424797,
      "num_tokens": 7053974.0,
      "step": 6600
    },
    {
      "entropy": 5.323807754516602,
      "epoch": 3.828439838802533,
      "grad_norm": 2.6806318759918213,
      "learning_rate": 4.799057159693577e-07,
      "loss": 5.1653,
      "mean_token_accuracy": 0.3294159671664238,
      "num_tokens": 7107061.0,
      "step": 6650
    },
    {
      "entropy": 5.4716163873672485,
      "epoch": 3.857225100748417,
      "grad_norm": 1.8264856338500977,
      "learning_rate": 4.6812021213906895e-07,
      "loss": 5.3124,
      "mean_token_accuracy": 0.3109353107213974,
      "num_tokens": 7161697.0,
      "step": 6700
    },
    {
      "entropy": 5.382365622520447,
      "epoch": 3.8860103626943006,
      "grad_norm": 0.9954923987388611,
      "learning_rate": 4.563347083087802e-07,
      "loss": 5.2237,
      "mean_token_accuracy": 0.32161149621009827,
      "num_tokens": 7215524.0,
      "step": 6750
    },
    {
      "entropy": 5.277496585845947,
      "epoch": 3.9147956246401843,
      "grad_norm": 1.267786979675293,
      "learning_rate": 4.445492044784914e-07,
      "loss": 5.1265,
      "mean_token_accuracy": 0.3319795566797257,
      "num_tokens": 7267329.0,
      "step": 6800
    },
    {
      "entropy": 5.550942025184631,
      "epoch": 3.943580886586068,
      "grad_norm": 0.9425063133239746,
      "learning_rate": 4.3276370064820265e-07,
      "loss": 5.3898,
      "mean_token_accuracy": 0.30050904959440233,
      "num_tokens": 7324070.0,
      "step": 6850
    },
    {
      "entropy": 5.125799627304077,
      "epoch": 3.9723661485319517,
      "grad_norm": 5.447021007537842,
      "learning_rate": 4.20978196817914e-07,
      "loss": 4.9781,
      "mean_token_accuracy": 0.3520450854301453,
      "num_tokens": 7375083.0,
      "step": 6900
    },
    {
      "epoch": 4.0,
      "eval_entropy": 5.6681923492712905,
      "eval_loss": 5.525067329406738,
      "eval_mean_token_accuracy": 0.2779707208893816,
      "eval_model_preparation_time": 0.0045,
      "eval_num_tokens": 7425448.0,
      "eval_runtime": 49.7944,
      "eval_samples_per_second": 8.716,
      "eval_steps_per_second": 4.358,
      "step": 6948
    }
  ],
  "logging_steps": 50,
  "max_steps": 8685,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.016969752533504e+17,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}