{
  "best_global_step": 4000,
  "best_metric": 0.9059992432594299,
  "best_model_checkpoint": "/output/checkpoint-4000",
  "epoch": 1.0,
  "eval_steps": 1000,
  "global_step": 6875,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "entropy": 1.9690187186002732,
      "epoch": 0.0014545454545454545,
      "grad_norm": 3.43977427482605,
      "learning_rate": 3.6e-06,
      "loss": 3.0029,
      "mean_token_accuracy": 0.5691731169819831,
      "num_tokens": 79064.0,
      "step": 10
    },
    {
      "entropy": 1.9884992867708207,
      "epoch": 0.002909090909090909,
      "grad_norm": 2.83909010887146,
      "learning_rate": 7.6e-06,
      "loss": 2.8692,
      "mean_token_accuracy": 0.5725244455039501,
      "num_tokens": 154393.0,
      "step": 20
    },
    {
      "entropy": 1.9358498841524123,
      "epoch": 0.004363636363636364,
      "grad_norm": 1.7116823196411133,
      "learning_rate": 1.16e-05,
      "loss": 2.3806,
      "mean_token_accuracy": 0.6368451230227947,
      "num_tokens": 228943.0,
      "step": 30
    },
    {
      "entropy": 1.7093544438481332,
      "epoch": 0.005818181818181818,
      "grad_norm": 1.6132644414901733,
      "learning_rate": 1.56e-05,
      "loss": 1.8428,
      "mean_token_accuracy": 0.7234954014420509,
      "num_tokens": 306256.0,
      "step": 40
    },
    {
      "entropy": 1.2660569161176682,
      "epoch": 0.007272727272727273,
      "grad_norm": 0.6458194255828857,
      "learning_rate": 1.9600000000000002e-05,
      "loss": 1.4382,
      "mean_token_accuracy": 0.7754112549126149,
      "num_tokens": 384266.0,
      "step": 50
    },
    {
      "entropy": 1.1226252019405365,
      "epoch": 0.008727272727272728,
      "grad_norm": 0.6182602047920227,
      "learning_rate": 2.36e-05,
      "loss": 1.3458,
      "mean_token_accuracy": 0.7865310192108155,
      "num_tokens": 461739.0,
      "step": 60
    },
    {
      "entropy": 1.125070570409298,
      "epoch": 0.010181818181818183,
      "grad_norm": 0.6158664226531982,
      "learning_rate": 2.7600000000000003e-05,
      "loss": 1.2877,
      "mean_token_accuracy": 0.7905296355485916,
      "num_tokens": 537886.0,
      "step": 70
    },
    {
      "entropy": 1.1717857442796231,
      "epoch": 0.011636363636363636,
      "grad_norm": 0.7183849811553955,
      "learning_rate": 3.16e-05,
      "loss": 1.2715,
      "mean_token_accuracy": 0.7846172668039799,
      "num_tokens": 615757.0,
      "step": 80
    },
    {
      "entropy": 1.1927465781569482,
      "epoch": 0.01309090909090909,
      "grad_norm": 0.8525499105453491,
      "learning_rate": 3.56e-05,
      "loss": 1.1248,
      "mean_token_accuracy": 0.7961836472153664,
      "num_tokens": 695863.0,
      "step": 90
    },
    {
      "entropy": 1.1620682552456856,
      "epoch": 0.014545454545454545,
      "grad_norm": 0.5231317281723022,
      "learning_rate": 3.960000000000001e-05,
      "loss": 1.0643,
      "mean_token_accuracy": 0.7999849386513234,
      "num_tokens": 772443.0,
      "step": 100
    },
    {
      "entropy": 1.045405673980713,
      "epoch": 0.016,
      "grad_norm": 0.49973607063293457,
      "learning_rate": 4.36e-05,
      "loss": 0.979,
      "mean_token_accuracy": 0.8065101094543934,
      "num_tokens": 846969.0,
      "step": 110
    },
    {
      "entropy": 0.9968309663236141,
      "epoch": 0.017454545454545455,
      "grad_norm": 0.3941137492656708,
      "learning_rate": 4.76e-05,
      "loss": 0.9938,
      "mean_token_accuracy": 0.8016430333256721,
      "num_tokens": 923000.0,
      "step": 120
    },
    {
      "entropy": 0.9893214151263237,
      "epoch": 0.01890909090909091,
      "grad_norm": 0.37142255902290344,
      "learning_rate": 5.16e-05,
      "loss": 0.9833,
      "mean_token_accuracy": 0.8012244574725628,
      "num_tokens": 1000553.0,
      "step": 130
    },
    {
      "entropy": 0.9901775293052196,
      "epoch": 0.020363636363636365,
      "grad_norm": 0.4280157685279846,
      "learning_rate": 5.560000000000001e-05,
      "loss": 0.9994,
      "mean_token_accuracy": 0.7991660989820957,
      "num_tokens": 1080518.0,
      "step": 140
    },
    {
      "entropy": 1.0011372923851014,
      "epoch": 0.02181818181818182,
      "grad_norm": 0.46264714002609253,
      "learning_rate": 5.96e-05,
      "loss": 0.9677,
      "mean_token_accuracy": 0.8027120277285575,
      "num_tokens": 1157512.0,
      "step": 150
    },
    {
      "entropy": 1.0423012726008891,
      "epoch": 0.02327272727272727,
      "grad_norm": 0.5806897282600403,
      "learning_rate": 6.36e-05,
      "loss": 1.0167,
      "mean_token_accuracy": 0.7943629838526249,
      "num_tokens": 1237133.0,
      "step": 160
    },
    {
      "entropy": 0.9785076573491096,
      "epoch": 0.024727272727272726,
      "grad_norm": 0.4361409544944763,
      "learning_rate": 6.76e-05,
      "loss": 0.9548,
      "mean_token_accuracy": 0.805290812253952,
      "num_tokens": 1313695.0,
      "step": 170
    },
    {
      "entropy": 1.0021748550236225,
      "epoch": 0.02618181818181818,
      "grad_norm": 0.33419522643089294,
      "learning_rate": 7.16e-05,
      "loss": 0.9693,
      "mean_token_accuracy": 0.8024131752550602,
      "num_tokens": 1391060.0,
      "step": 180
    },
    {
      "entropy": 0.9982842341065407,
      "epoch": 0.027636363636363636,
      "grad_norm": 0.3242931663990021,
      "learning_rate": 7.560000000000001e-05,
      "loss": 0.9685,
      "mean_token_accuracy": 0.8046549871563912,
      "num_tokens": 1466252.0,
      "step": 190
    },
    {
      "entropy": 1.0053840905427933,
      "epoch": 0.02909090909090909,
      "grad_norm": 0.37451791763305664,
      "learning_rate": 7.960000000000001e-05,
      "loss": 0.9732,
      "mean_token_accuracy": 0.8007707796990872,
      "num_tokens": 1546051.0,
      "step": 200
    },
    {
      "entropy": 1.0307183906435966,
      "epoch": 0.030545454545454546,
      "grad_norm": 0.36106330156326294,
      "learning_rate": 8.36e-05,
      "loss": 0.9932,
      "mean_token_accuracy": 0.7989874318242073,
      "num_tokens": 1623829.0,
      "step": 210
    },
    {
      "entropy": 1.004240720719099,
      "epoch": 0.032,
      "grad_norm": 0.37988972663879395,
      "learning_rate": 8.76e-05,
      "loss": 0.9644,
      "mean_token_accuracy": 0.8057900808751584,
      "num_tokens": 1700491.0,
      "step": 220
    },
    {
      "entropy": 0.9733926706016064,
      "epoch": 0.03345454545454545,
      "grad_norm": 0.3222545087337494,
      "learning_rate": 9.16e-05,
      "loss": 0.9392,
      "mean_token_accuracy": 0.8078257352113724,
      "num_tokens": 1774420.0,
      "step": 230
    },
    {
      "entropy": 1.0155424050986768,
      "epoch": 0.03490909090909091,
      "grad_norm": 0.3336285352706909,
      "learning_rate": 9.56e-05,
      "loss": 0.9734,
      "mean_token_accuracy": 0.8015428952872753,
      "num_tokens": 1851932.0,
      "step": 240
    },
    {
      "entropy": 0.965352301299572,
      "epoch": 0.03636363636363636,
      "grad_norm": 0.43020081520080566,
      "learning_rate": 9.960000000000001e-05,
      "loss": 0.9282,
      "mean_token_accuracy": 0.8091536939144135,
      "num_tokens": 1927921.0,
      "step": 250
    },
    {
      "entropy": 0.9632523223757744,
      "epoch": 0.03781818181818182,
      "grad_norm": 0.32411623001098633,
      "learning_rate": 0.00010360000000000001,
      "loss": 0.9377,
      "mean_token_accuracy": 0.8093271479010582,
      "num_tokens": 2002244.0,
      "step": 260
    },
    {
      "entropy": 0.9855949737131595,
      "epoch": 0.03927272727272727,
      "grad_norm": 0.29636406898498535,
      "learning_rate": 0.00010760000000000001,
      "loss": 0.9383,
      "mean_token_accuracy": 0.8091096922755241,
      "num_tokens": 2076471.0,
      "step": 270
    },
    {
      "entropy": 0.9959837473928929,
      "epoch": 0.04072727272727273,
      "grad_norm": 0.31697383522987366,
      "learning_rate": 0.00011160000000000002,
      "loss": 0.9631,
      "mean_token_accuracy": 0.8035324491560459,
      "num_tokens": 2152670.0,
      "step": 280
    },
    {
      "entropy": 0.9984123289585114,
      "epoch": 0.04218181818181818,
      "grad_norm": 0.30740225315093994,
      "learning_rate": 0.00011559999999999999,
      "loss": 0.9652,
      "mean_token_accuracy": 0.8041438944637775,
      "num_tokens": 2227827.0,
      "step": 290
    },
    {
      "entropy": 0.9823802955448627,
      "epoch": 0.04363636363636364,
      "grad_norm": 0.5217260122299194,
      "learning_rate": 0.00011960000000000001,
      "loss": 0.9501,
      "mean_token_accuracy": 0.8054948531091213,
      "num_tokens": 2304457.0,
      "step": 300
    },
    {
      "entropy": 0.986659524589777,
      "epoch": 0.04509090909090909,
      "grad_norm": 0.37950316071510315,
      "learning_rate": 0.0001236,
      "loss": 0.9508,
      "mean_token_accuracy": 0.8070364102721215,
      "num_tokens": 2381034.0,
      "step": 310
    },
    {
      "entropy": 0.9770919024944306,
      "epoch": 0.04654545454545454,
      "grad_norm": 0.2790478467941284,
      "learning_rate": 0.0001276,
      "loss": 0.9416,
      "mean_token_accuracy": 0.8086497105658055,
      "num_tokens": 2457635.0,
      "step": 320
    },
    {
      "entropy": 0.9718254946172238,
      "epoch": 0.048,
      "grad_norm": 0.29544803500175476,
      "learning_rate": 0.0001316,
      "loss": 0.9455,
      "mean_token_accuracy": 0.8079691752791405,
      "num_tokens": 2534043.0,
      "step": 330
    },
    {
      "entropy": 0.9868722438812256,
      "epoch": 0.04945454545454545,
      "grad_norm": 0.3150554895401001,
      "learning_rate": 0.00013560000000000002,
      "loss": 0.9516,
      "mean_token_accuracy": 0.8049276761710644,
      "num_tokens": 2610676.0,
      "step": 340
    },
    {
      "entropy": 0.9821666151285171,
      "epoch": 0.05090909090909091,
      "grad_norm": 0.30912908911705017,
      "learning_rate": 0.0001396,
      "loss": 0.9413,
      "mean_token_accuracy": 0.80631243288517,
      "num_tokens": 2687398.0,
      "step": 350
    },
    {
      "entropy": 0.9813764326274395,
      "epoch": 0.05236363636363636,
      "grad_norm": 0.3392393887042999,
      "learning_rate": 0.0001436,
      "loss": 0.9551,
      "mean_token_accuracy": 0.8057433366775513,
      "num_tokens": 2763780.0,
      "step": 360
    },
    {
      "entropy": 0.9714535295963287,
      "epoch": 0.05381818181818182,
      "grad_norm": 0.593127429485321,
      "learning_rate": 0.0001476,
      "loss": 0.9446,
      "mean_token_accuracy": 0.8071163192391395,
      "num_tokens": 2841687.0,
      "step": 370
    },
    {
      "entropy": 0.9950341284275055,
      "epoch": 0.05527272727272727,
      "grad_norm": 0.5920702219009399,
      "learning_rate": 0.0001516,
      "loss": 0.9567,
      "mean_token_accuracy": 0.8022288359701634,
      "num_tokens": 2921696.0,
      "step": 380
    },
    {
      "entropy": 0.9623408317565918,
      "epoch": 0.05672727272727273,
      "grad_norm": 0.26438552141189575,
      "learning_rate": 0.00015560000000000001,
      "loss": 0.9451,
      "mean_token_accuracy": 0.8056230574846268,
      "num_tokens": 2999827.0,
      "step": 390
    },
    {
      "entropy": 0.969329858571291,
      "epoch": 0.05818181818181818,
      "grad_norm": 0.28617411851882935,
      "learning_rate": 0.0001596,
      "loss": 0.9311,
      "mean_token_accuracy": 0.8082362949848175,
      "num_tokens": 3075878.0,
      "step": 400
    },
    {
      "entropy": 0.9626947619020939,
      "epoch": 0.05963636363636363,
      "grad_norm": 0.29560142755508423,
      "learning_rate": 0.0001636,
      "loss": 0.9325,
      "mean_token_accuracy": 0.8075426079332828,
      "num_tokens": 3149659.0,
      "step": 410
    },
    {
      "entropy": 0.9842231079936028,
      "epoch": 0.06109090909090909,
      "grad_norm": 0.36081886291503906,
      "learning_rate": 0.0001676,
      "loss": 0.9597,
      "mean_token_accuracy": 0.8048615127801895,
      "num_tokens": 3224908.0,
      "step": 420
    },
    {
      "entropy": 0.9678519986569881,
      "epoch": 0.06254545454545454,
      "grad_norm": 0.32404276728630066,
      "learning_rate": 0.0001716,
      "loss": 0.9501,
      "mean_token_accuracy": 0.8070563480257988,
      "num_tokens": 3300465.0,
      "step": 430
    },
    {
      "entropy": 0.9691661924123764,
      "epoch": 0.064,
      "grad_norm": 0.3594436049461365,
      "learning_rate": 0.0001756,
      "loss": 0.9418,
      "mean_token_accuracy": 0.80767857208848,
      "num_tokens": 3378460.0,
      "step": 440
    },
    {
      "entropy": 1.000221849232912,
      "epoch": 0.06545454545454546,
      "grad_norm": 0.27946728467941284,
      "learning_rate": 0.0001796,
      "loss": 0.9784,
      "mean_token_accuracy": 0.8053525581955909,
      "num_tokens": 3457035.0,
      "step": 450
    },
    {
      "entropy": 0.9556974627077579,
      "epoch": 0.0669090909090909,
      "grad_norm": 0.3116967976093292,
      "learning_rate": 0.00018360000000000002,
      "loss": 0.9295,
      "mean_token_accuracy": 0.8102393008768558,
      "num_tokens": 3531556.0,
      "step": 460
    },
    {
      "entropy": 1.0009116798639297,
      "epoch": 0.06836363636363636,
      "grad_norm": 0.24993526935577393,
      "learning_rate": 0.0001876,
      "loss": 0.9652,
      "mean_token_accuracy": 0.8023554868996143,
      "num_tokens": 3608896.0,
      "step": 470
    },
    {
      "entropy": 0.9631584003567696,
      "epoch": 0.06981818181818182,
      "grad_norm": 0.39844977855682373,
      "learning_rate": 0.0001916,
      "loss": 0.9402,
      "mean_token_accuracy": 0.8080028325319291,
      "num_tokens": 3685526.0,
      "step": 480
    },
    {
      "entropy": 0.9812145739793777,
      "epoch": 0.07127272727272728,
      "grad_norm": 0.25667306780815125,
      "learning_rate": 0.0001956,
      "loss": 0.9498,
      "mean_token_accuracy": 0.8053297907114029,
      "num_tokens": 3762930.0,
      "step": 490
    },
    {
      "entropy": 0.9755678996443748,
      "epoch": 0.07272727272727272,
      "grad_norm": 0.25863754749298096,
      "learning_rate": 0.0001996,
      "loss": 0.9655,
      "mean_token_accuracy": 0.8062391959130764,
      "num_tokens": 3841245.0,
      "step": 500
    },
    {
      "entropy": 0.9802602089941501,
      "epoch": 0.07418181818181818,
      "grad_norm": 0.36054691672325134,
      "learning_rate": 0.00019999901694803208,
      "loss": 0.9497,
      "mean_token_accuracy": 0.8051057532429695,
      "num_tokens": 3918834.0,
      "step": 510
    },
    {
      "entropy": 0.9571634404361248,
      "epoch": 0.07563636363636364,
      "grad_norm": 0.4442167580127716,
      "learning_rate": 0.00019999561876852487,
      "loss": 0.9338,
      "mean_token_accuracy": 0.8084112584590912,
      "num_tokens": 3995115.0,
      "step": 520
    },
    {
      "entropy": 0.963586512953043,
      "epoch": 0.07709090909090908,
      "grad_norm": 0.31452101469039917,
      "learning_rate": 0.0001999897934075395,
      "loss": 0.9485,
      "mean_token_accuracy": 0.8092979274690151,
      "num_tokens": 4072831.0,
      "step": 530
    },
    {
      "entropy": 0.9710368476808071,
      "epoch": 0.07854545454545454,
      "grad_norm": 0.271147757768631,
      "learning_rate": 0.0001999815410065451,
      "loss": 0.9452,
      "mean_token_accuracy": 0.8060981146991253,
      "num_tokens": 4149986.0,
      "step": 540
    },
    {
      "entropy": 0.9665751494467258,
      "epoch": 0.08,
      "grad_norm": 0.34051278233528137,
      "learning_rate": 0.0001999708617659513,
      "loss": 0.9449,
      "mean_token_accuracy": 0.8084499321877956,
      "num_tokens": 4225284.0,
      "step": 550
    },
    {
      "entropy": 0.9917739503085613,
      "epoch": 0.08145454545454546,
      "grad_norm": 0.28123149275779724,
      "learning_rate": 0.00019995775594510375,
      "loss": 0.965,
      "mean_token_accuracy": 0.8031023532152176,
      "num_tokens": 4304103.0,
      "step": 560
    },
    {
      "entropy": 0.9659538678824902,
      "epoch": 0.0829090909090909,
      "grad_norm": 0.3097119927406311,
      "learning_rate": 0.00019994222386227756,
      "loss": 0.9469,
      "mean_token_accuracy": 0.8073901452124119,
      "num_tokens": 4378474.0,
      "step": 570
    },
    {
      "entropy": 0.9613742314279079,
      "epoch": 0.08436363636363636,
      "grad_norm": 0.36504626274108887,
      "learning_rate": 0.0001999242658946697,
      "loss": 0.9362,
      "mean_token_accuracy": 0.8093178525567055,
      "num_tokens": 4455068.0,
      "step": 580
    },
    {
      "entropy": 0.9502459809184074,
      "epoch": 0.08581818181818182,
      "grad_norm": 0.30682137608528137,
      "learning_rate": 0.00019990388247838976,
      "loss": 0.9296,
      "mean_token_accuracy": 0.8076415918767452,
      "num_tokens": 4532441.0,
      "step": 590
    },
    {
      "entropy": 0.9592754155397415,
      "epoch": 0.08727272727272728,
      "grad_norm": 0.2672958970069885,
      "learning_rate": 0.0001998810741084495,
      "loss": 0.9517,
      "mean_token_accuracy": 0.8078010775148868,
      "num_tokens": 4608937.0,
      "step": 600
    },
    {
      "entropy": 0.9618418499827385,
      "epoch": 0.08872727272727272,
      "grad_norm": 0.3508414328098297,
      "learning_rate": 0.00019985584133875062,
      "loss": 0.929,
      "mean_token_accuracy": 0.8108504980802536,
      "num_tokens": 4683807.0,
      "step": 610
    },
    {
      "entropy": 0.9733804412186146,
      "epoch": 0.09018181818181818,
      "grad_norm": 0.3088361620903015,
      "learning_rate": 0.00019982818478207146,
      "loss": 0.9496,
      "mean_token_accuracy": 0.8067471913993358,
      "num_tokens": 4761822.0,
      "step": 620
    },
    {
      "entropy": 0.948646043241024,
      "epoch": 0.09163636363636364,
      "grad_norm": 0.29966917634010315,
      "learning_rate": 0.00019979810511005215,
      "loss": 0.9269,
      "mean_token_accuracy": 0.8102207146584988,
      "num_tokens": 4836571.0,
      "step": 630
    },
    {
      "entropy": 0.9861510775983333,
      "epoch": 0.09309090909090909,
      "grad_norm": 0.2919958233833313,
      "learning_rate": 0.0001997656030531781,
      "loss": 0.97,
      "mean_token_accuracy": 0.8068868108093739,
      "num_tokens": 4912711.0,
      "step": 640
    },
    {
      "entropy": 0.9733738884329796,
      "epoch": 0.09454545454545454,
      "grad_norm": 0.2824857234954834,
      "learning_rate": 0.00019973067940076252,
      "loss": 0.9599,
      "mean_token_accuracy": 0.8068003825843334,
      "num_tokens": 4990087.0,
      "step": 650
    },
    {
      "entropy": 0.9654402084648609,
      "epoch": 0.096,
      "grad_norm": 0.5585899353027344,
      "learning_rate": 0.00019969333500092698,
      "loss": 0.9415,
      "mean_token_accuracy": 0.8078834369778634,
      "num_tokens": 5068076.0,
      "step": 660
    },
    {
      "entropy": 0.9511952914297581,
      "epoch": 0.09745454545454546,
      "grad_norm": 0.27784818410873413,
      "learning_rate": 0.00019965357076058104,
      "loss": 0.9276,
      "mean_token_accuracy": 0.810407143086195,
      "num_tokens": 5144288.0,
      "step": 670
    },
    {
      "entropy": 0.9477476179599762,
      "epoch": 0.0989090909090909,
      "grad_norm": 0.39101967215538025,
      "learning_rate": 0.0001996113876454002,
      "loss": 0.9259,
      "mean_token_accuracy": 0.811048673093319,
      "num_tokens": 5218249.0,
      "step": 680
    },
    {
      "entropy": 0.9532744571566582,
      "epoch": 0.10036363636363636,
      "grad_norm": 0.32930827140808105,
      "learning_rate": 0.00019956678667980217,
      "loss": 0.9209,
      "mean_token_accuracy": 0.8119081415235996,
      "num_tokens": 5294409.0,
      "step": 690
    },
    {
      "entropy": 0.9863594181835651,
      "epoch": 0.10181818181818182,
      "grad_norm": 0.31709808111190796,
      "learning_rate": 0.0001995197689469225,
      "loss": 0.9568,
      "mean_token_accuracy": 0.8043609701097012,
      "num_tokens": 5373459.0,
      "step": 700
    },
    {
      "entropy": 0.9500035136938095,
      "epoch": 0.10327272727272727,
      "grad_norm": 0.35338568687438965,
      "learning_rate": 0.00019947033558858775,
      "loss": 0.9333,
      "mean_token_accuracy": 0.8085519306361675,
      "num_tokens": 5450783.0,
      "step": 710
    },
    {
      "entropy": 0.9919221103191376,
      "epoch": 0.10472727272727272,
      "grad_norm": 0.3136957287788391,
      "learning_rate": 0.00019941848780528806,
      "loss": 0.9667,
      "mean_token_accuracy": 0.8043533161282539,
      "num_tokens": 5527972.0,
      "step": 720
    },
    {
      "entropy": 0.9358678698539734,
      "epoch": 0.10618181818181818,
      "grad_norm": 0.30247825384140015,
      "learning_rate": 0.00019936422685614799,
      "loss": 0.9128,
      "mean_token_accuracy": 0.812498589605093,
      "num_tokens": 5601051.0,
      "step": 730
    },
    {
      "entropy": 0.9480448178946972,
      "epoch": 0.10763636363636364,
      "grad_norm": 0.3738067150115967,
      "learning_rate": 0.00019930755405889582,
      "loss": 0.927,
      "mean_token_accuracy": 0.8110132366418839,
      "num_tokens": 5674547.0,
      "step": 740
    },
    {
      "entropy": 0.9623192340135575,
      "epoch": 0.10909090909090909,
      "grad_norm": 0.341701865196228,
      "learning_rate": 0.0001992484707898317,
      "loss": 0.9471,
      "mean_token_accuracy": 0.808195649087429,
      "num_tokens": 5753663.0,
      "step": 750
    },
    {
      "entropy": 0.9636863119900226,
      "epoch": 0.11054545454545454,
      "grad_norm": 0.3272800147533417,
      "learning_rate": 0.00019918697848379398,
      "loss": 0.9372,
      "mean_token_accuracy": 0.8089156173169613,
      "num_tokens": 5829244.0,
      "step": 760
    },
    {
      "entropy": 0.952582273632288,
      "epoch": 0.112,
      "grad_norm": 0.2747955024242401,
      "learning_rate": 0.00019912307863412468,
      "loss": 0.9246,
      "mean_token_accuracy": 0.8115990363061428,
      "num_tokens": 5905553.0,
      "step": 770
    },
    {
      "entropy": 0.956246180832386,
      "epoch": 0.11345454545454546,
      "grad_norm": 0.28963592648506165,
      "learning_rate": 0.00019905677279263303,
      "loss": 0.9513,
      "mean_token_accuracy": 0.8093406230211257,
      "num_tokens": 5982595.0,
      "step": 780
    },
    {
      "entropy": 0.9501231409609318,
      "epoch": 0.1149090909090909,
      "grad_norm": 0.2975050210952759,
      "learning_rate": 0.0001989880625695577,
      "loss": 0.9227,
      "mean_token_accuracy": 0.8131072640419006,
      "num_tokens": 6058325.0,
      "step": 790
    },
    {
      "entropy": 0.9440824434161186,
      "epoch": 0.11636363636363636,
      "grad_norm": 0.34666338562965393,
      "learning_rate": 0.00019891694963352803,
      "loss": 0.9146,
      "mean_token_accuracy": 0.8128655597567558,
      "num_tokens": 6133747.0,
      "step": 800
    },
    {
      "entropy": 0.9762783326208592,
      "epoch": 0.11781818181818182,
      "grad_norm": 0.4371774196624756,
      "learning_rate": 0.00019884343571152313,
      "loss": 0.9608,
      "mean_token_accuracy": 0.8071666911244393,
      "num_tokens": 6212280.0,
      "step": 810
    },
    {
      "entropy": 0.9690908081829548,
      "epoch": 0.11927272727272727,
      "grad_norm": 0.3035489618778229,
      "learning_rate": 0.0001987675225888302,
      "loss": 0.9328,
      "mean_token_accuracy": 0.8099577404558659,
      "num_tokens": 6288436.0,
      "step": 820
    },
    {
      "entropy": 0.9473535366356373,
      "epoch": 0.12072727272727272,
      "grad_norm": 0.2877722680568695,
      "learning_rate": 0.000198689212109001,
      "loss": 0.9367,
      "mean_token_accuracy": 0.8102696388959885,
      "num_tokens": 6365114.0,
      "step": 830
    },
    {
      "entropy": 0.9800632715225219,
      "epoch": 0.12218181818181818,
      "grad_norm": 0.3979131877422333,
      "learning_rate": 0.0001986085061738072,
      "loss": 0.9526,
      "mean_token_accuracy": 0.807623814046383,
      "num_tokens": 6442583.0,
      "step": 840
    },
    {
      "entropy": 0.9318563066422939,
      "epoch": 0.12363636363636364,
      "grad_norm": 0.36992502212524414,
      "learning_rate": 0.00019852540674319415,
      "loss": 0.9192,
      "mean_token_accuracy": 0.8144332192838192,
      "num_tokens": 6518109.0,
      "step": 850
    },
    {
      "entropy": 0.9581544451415539,
      "epoch": 0.12509090909090909,
      "grad_norm": 0.2897164225578308,
      "learning_rate": 0.00019843991583523332,
      "loss": 0.925,
      "mean_token_accuracy": 0.8085106551647187,
      "num_tokens": 6598310.0,
      "step": 860
    },
    {
      "entropy": 0.9448593385517597,
      "epoch": 0.12654545454545454,
      "grad_norm": 0.2775498926639557,
      "learning_rate": 0.0001983520355260732,
      "loss": 0.9303,
      "mean_token_accuracy": 0.8096731394529343,
      "num_tokens": 6676666.0,
      "step": 870
    },
    {
      "entropy": 0.9634767659008503,
      "epoch": 0.128,
      "grad_norm": 0.3571963906288147,
      "learning_rate": 0.000198261767949889,
      "loss": 0.9336,
      "mean_token_accuracy": 0.809954322874546,
      "num_tokens": 6753460.0,
      "step": 880
    },
    {
      "entropy": 0.9582797236740589,
      "epoch": 0.12945454545454546,
      "grad_norm": 0.3039088845252991,
      "learning_rate": 0.0001981691152988307,
      "loss": 0.9457,
      "mean_token_accuracy": 0.8096859961748123,
      "num_tokens": 6830638.0,
      "step": 890
    },
    {
      "entropy": 0.941552960127592,
      "epoch": 0.13090909090909092,
      "grad_norm": 0.32802486419677734,
      "learning_rate": 0.00019807407982296992,
      "loss": 0.9149,
      "mean_token_accuracy": 0.8126296833157539,
      "num_tokens": 6906602.0,
      "step": 900
    },
    {
      "entropy": 0.9520909287035465,
      "epoch": 0.13236363636363635,
      "grad_norm": 0.3812902569770813,
      "learning_rate": 0.00019797666383024516,
      "loss": 0.9332,
      "mean_token_accuracy": 0.8106161102652549,
      "num_tokens": 6981252.0,
      "step": 910
    },
    {
      "entropy": 0.9571222126483917,
      "epoch": 0.1338181818181818,
      "grad_norm": 0.33096054196357727,
      "learning_rate": 0.00019787686968640594,
      "loss": 0.933,
      "mean_token_accuracy": 0.8085237428545952,
      "num_tokens": 7059497.0,
      "step": 920
    },
    {
      "entropy": 0.9404919967055321,
      "epoch": 0.13527272727272727,
      "grad_norm": 0.3364405333995819,
      "learning_rate": 0.00019777469981495515,
      "loss": 0.9238,
      "mean_token_accuracy": 0.8110041253268718,
      "num_tokens": 7135602.0,
      "step": 930
    },
    {
      "entropy": 0.9433817587792873,
      "epoch": 0.13672727272727273,
      "grad_norm": 0.28743976354599,
      "learning_rate": 0.0001976701566970903,
      "loss": 0.9178,
      "mean_token_accuracy": 0.8115247070789338,
      "num_tokens": 7211560.0,
      "step": 940
    },
    {
      "entropy": 0.9528502456843853,
      "epoch": 0.13818181818181818,
      "grad_norm": 0.28429099917411804,
      "learning_rate": 0.0001975632428716432,
      "loss": 0.9356,
      "mean_token_accuracy": 0.81063963919878,
      "num_tokens": 7289139.0,
      "step": 950
    },
    {
      "entropy": 0.9531728699803352,
      "epoch": 0.13963636363636364,
      "grad_norm": 0.30310696363449097,
      "learning_rate": 0.0001974539609350184,
      "loss": 0.9214,
      "mean_token_accuracy": 0.8126495592296124,
      "num_tokens": 7365669.0,
      "step": 960
    },
    {
      "entropy": 0.9738307923078537,
      "epoch": 0.1410909090909091,
      "grad_norm": 0.3467698395252228,
      "learning_rate": 0.00019734231354113003,
      "loss": 0.9643,
      "mean_token_accuracy": 0.8050346717238426,
      "num_tokens": 7443919.0,
      "step": 970
    },
    {
      "entropy": 0.9685733772814273,
      "epoch": 0.14254545454545456,
      "grad_norm": 0.29552823305130005,
      "learning_rate": 0.00019722830340133748,
      "loss": 0.942,
      "mean_token_accuracy": 0.8079716853797436,
      "num_tokens": 7520114.0,
      "step": 980
    },
    {
      "entropy": 0.9668281979858875,
      "epoch": 0.144,
      "grad_norm": 0.2899906039237976,
      "learning_rate": 0.00019711193328437936,
      "loss": 0.9367,
      "mean_token_accuracy": 0.8102179229259491,
      "num_tokens": 7597364.0,
      "step": 990
    },
    {
      "entropy": 0.9519609495997429,
      "epoch": 0.14545454545454545,
      "grad_norm": 0.2897755205631256,
      "learning_rate": 0.0001969932060163065,
      "loss": 0.9446,
      "mean_token_accuracy": 0.8093790285289287,
      "num_tokens": 7674378.0,
      "step": 1000
    },
    {
      "epoch": 0.14545454545454545,
      "eval_entropy": 0.9725983184814453,
      "eval_loss": 0.9227215051651001,
      "eval_mean_token_accuracy": 0.8082753699302674,
      "eval_num_tokens": 7674378.0,
      "eval_runtime": 78.3318,
      "eval_samples_per_second": 63.831,
      "eval_steps_per_second": 7.979,
      "step": 1000
    },
    {
      "entropy": 0.9554769098758698,
      "epoch": 0.1469090909090909,
      "grad_norm": 0.40527433156967163,
      "learning_rate": 0.00019687212448041305,
      "loss": 0.9341,
      "mean_token_accuracy": 0.8105138674378395,
      "num_tokens": 7749244.0,
      "step": 1010
    },
    {
      "entropy": 0.9631451539695263,
      "epoch": 0.14836363636363636,
      "grad_norm": 0.2934854030609131,
      "learning_rate": 0.00019674869161716673,
      "loss": 0.9429,
      "mean_token_accuracy": 0.8103604264557361,
      "num_tokens": 7827130.0,
      "step": 1020
    },
    {
      "entropy": 0.9962614960968494,
      "epoch": 0.14981818181818182,
      "grad_norm": 0.29253679513931274,
      "learning_rate": 0.00019662291042413726,
      "loss": 0.9789,
      "mean_token_accuracy": 0.8047004818916321,
      "num_tokens": 7905729.0,
      "step": 1030
    },
    {
      "entropy": 0.9482852481305599,
      "epoch": 0.15127272727272728,
      "grad_norm": 0.29692456126213074,
      "learning_rate": 0.00019649478395592355,
      "loss": 0.9362,
      "mean_token_accuracy": 0.8123825170099735,
      "num_tokens": 7980537.0,
      "step": 1040
    },
    {
      "entropy": 0.955985976755619,
      "epoch": 0.15272727272727274,
      "grad_norm": 0.4257398843765259,
      "learning_rate": 0.00019636431532407968,
      "loss": 0.9342,
      "mean_token_accuracy": 0.8100227758288383,
      "num_tokens": 8056340.0,
      "step": 1050
    },
    {
      "entropy": 0.9584499664604664,
      "epoch": 0.15418181818181817,
      "grad_norm": 0.29070353507995605,
      "learning_rate": 0.00019623150769703908,
      "loss": 0.9383,
      "mean_token_accuracy": 0.810140723735094,
      "num_tokens": 8132679.0,
      "step": 1060
    },
    {
      "entropy": 0.9684336230158805,
      "epoch": 0.15563636363636363,
      "grad_norm": 0.3143501579761505,
      "learning_rate": 0.00019609636430003785,
      "loss": 0.9499,
      "mean_token_accuracy": 0.8064659915864467,
      "num_tokens": 8208680.0,
      "step": 1070
    },
    {
      "entropy": 0.9466678649187088,
      "epoch": 0.1570909090909091,
      "grad_norm": 0.29853272438049316,
      "learning_rate": 0.00019595888841503618,
      "loss": 0.925,
      "mean_token_accuracy": 0.8112887017428875,
      "num_tokens": 8287768.0,
      "step": 1080
    },
    {
      "entropy": 0.950913506001234,
      "epoch": 0.15854545454545454,
      "grad_norm": 0.3101540505886078,
      "learning_rate": 0.00019581908338063897,
      "loss": 0.9298,
      "mean_token_accuracy": 0.8090488716959954,
      "num_tokens": 8366655.0,
      "step": 1090
    },
    {
      "entropy": 0.9407593891024589,
      "epoch": 0.16,
      "grad_norm": 0.29976364970207214,
      "learning_rate": 0.00019567695259201434,
      "loss": 0.9345,
      "mean_token_accuracy": 0.8117971464991569,
      "num_tokens": 8442053.0,
      "step": 1100
    },
    {
      "entropy": 0.955774050205946,
      "epoch": 0.16145454545454546,
      "grad_norm": 0.34253445267677307,
      "learning_rate": 0.00019553249950081164,
      "loss": 0.9431,
      "mean_token_accuracy": 0.8093701273202896,
      "num_tokens": 8518218.0,
      "step": 1110
    },
    {
      "entropy": 0.98501892760396,
      "epoch": 0.16290909090909092,
      "grad_norm": 0.33748891949653625,
      "learning_rate": 0.00019538572761507714,
      "loss": 0.9831,
      "mean_token_accuracy": 0.8018230192363263,
      "num_tokens": 8597394.0,
      "step": 1120
    },
    {
      "entropy": 0.9631325028836727,
      "epoch": 0.16436363636363635,
      "grad_norm": 0.38818150758743286,
      "learning_rate": 0.0001952366404991693,
      "loss": 0.9509,
      "mean_token_accuracy": 0.8064041070640087,
      "num_tokens": 8675541.0,
      "step": 1130
    },
    {
      "entropy": 0.9727548152208328,
      "epoch": 0.1658181818181818,
      "grad_norm": 0.3559781312942505,
      "learning_rate": 0.00019508524177367184,
      "loss": 0.949,
      "mean_token_accuracy": 0.8060977406799793,
      "num_tokens": 8755075.0,
      "step": 1140
    },
    {
      "entropy": 0.9649280205368995,
      "epoch": 0.16727272727272727,
      "grad_norm": 0.30704405903816223,
      "learning_rate": 0.0001949315351153061,
      "loss": 0.9547,
      "mean_token_accuracy": 0.8092526689171791,
      "num_tokens": 8831736.0,
      "step": 1150
    },
    {
      "entropy": 0.9345536336302758,
      "epoch": 0.16872727272727273,
      "grad_norm": 0.30081990361213684,
      "learning_rate": 0.00019477552425684148,
      "loss": 0.9178,
      "mean_token_accuracy": 0.8131528303027153,
      "num_tokens": 8911228.0,
      "step": 1160
    },
    {
      "entropy": 0.9596106916666031,
      "epoch": 0.17018181818181818,
      "grad_norm": 0.3070826232433319,
      "learning_rate": 0.000194617212987005,
      "loss": 0.9362,
      "mean_token_accuracy": 0.8075219623744487,
      "num_tokens": 8988216.0,
      "step": 1170
    },
    {
      "entropy": 0.9631124071776866,
      "epoch": 0.17163636363636364,
      "grad_norm": 0.3408432900905609,
      "learning_rate": 0.00019445660515038927,
      "loss": 0.9423,
      "mean_token_accuracy": 0.8091696575284004,
      "num_tokens": 9065626.0,
      "step": 1180
    },
    {
      "entropy": 0.9516327641904354,
      "epoch": 0.1730909090909091,
      "grad_norm": 0.31564784049987793,
      "learning_rate": 0.00019429370464735897,
      "loss": 0.9521,
      "mean_token_accuracy": 0.8093578301370143,
      "num_tokens": 9142973.0,
      "step": 1190
    },
    {
      "entropy": 0.9371618889272213,
      "epoch": 0.17454545454545456,
      "grad_norm": 0.3847862184047699,
      "learning_rate": 0.0001941285154339563,
      "loss": 0.9091,
      "mean_token_accuracy": 0.812215743213892,
      "num_tokens": 9218296.0,
      "step": 1200
    },
    {
      "entropy": 0.9448715083301067,
      "epoch": 0.176,
      "grad_norm": 0.3291180431842804,
      "learning_rate": 0.0001939610415218048,
      "loss": 0.9186,
      "mean_token_accuracy": 0.8123110935091973,
      "num_tokens": 9293398.0,
      "step": 1210
    },
    {
      "entropy": 0.966688671708107,
      "epoch": 0.17745454545454545,
      "grad_norm": 0.30451062321662903,
      "learning_rate": 0.00019379128697801206,
      "loss": 0.951,
      "mean_token_accuracy": 0.8065337009727955,
      "num_tokens": 9373633.0,
      "step": 1220
    },
    {
      "entropy": 0.9484643906354904,
      "epoch": 0.1789090909090909,
      "grad_norm": 0.2932063341140747,
      "learning_rate": 0.0001936192559250707,
      "loss": 0.9325,
      "mean_token_accuracy": 0.8090389378368854,
      "num_tokens": 9451080.0,
      "step": 1230
    },
    {
      "entropy": 0.9615976348519325,
      "epoch": 0.18036363636363636,
      "grad_norm": 0.33180785179138184,
      "learning_rate": 0.00019344495254075854,
      "loss": 0.9344,
      "mean_token_accuracy": 0.8113547652959824,
      "num_tokens": 9526463.0,
      "step": 1240
    },
    {
      "entropy": 0.9671885691583156,
      "epoch": 0.18181818181818182,
      "grad_norm": 0.30535823106765747,
      "learning_rate": 0.000193268381058037,
      "loss": 0.9717,
      "mean_token_accuracy": 0.8057887375354766,
      "num_tokens": 9605802.0,
      "step": 1250
    },
    {
      "entropy": 0.9746878013014794,
      "epoch": 0.18327272727272728,
      "grad_norm": 0.3194335699081421,
      "learning_rate": 0.00019308954576494826,
      "loss": 0.948,
      "mean_token_accuracy": 0.8088590256869793,
      "num_tokens": 9683267.0,
      "step": 1260
    },
    {
      "entropy": 0.925223445147276,
      "epoch": 0.18472727272727274,
      "grad_norm": 0.3094640076160431,
      "learning_rate": 0.00019290845100451123,
      "loss": 0.9105,
      "mean_token_accuracy": 0.8131582289934158,
      "num_tokens": 9760203.0,
      "step": 1270
    },
    {
      "entropy": 0.9676970973610878,
      "epoch": 0.18618181818181817,
      "grad_norm": 0.3541739284992218,
      "learning_rate": 0.000192725101174616,
      "loss": 0.9339,
      "mean_token_accuracy": 0.8105718605220318,
      "num_tokens": 9837897.0,
      "step": 1280
    },
    {
      "entropy": 0.9446438558399677,
      "epoch": 0.18763636363636363,
      "grad_norm": 0.31916889548301697,
      "learning_rate": 0.00019253950072791712,
      "loss": 0.9307,
      "mean_token_accuracy": 0.8087717853486538,
      "num_tokens": 9914578.0,
      "step": 1290
    },
    {
      "entropy": 0.9461817003786563,
      "epoch": 0.1890909090909091,
      "grad_norm": 0.30146437883377075,
      "learning_rate": 0.00019235165417172536,
      "loss": 0.9266,
      "mean_token_accuracy": 0.8111361682415008,
      "num_tokens": 9991551.0,
      "step": 1300
    },
    {
      "entropy": 0.9666557475924492,
      "epoch": 0.19054545454545455,
      "grad_norm": 0.32777443528175354,
      "learning_rate": 0.00019216156606789836,
      "loss": 0.9374,
      "mean_token_accuracy": 0.8101299934089183,
      "num_tokens": 10067407.0,
      "step": 1310
    },
    {
      "entropy": 0.9579305075109005,
      "epoch": 0.192,
      "grad_norm": 0.3164553642272949,
      "learning_rate": 0.00019196924103272972,
      "loss": 0.9383,
      "mean_token_accuracy": 0.8088756151497364,
      "num_tokens": 10143656.0,
      "step": 1320
    },
    {
      "entropy": 0.9566913366317749,
      "epoch": 0.19345454545454546,
      "grad_norm": 0.3002376854419708,
      "learning_rate": 0.00019177468373683708,
      "loss": 0.9359,
      "mean_token_accuracy": 0.8092375807464123,
      "num_tokens": 10221715.0,
      "step": 1330
    },
    {
      "entropy": 0.9349559977650642,
      "epoch": 0.19490909090909092,
      "grad_norm": 0.3056219518184662,
      "learning_rate": 0.00019157789890504847,
      "loss": 0.9207,
      "mean_token_accuracy": 0.8103202067315578,
      "num_tokens": 10299354.0,
      "step": 1340
    },
    {
      "entropy": 0.9478652633726596,
      "epoch": 0.19636363636363635,
      "grad_norm": 0.3512885272502899,
      "learning_rate": 0.0001913788913162877,
      "loss": 0.9331,
      "mean_token_accuracy": 0.8110518783330918,
      "num_tokens": 10376102.0,
      "step": 1350
    },
    {
      "entropy": 0.9545814357697964,
      "epoch": 0.1978181818181818,
      "grad_norm": 0.3496834635734558,
      "learning_rate": 0.00019117766580345832,
      "loss": 0.9344,
      "mean_token_accuracy": 0.8108161889016628,
      "num_tokens": 10452922.0,
      "step": 1360
    },
    {
      "entropy": 0.9444903768599033,
      "epoch": 0.19927272727272727,
      "grad_norm": 0.33390098810195923,
      "learning_rate": 0.0001909742272533262,
      "loss": 0.9189,
      "mean_token_accuracy": 0.8123569242656231,
      "num_tokens": 10529714.0,
      "step": 1370
    },
    {
      "entropy": 0.9337248101830482,
      "epoch": 0.20072727272727273,
      "grad_norm": 0.4034648537635803,
      "learning_rate": 0.00019076858060640083,
      "loss": 0.9289,
      "mean_token_accuracy": 0.8110757023096085,
      "num_tokens": 10606511.0,
      "step": 1380
    },
    {
      "entropy": 0.9631103947758675,
      "epoch": 0.20218181818181818,
      "grad_norm": 0.32546374201774597,
      "learning_rate": 0.00019056073085681543,
      "loss": 0.9318,
      "mean_token_accuracy": 0.8103945441544056,
      "num_tokens": 10682172.0,
      "step": 1390
    },
    {
      "entropy": 0.9652538903057575,
      "epoch": 0.20363636363636364,
      "grad_norm": 0.3356182277202606,
      "learning_rate": 0.00019035068305220555,
      "loss": 0.954,
      "mean_token_accuracy": 0.806369586288929,
      "num_tokens": 10760611.0,
      "step": 1400
    },
    {
      "entropy": 0.959171961992979,
      "epoch": 0.2050909090909091,
      "grad_norm": 0.4710279405117035,
      "learning_rate": 0.00019013844229358667,
      "loss": 0.9416,
      "mean_token_accuracy": 0.8092709675431251,
      "num_tokens": 10839415.0,
      "step": 1410
    },
    {
      "entropy": 0.9554042249917984,
      "epoch": 0.20654545454545453,
      "grad_norm": 0.39765721559524536,
      "learning_rate": 0.00018992401373523006,
      "loss": 0.9311,
      "mean_token_accuracy": 0.8113908626139164,
      "num_tokens": 10917800.0,
      "step": 1420
    },
    {
      "entropy": 0.9345662452280521,
      "epoch": 0.208,
      "grad_norm": 0.3606942892074585,
      "learning_rate": 0.00018970740258453784,
      "loss": 0.9218,
      "mean_token_accuracy": 0.8098172180354595,
      "num_tokens": 10994499.0,
      "step": 1430
    },
    {
      "entropy": 0.9311383411288261,
      "epoch": 0.20945454545454545,
      "grad_norm": 0.3340699374675751,
      "learning_rate": 0.0001894886141019164,
      "loss": 0.9041,
      "mean_token_accuracy": 0.8147461965680123,
      "num_tokens": 11069175.0,
      "step": 1440
    },
    {
      "entropy": 0.9625661939382553,
      "epoch": 0.2109090909090909,
      "grad_norm": 0.3354319632053375,
      "learning_rate": 0.00018926765360064877,
      "loss": 0.9407,
      "mean_token_accuracy": 0.8064736150205135,
      "num_tokens": 11146085.0,
      "step": 1450
    },
    {
      "entropy": 0.9429851680994034,
      "epoch": 0.21236363636363637,
      "grad_norm": 0.32050856947898865,
      "learning_rate": 0.0001890445264467653,
      "loss": 0.9253,
      "mean_token_accuracy": 0.8105651929974556,
      "num_tokens": 11221965.0,
      "step": 1460
    },
    {
      "entropy": 0.930499044060707,
      "epoch": 0.21381818181818182,
      "grad_norm": 0.35083913803100586,
      "learning_rate": 0.0001888192380589138,
      "loss": 0.9082,
      "mean_token_accuracy": 0.8133820489048957,
      "num_tokens": 11297583.0,
      "step": 1470
    },
    {
      "entropy": 0.9538578160107136,
      "epoch": 0.21527272727272728,
      "grad_norm": 0.3205820918083191,
      "learning_rate": 0.00018859179390822746,
      "loss": 0.9313,
      "mean_token_accuracy": 0.8103922225534916,
      "num_tokens": 11373362.0,
      "step": 1480
    },
    {
      "entropy": 0.9458141349256038,
      "epoch": 0.21672727272727274,
      "grad_norm": 0.34689080715179443,
      "learning_rate": 0.00018836219951819236,
      "loss": 0.9188,
      "mean_token_accuracy": 0.8122154355049134,
      "num_tokens": 11449026.0,
      "step": 1490
    },
    {
      "entropy": 0.9495655484497547,
      "epoch": 0.21818181818181817,
      "grad_norm": 0.3351272642612457,
      "learning_rate": 0.00018813046046451312,
      "loss": 0.9309,
      "mean_token_accuracy": 0.8108266346156597,
      "num_tokens": 11529710.0,
      "step": 1500
    },
    {
      "entropy": 0.9607556290924549,
      "epoch": 0.21963636363636363,
      "grad_norm": 0.2975183129310608,
      "learning_rate": 0.00018789658237497765,
      "loss": 0.9472,
      "mean_token_accuracy": 0.8094669461250306,
      "num_tokens": 11606684.0,
      "step": 1510
    },
    {
      "entropy": 0.9780306808650494,
      "epoch": 0.2210909090909091,
      "grad_norm": 0.30866914987564087,
      "learning_rate": 0.00018766057092932026,
      "loss": 0.9885,
      "mean_token_accuracy": 0.8058537535369397,
      "num_tokens": 11686025.0,
      "step": 1520
    },
    {
      "entropy": 0.9589258328080177,
      "epoch": 0.22254545454545455,
      "grad_norm": 0.2969805896282196,
      "learning_rate": 0.00018742243185908397,
      "loss": 0.928,
      "mean_token_accuracy": 0.8105280227959156,
      "num_tokens": 11761670.0,
      "step": 1530
    },
    {
      "entropy": 0.9614237651228905,
      "epoch": 0.224,
      "grad_norm": 0.3447507321834564,
      "learning_rate": 0.00018718217094748117,
      "loss": 0.9589,
      "mean_token_accuracy": 0.8076143585145473,
      "num_tokens": 11840949.0,
      "step": 1540
    },
    {
      "entropy": 0.9415218912065029,
      "epoch": 0.22545454545454546,
      "grad_norm": 0.4087255597114563,
      "learning_rate": 0.00018693979402925316,
      "loss": 0.9412,
      "mean_token_accuracy": 0.8099103234708309,
      "num_tokens": 11918454.0,
      "step": 1550
    },
    {
      "entropy": 0.9644451148808002,
      "epoch": 0.22690909090909092,
      "grad_norm": 0.36078566312789917,
      "learning_rate": 0.00018669530699052863,
      "loss": 0.9429,
      "mean_token_accuracy": 0.8090198636054993,
      "num_tokens": 11997450.0,
      "step": 1560
    },
    {
      "entropy": 0.9489136710762978,
      "epoch": 0.22836363636363635,
      "grad_norm": 0.3263258635997772,
      "learning_rate": 0.00018644871576868047,
      "loss": 0.9294,
      "mean_token_accuracy": 0.8090980783104896,
      "num_tokens": 12073862.0,
      "step": 1570
    },
    {
      "entropy": 0.9820632249116897,
      "epoch": 0.2298181818181818,
      "grad_norm": 0.3392266631126404,
      "learning_rate": 0.00018620002635218175,
      "loss": 0.9619,
      "mean_token_accuracy": 0.8049184948205947,
      "num_tokens": 12154103.0,
      "step": 1580
    },
    {
      "entropy": 0.9613321676850319,
      "epoch": 0.23127272727272727,
      "grad_norm": 0.3108541965484619,
      "learning_rate": 0.00018594924478046026,
      "loss": 0.9361,
      "mean_token_accuracy": 0.8093224719166756,
      "num_tokens": 12230848.0,
      "step": 1590
    },
    {
      "entropy": 0.9354442425072194,
      "epoch": 0.23272727272727273,
      "grad_norm": 0.3270244896411896,
      "learning_rate": 0.00018569637714375178,
      "loss": 0.919,
      "mean_token_accuracy": 0.810560081154108,
      "num_tokens": 12308605.0,
      "step": 1600
    },
    {
      "entropy": 0.990824494510889,
      "epoch": 0.23418181818181819,
      "grad_norm": 0.31900009512901306,
      "learning_rate": 0.00018544142958295225,
      "loss": 0.9741,
      "mean_token_accuracy": 0.8050805293023586,
      "num_tokens": 12387234.0,
      "step": 1610
    },
    {
      "entropy": 0.932359478622675,
      "epoch": 0.23563636363636364,
      "grad_norm": 0.41305163502693176,
      "learning_rate": 0.00018518440828946863,
      "loss": 0.9181,
      "mean_token_accuracy": 0.8122582986950875,
      "num_tokens": 12463203.0,
      "step": 1620
    },
    {
      "entropy": 0.9484110824763775,
      "epoch": 0.2370909090909091,
      "grad_norm": 0.3127359449863434,
      "learning_rate": 0.0001849253195050685,
      "loss": 0.9259,
      "mean_token_accuracy": 0.8118439868092537,
      "num_tokens": 12539819.0,
      "step": 1630
    },
    {
      "entropy": 0.9479153640568256,
      "epoch": 0.23854545454545453,
      "grad_norm": 0.3151299059391022,
      "learning_rate": 0.00018466416952172846,
      "loss": 0.9253,
      "mean_token_accuracy": 0.808732783049345,
      "num_tokens": 12619081.0,
      "step": 1640
    },
    {
      "entropy": 0.9727005705237388,
      "epoch": 0.24,
      "grad_norm": 0.32946428656578064,
      "learning_rate": 0.0001844009646814815,
      "loss": 0.9607,
      "mean_token_accuracy": 0.8065188810229301,
      "num_tokens": 12696222.0,
      "step": 1650
    },
    {
      "entropy": 0.9596579805016517,
      "epoch": 0.24145454545454545,
      "grad_norm": 0.3054264783859253,
      "learning_rate": 0.00018413571137626265,
      "loss": 0.9346,
      "mean_token_accuracy": 0.8099383614957333,
      "num_tokens": 12770009.0,
      "step": 1660
    },
    {
      "entropy": 0.9318838283419609,
      "epoch": 0.2429090909090909,
      "grad_norm": 0.38207921385765076,
      "learning_rate": 0.00018386841604775415,
      "loss": 0.9114,
      "mean_token_accuracy": 0.8137748256325722,
      "num_tokens": 12845015.0,
      "step": 1670
    },
    {
      "entropy": 0.9751473486423492,
      "epoch": 0.24436363636363637,
      "grad_norm": 0.3861842751502991,
      "learning_rate": 0.0001835990851872287,
      "loss": 0.9493,
      "mean_token_accuracy": 0.8080014623701572,
      "num_tokens": 12922447.0,
      "step": 1680
    },
    {
      "entropy": 0.9198803529143333,
      "epoch": 0.24581818181818182,
      "grad_norm": 0.3217521011829376,
      "learning_rate": 0.000183327725335392,
      "loss": 0.9063,
      "mean_token_accuracy": 0.8148512996733188,
      "num_tokens": 12999334.0,
      "step": 1690
    },
    {
      "entropy": 0.9567082822322845,
      "epoch": 0.24727272727272728,
      "grad_norm": 0.34669220447540283,
      "learning_rate": 0.00018305434308222383,
      "loss": 0.9382,
      "mean_token_accuracy": 0.8091933868825436,
      "num_tokens": 13076110.0,
      "step": 1700
    },
    {
      "entropy": 0.9501121707260609,
      "epoch": 0.2487272727272727,
      "grad_norm": 0.3057405352592468,
      "learning_rate": 0.000182778945066818,
      "loss": 0.9297,
      "mean_token_accuracy": 0.81047957316041,
      "num_tokens": 13151198.0,
      "step": 1710
    },
    {
      "entropy": 0.9429646603763103,
      "epoch": 0.25018181818181817,
      "grad_norm": 0.3221091628074646,
      "learning_rate": 0.00018250153797722118,
      "loss": 0.9311,
      "mean_token_accuracy": 0.8111985780298709,
      "num_tokens": 13226772.0,
      "step": 1720
    },
    {
      "entropy": 0.9599173679947853,
      "epoch": 0.25163636363636366,
      "grad_norm": 0.31713050603866577,
      "learning_rate": 0.00018222212855027045,
      "loss": 0.9345,
      "mean_token_accuracy": 0.8075641810894012,
      "num_tokens": 13306183.0,
      "step": 1730
    },
    {
      "entropy": 0.9429551161825657,
      "epoch": 0.2530909090909091,
      "grad_norm": 0.3050651252269745,
      "learning_rate": 0.00018194072357142967,
      "loss": 0.9326,
      "mean_token_accuracy": 0.8090726666152477,
      "num_tokens": 13384697.0,
      "step": 1740
    },
    {
      "entropy": 0.9429156497120857,
      "epoch": 0.2545454545454545,
      "grad_norm": 0.404987633228302,
      "learning_rate": 0.00018165732987462482,
      "loss": 0.9109,
      "mean_token_accuracy": 0.8144759446382522,
      "num_tokens": 13460101.0,
      "step": 1750
    },
    {
      "entropy": 0.9501292340457439,
      "epoch": 0.256,
      "grad_norm": 0.3385776877403259,
      "learning_rate": 0.00018137195434207772,
      "loss": 0.9283,
      "mean_token_accuracy": 0.8100073337554932,
      "num_tokens": 13536659.0,
      "step": 1760
    },
    {
      "entropy": 0.9687538675963878,
      "epoch": 0.25745454545454544,
      "grad_norm": 0.3240547478199005,
      "learning_rate": 0.00018108460390413937,
      "loss": 0.9518,
      "mean_token_accuracy": 0.806635819375515,
      "num_tokens": 13615612.0,
      "step": 1770
    },
    {
      "entropy": 0.9552951484918595,
      "epoch": 0.2589090909090909,
      "grad_norm": 0.32524409890174866,
      "learning_rate": 0.0001807952855391212,
      "loss": 0.9224,
      "mean_token_accuracy": 0.8102507315576076,
      "num_tokens": 13691703.0,
      "step": 1780
    },
    {
      "entropy": 0.9637807898223401,
      "epoch": 0.26036363636363635,
      "grad_norm": 0.32301396131515503,
      "learning_rate": 0.00018050400627312588,
      "loss": 0.9489,
      "mean_token_accuracy": 0.8067096285521984,
      "num_tokens": 13770988.0,
      "step": 1790
    },
    {
      "entropy": 0.9319825753569603,
      "epoch": 0.26181818181818184,
      "grad_norm": 0.3573399484157562,
      "learning_rate": 0.00018021077317987655,
      "loss": 0.9117,
      "mean_token_accuracy": 0.8137773007154465,
      "num_tokens": 13845817.0,
      "step": 1800
    },
    {
      "entropy": 0.9618785664439201,
      "epoch": 0.26327272727272727,
      "grad_norm": 0.33567363023757935,
      "learning_rate": 0.00017991559338054517,
      "loss": 0.9512,
      "mean_token_accuracy": 0.8086851388216019,
      "num_tokens": 13921319.0,
      "step": 1810
    },
    {
      "entropy": 0.960854459553957,
      "epoch": 0.2647272727272727,
      "grad_norm": 0.37292754650115967,
      "learning_rate": 0.00017961847404357944,
      "loss": 0.938,
      "mean_token_accuracy": 0.8094214454293251,
      "num_tokens": 13998752.0,
      "step": 1820
    },
    {
      "entropy": 0.9524516589939594,
      "epoch": 0.2661818181818182,
      "grad_norm": 0.3304807245731354,
      "learning_rate": 0.00017931942238452874,
      "loss": 0.938,
      "mean_token_accuracy": 0.8107712835073471,
      "num_tokens": 14077253.0,
      "step": 1830
    },
    {
      "entropy": 0.9252058751881123,
      "epoch": 0.2676363636363636,
      "grad_norm": 0.33476701378822327,
      "learning_rate": 0.000179018445665869,
      "loss": 0.8945,
      "mean_token_accuracy": 0.8162094913423061,
      "num_tokens": 14150885.0,
      "step": 1840
    },
    {
      "entropy": 1.0102965272963047,
      "epoch": 0.2690909090909091,
      "grad_norm": 0.3192066550254822,
      "learning_rate": 0.00017871555119682625,
      "loss": 1.0206,
      "mean_token_accuracy": 0.8030799396336079,
      "num_tokens": 14229299.0,
      "step": 1850
    },
    {
      "entropy": 0.9697954557836056,
      "epoch": 0.27054545454545453,
      "grad_norm": 0.34754517674446106,
      "learning_rate": 0.00017841074633319904,
      "loss": 0.9557,
      "mean_token_accuracy": 0.8078889131546021,
      "num_tokens": 14304823.0,
      "step": 1860
    },
    {
      "entropy": 0.9449729457497597,
      "epoch": 0.272,
      "grad_norm": 0.3253391981124878,
      "learning_rate": 0.0001781040384771801,
      "loss": 0.918,
      "mean_token_accuracy": 0.8116657704114913,
      "num_tokens": 14380159.0,
      "step": 1870
    },
    {
      "entropy": 0.9418261803686618,
      "epoch": 0.27345454545454545,
      "grad_norm": 0.30982857942581177,
      "learning_rate": 0.00017779543507717612,
      "loss": 0.9228,
      "mean_token_accuracy": 0.8130053669214249,
      "num_tokens": 14455983.0,
      "step": 1880
    },
    {
      "entropy": 0.9387239389121532,
      "epoch": 0.27490909090909094,
      "grad_norm": 0.35093173384666443,
      "learning_rate": 0.0001774849436276273,
      "loss": 0.9102,
      "mean_token_accuracy": 0.8151706494390965,
      "num_tokens": 14531144.0,
      "step": 1890
    },
    {
      "entropy": 0.9262372322380543,
      "epoch": 0.27636363636363637,
      "grad_norm": 0.3041590452194214,
      "learning_rate": 0.00017717257166882503,
      "loss": 0.9008,
      "mean_token_accuracy": 0.8168079420924187,
      "num_tokens": 14604877.0,
      "step": 1900
    },
    {
      "entropy": 0.9444479495286942,
      "epoch": 0.2778181818181818,
      "grad_norm": 0.3268347382545471,
      "learning_rate": 0.00017685832678672906,
      "loss": 0.9324,
      "mean_token_accuracy": 0.8125970877707005,
      "num_tokens": 14682285.0,
      "step": 1910
    },
    {
      "entropy": 0.9417746476829052,
      "epoch": 0.2792727272727273,
      "grad_norm": 0.34488826990127563,
      "learning_rate": 0.000176542216612783,
      "loss": 0.9103,
      "mean_token_accuracy": 0.8135712221264839,
      "num_tokens": 14758010.0,
      "step": 1920
    },
    {
      "entropy": 0.9153248682618141,
      "epoch": 0.2807272727272727,
      "grad_norm": 0.327629953622818,
      "learning_rate": 0.00017622424882372912,
      "loss": 0.8994,
      "mean_token_accuracy": 0.8184183113276958,
      "num_tokens": 14831408.0,
      "step": 1930
    },
    {
      "entropy": 0.9517436102032661,
      "epoch": 0.2821818181818182,
      "grad_norm": 0.3337823152542114,
      "learning_rate": 0.000175904431141422,
      "loss": 0.927,
      "mean_token_accuracy": 0.8096821196377277,
      "num_tokens": 14910742.0,
      "step": 1940
    },
    {
      "entropy": 0.9344942085444927,
      "epoch": 0.28363636363636363,
      "grad_norm": 0.38580557703971863,
      "learning_rate": 0.0001755827713326408,
      "loss": 0.9282,
      "mean_token_accuracy": 0.8125026755034923,
      "num_tokens": 14984373.0,
      "step": 1950
    },
    {
      "entropy": 0.9408765397965908,
      "epoch": 0.2850909090909091,
      "grad_norm": 0.35432329773902893,
      "learning_rate": 0.00017525927720890086,
      "loss": 0.9051,
      "mean_token_accuracy": 0.8168821662664414,
      "num_tokens": 15058109.0,
      "step": 1960
    },
    {
      "entropy": 0.9451767437160015,
      "epoch": 0.28654545454545455,
      "grad_norm": 0.37497565150260925,
      "learning_rate": 0.00017493395662626383,
      "loss": 0.927,
      "mean_token_accuracy": 0.8118024595081806,
      "num_tokens": 15134759.0,
      "step": 1970
    },
    {
      "entropy": 0.9334485501050949,
      "epoch": 0.288,
      "grad_norm": 0.3791519105434418,
      "learning_rate": 0.00017460681748514701,
      "loss": 0.9232,
      "mean_token_accuracy": 0.811272544413805,
      "num_tokens": 15212159.0,
      "step": 1980
    },
    {
      "entropy": 0.965351027995348,
      "epoch": 0.28945454545454546,
      "grad_norm": 0.32630297541618347,
      "learning_rate": 0.0001742778677301314,
      "loss": 0.9295,
      "mean_token_accuracy": 0.8107661835849285,
      "num_tokens": 15290445.0,
      "step": 1990
    },
    {
      "entropy": 0.960039559751749,
      "epoch": 0.2909090909090909,
      "grad_norm": 0.3436606526374817,
      "learning_rate": 0.00017394711534976875,
      "loss": 0.9372,
      "mean_token_accuracy": 0.8104394495487213,
      "num_tokens": 15368377.0,
      "step": 2000
    },
    {
      "epoch": 0.2909090909090909,
      "eval_entropy": 0.9545571558952332,
      "eval_loss": 0.9135851860046387,
      "eval_mean_token_accuracy": 0.8095944012641907,
      "eval_num_tokens": 15368377.0,
      "eval_runtime": 78.1006,
      "eval_samples_per_second": 64.02,
      "eval_steps_per_second": 8.002,
      "step": 2000
    },
    {
      "entropy": 0.93552967607975,
      "epoch": 0.2923636363636364,
      "grad_norm": 0.3713809847831726,
      "learning_rate": 0.00017361456837638773,
      "loss": 0.9054,
      "mean_token_accuracy": 0.8160525418817997,
      "num_tokens": 15443459.0,
      "step": 2010
    },
    {
      "entropy": 0.9752048850059509,
      "epoch": 0.2938181818181818,
      "grad_norm": 0.3389149308204651,
      "learning_rate": 0.0001732802348858986,
      "loss": 0.9717,
      "mean_token_accuracy": 0.8063795827329159,
      "num_tokens": 15521440.0,
      "step": 2020
    },
    {
      "entropy": 0.9328745149075985,
      "epoch": 0.2952727272727273,
      "grad_norm": 0.307253360748291,
      "learning_rate": 0.0001729441229975973,
      "loss": 0.9088,
      "mean_token_accuracy": 0.8136720769107342,
      "num_tokens": 15596124.0,
      "step": 2030
    },
    {
      "entropy": 0.9618542812764644,
      "epoch": 0.29672727272727273,
      "grad_norm": 0.3493705093860626,
      "learning_rate": 0.0001726062408739682,
      "loss": 0.9413,
      "mean_token_accuracy": 0.8102415710687637,
      "num_tokens": 15673126.0,
      "step": 2040
    },
    {
      "entropy": 0.9472650431096554,
      "epoch": 0.29818181818181816,
      "grad_norm": 0.3055593967437744,
      "learning_rate": 0.0001722665967204859,
      "loss": 0.93,
      "mean_token_accuracy": 0.8135502748191357,
      "num_tokens": 15748318.0,
      "step": 2050
    },
    {
      "entropy": 0.9465427048504352,
      "epoch": 0.29963636363636365,
      "grad_norm": 0.3755161762237549,
      "learning_rate": 0.0001719251987854158,
      "loss": 0.9201,
      "mean_token_accuracy": 0.8124631099402905,
      "num_tokens": 15824620.0,
      "step": 2060
    },
    {
      "entropy": 0.9456887245178223,
      "epoch": 0.3010909090909091,
      "grad_norm": 0.32835713028907776,
      "learning_rate": 0.0001715820553596141,
      "loss": 0.9368,
      "mean_token_accuracy": 0.8130617618560791,
      "num_tokens": 15899968.0,
      "step": 2070
    },
    {
      "entropy": 0.9543628983199597,
      "epoch": 0.30254545454545456,
      "grad_norm": 0.3607058525085449,
      "learning_rate": 0.00017123717477632617,
      "loss": 0.925,
      "mean_token_accuracy": 0.8120505645871162,
      "num_tokens": 15979356.0,
      "step": 2080
    },
    {
      "entropy": 0.9334946945309639,
      "epoch": 0.304,
      "grad_norm": 0.3168565332889557,
      "learning_rate": 0.00017089056541098432,
      "loss": 0.9073,
      "mean_token_accuracy": 0.8149824872612953,
      "num_tokens": 16056328.0,
      "step": 2090
    },
    {
      "entropy": 0.9526524029672145,
      "epoch": 0.3054545454545455,
      "grad_norm": 0.3339099586009979,
      "learning_rate": 0.00017054223568100432,
      "loss": 0.9332,
      "mean_token_accuracy": 0.8111643262207509,
      "num_tokens": 16133661.0,
      "step": 2100
    },
    {
      "entropy": 0.9297716051340104,
      "epoch": 0.3069090909090909,
      "grad_norm": 0.37411344051361084,
      "learning_rate": 0.00017019219404558112,
      "loss": 0.9131,
      "mean_token_accuracy": 0.8139175310730934,
      "num_tokens": 16208975.0,
      "step": 2110
    },
    {
      "entropy": 0.9196740917861461,
      "epoch": 0.30836363636363634,
      "grad_norm": 0.3221561908721924,
      "learning_rate": 0.00016984044900548325,
      "loss": 0.892,
      "mean_token_accuracy": 0.8175107464194298,
      "num_tokens": 16285565.0,
      "step": 2120
    },
    {
      "entropy": 0.9326213136315346,
      "epoch": 0.3098181818181818,
      "grad_norm": 0.3141707181930542,
      "learning_rate": 0.00016948700910284653,
      "loss": 0.9109,
      "mean_token_accuracy": 0.814671640843153,
      "num_tokens": 16360976.0,
      "step": 2130
    },
    {
      "entropy": 0.9365306921303272,
      "epoch": 0.31127272727272726,
      "grad_norm": 0.5048186779022217,
      "learning_rate": 0.0001691318829209665,
      "loss": 0.9152,
      "mean_token_accuracy": 0.8133588239550591,
      "num_tokens": 16435360.0,
      "step": 2140
    },
    {
      "entropy": 0.9475600443780422,
      "epoch": 0.31272727272727274,
      "grad_norm": 0.33213672041893005,
      "learning_rate": 0.00016877507908409,
      "loss": 0.9127,
      "mean_token_accuracy": 0.813559490442276,
      "num_tokens": 16510866.0,
      "step": 2150
    },
    {
      "entropy": 0.9257668554782867,
      "epoch": 0.3141818181818182,
      "grad_norm": 0.3788197934627533,
      "learning_rate": 0.00016841660625720585,
      "loss": 0.9233,
      "mean_token_accuracy": 0.811751264333725,
      "num_tokens": 16586819.0,
      "step": 2160
    },
    {
      "entropy": 0.9617751978337765,
      "epoch": 0.31563636363636366,
      "grad_norm": 0.4031602442264557,
      "learning_rate": 0.00016805647314583428,
      "loss": 0.9339,
      "mean_token_accuracy": 0.8118709713220597,
      "num_tokens": 16663718.0,
      "step": 2170
    },
    {
      "entropy": 0.9458882667124271,
      "epoch": 0.3170909090909091,
      "grad_norm": 0.3160066306591034,
      "learning_rate": 0.0001676946884958155,
      "loss": 0.9418,
      "mean_token_accuracy": 0.8106794603168964,
      "num_tokens": 16739810.0,
      "step": 2180
    },
    {
      "entropy": 0.9457833237946034,
      "epoch": 0.3185454545454545,
      "grad_norm": 0.34538161754608154,
      "learning_rate": 0.00016733126109309753,
      "loss": 0.9234,
      "mean_token_accuracy": 0.8111150979995727,
      "num_tokens": 16817590.0,
      "step": 2190
    },
    {
      "entropy": 0.9471365615725518,
      "epoch": 0.32,
      "grad_norm": 0.32485249638557434,
      "learning_rate": 0.0001669661997635225,
      "loss": 0.938,
      "mean_token_accuracy": 0.8131156258285046,
      "num_tokens": 16892279.0,
      "step": 2200
    },
    {
      "entropy": 0.9507883109152317,
      "epoch": 0.32145454545454544,
      "grad_norm": 0.3409273326396942,
      "learning_rate": 0.00016659951337261256,
      "loss": 0.9323,
      "mean_token_accuracy": 0.8105610504746437,
      "num_tokens": 16971570.0,
      "step": 2210
    },
    {
      "entropy": 0.9707157976925374,
      "epoch": 0.3229090909090909,
      "grad_norm": 0.4358591139316559,
      "learning_rate": 0.0001662312108253546,
      "loss": 0.9563,
      "mean_token_accuracy": 0.808458685129881,
      "num_tokens": 17048391.0,
      "step": 2220
    },
    {
      "entropy": 0.962416484951973,
      "epoch": 0.32436363636363635,
      "grad_norm": 0.32611650228500366,
      "learning_rate": 0.00016586130106598372,
      "loss": 0.9379,
      "mean_token_accuracy": 0.8086472131311894,
      "num_tokens": 17125614.0,
      "step": 2230
    },
    {
      "entropy": 0.923296982795,
      "epoch": 0.32581818181818184,
      "grad_norm": 0.38609233498573303,
      "learning_rate": 0.00016548979307776636,
      "loss": 0.9116,
      "mean_token_accuracy": 0.8152348309755325,
      "num_tokens": 17201164.0,
      "step": 2240
    },
    {
      "entropy": 0.9433801986277104,
      "epoch": 0.32727272727272727,
      "grad_norm": 0.3359317481517792,
      "learning_rate": 0.00016511669588278188,
      "loss": 0.9247,
      "mean_token_accuracy": 0.8113768070936203,
      "num_tokens": 17278206.0,
      "step": 2250
    },
    {
      "entropy": 0.9336811497807502,
      "epoch": 0.3287272727272727,
      "grad_norm": 0.31968075037002563,
      "learning_rate": 0.00016474201854170358,
      "loss": 0.9103,
      "mean_token_accuracy": 0.8159900955855847,
      "num_tokens": 17354469.0,
      "step": 2260
    },
    {
      "entropy": 0.9368410244584083,
      "epoch": 0.3301818181818182,
      "grad_norm": 0.32750600576400757,
      "learning_rate": 0.00016436577015357872,
      "loss": 0.9207,
      "mean_token_accuracy": 0.8116543501615524,
      "num_tokens": 17432087.0,
      "step": 2270
    },
    {
      "entropy": 0.9387645564973355,
      "epoch": 0.3316363636363636,
      "grad_norm": 0.32613465189933777,
      "learning_rate": 0.00016398795985560736,
      "loss": 0.9177,
      "mean_token_accuracy": 0.8138992644846439,
      "num_tokens": 17507413.0,
      "step": 2280
    },
    {
      "entropy": 0.955746828019619,
      "epoch": 0.3330909090909091,
      "grad_norm": 0.3252658545970917,
      "learning_rate": 0.00016360859682292066,
      "loss": 0.9273,
      "mean_token_accuracy": 0.8117515601217746,
      "num_tokens": 17586012.0,
      "step": 2290
    },
    {
      "entropy": 0.9415918923914433,
      "epoch": 0.33454545454545453,
      "grad_norm": 0.3188813030719757,
      "learning_rate": 0.00016322769026835794,
      "loss": 0.9323,
      "mean_token_accuracy": 0.8117557033896446,
      "num_tokens": 17664213.0,
      "step": 2300
    },
    {
      "entropy": 0.9359489880502224,
      "epoch": 0.336,
      "grad_norm": 0.33735203742980957,
      "learning_rate": 0.00016284524944224296,
      "loss": 0.9136,
      "mean_token_accuracy": 0.8152964189648628,
      "num_tokens": 17739740.0,
      "step": 2310
    },
    {
      "entropy": 0.9483966782689095,
      "epoch": 0.33745454545454545,
      "grad_norm": 0.33717650175094604,
      "learning_rate": 0.00016246128363215926,
      "loss": 0.9323,
      "mean_token_accuracy": 0.8132497929036617,
      "num_tokens": 17816245.0,
      "step": 2320
    },
    {
      "entropy": 0.9411058373749256,
      "epoch": 0.3389090909090909,
      "grad_norm": 0.32491806149482727,
      "learning_rate": 0.00016207580216272483,
      "loss": 0.9234,
      "mean_token_accuracy": 0.8120617903769016,
      "num_tokens": 17894511.0,
      "step": 2330
    },
    {
      "entropy": 0.9490449778735638,
      "epoch": 0.34036363636363637,
      "grad_norm": 0.3255004286766052,
      "learning_rate": 0.00016168881439536527,
      "loss": 0.9261,
      "mean_token_accuracy": 0.8118731305003166,
      "num_tokens": 17971114.0,
      "step": 2340
    },
    {
      "entropy": 0.9435882270336151,
      "epoch": 0.3418181818181818,
      "grad_norm": 0.3351687490940094,
      "learning_rate": 0.0001613003297280868,
      "loss": 0.9258,
      "mean_token_accuracy": 0.8116018794476986,
      "num_tokens": 18049687.0,
      "step": 2350
    },
    {
      "entropy": 0.9387622274458408,
      "epoch": 0.3432727272727273,
      "grad_norm": 0.358738511800766,
      "learning_rate": 0.0001609103575952478,
      "loss": 0.9196,
      "mean_token_accuracy": 0.813437120616436,
      "num_tokens": 18124677.0,
      "step": 2360
    },
    {
      "entropy": 0.92884556427598,
      "epoch": 0.3447272727272727,
      "grad_norm": 0.38657838106155396,
      "learning_rate": 0.00016051890746732978,
      "loss": 0.9007,
      "mean_token_accuracy": 0.8151167511940003,
      "num_tokens": 18201388.0,
      "step": 2370
    },
    {
      "entropy": 0.9288479879498481,
      "epoch": 0.3461818181818182,
      "grad_norm": 0.387179970741272,
      "learning_rate": 0.00016012598885070748,
      "loss": 0.9131,
      "mean_token_accuracy": 0.8137955233454705,
      "num_tokens": 18275995.0,
      "step": 2380
    },
    {
      "entropy": 0.9669632643461228,
      "epoch": 0.34763636363636363,
      "grad_norm": 0.3194025158882141,
      "learning_rate": 0.0001597316112874178,
      "loss": 0.9505,
      "mean_token_accuracy": 0.8098523296415806,
      "num_tokens": 18351968.0,
      "step": 2390
    },
    {
      "entropy": 0.9311579629778862,
      "epoch": 0.3490909090909091,
      "grad_norm": 0.3317321538925171,
      "learning_rate": 0.00015933578435492834,
      "loss": 0.9173,
      "mean_token_accuracy": 0.8136695794761181,
      "num_tokens": 18429882.0,
      "step": 2400
    },
    {
      "entropy": 0.9734658777713776,
      "epoch": 0.35054545454545455,
      "grad_norm": 0.33316028118133545,
      "learning_rate": 0.00015893851766590454,
      "loss": 0.9715,
      "mean_token_accuracy": 0.8081215254962444,
      "num_tokens": 18506409.0,
      "step": 2410
    },
    {
      "entropy": 0.9483408346772194,
      "epoch": 0.352,
      "grad_norm": 0.3269667625427246,
      "learning_rate": 0.00015853982086797632,
      "loss": 0.9327,
      "mean_token_accuracy": 0.8124340586364269,
      "num_tokens": 18584276.0,
      "step": 2420
    },
    {
      "entropy": 0.9598711885511875,
      "epoch": 0.35345454545454547,
      "grad_norm": 0.35559943318367004,
      "learning_rate": 0.00015813970364350404,
      "loss": 0.934,
      "mean_token_accuracy": 0.8129792243242264,
      "num_tokens": 18663675.0,
      "step": 2430
    },
    {
      "entropy": 0.9243545681238174,
      "epoch": 0.3549090909090909,
      "grad_norm": 0.3285573422908783,
      "learning_rate": 0.00015773817570934287,
      "loss": 0.9103,
      "mean_token_accuracy": 0.8160747058689595,
      "num_tokens": 18739787.0,
      "step": 2440
    },
    {
      "entropy": 0.9421109184622765,
      "epoch": 0.3563636363636364,
      "grad_norm": 0.3947218954563141,
      "learning_rate": 0.00015733524681660734,
      "loss": 0.9155,
      "mean_token_accuracy": 0.8172662526369094,
      "num_tokens": 18816021.0,
      "step": 2450
    },
    {
      "entropy": 0.94013966396451,
      "epoch": 0.3578181818181818,
      "grad_norm": 0.33459532260894775,
      "learning_rate": 0.0001569309267504341,
      "loss": 0.9249,
      "mean_token_accuracy": 0.8147119626402854,
      "num_tokens": 18894131.0,
      "step": 2460
    },
    {
      "entropy": 0.93881671205163,
      "epoch": 0.3592727272727273,
      "grad_norm": 0.532590389251709,
      "learning_rate": 0.00015652522532974463,
      "loss": 0.9183,
      "mean_token_accuracy": 0.8175694108009338,
      "num_tokens": 18971093.0,
      "step": 2470
    },
    {
      "entropy": 0.9401829384267331,
      "epoch": 0.36072727272727273,
      "grad_norm": 0.3573850691318512,
      "learning_rate": 0.00015611815240700657,
      "loss": 0.9116,
      "mean_token_accuracy": 0.8167402453720569,
      "num_tokens": 19046555.0,
      "step": 2480
    },
    {
      "entropy": 0.9053499557077884,
      "epoch": 0.36218181818181816,
      "grad_norm": 0.37487003207206726,
      "learning_rate": 0.0001557097178679944,
      "loss": 0.8757,
      "mean_token_accuracy": 0.8230059690773487,
      "num_tokens": 19119410.0,
      "step": 2490
    },
    {
      "entropy": 0.9198244661092758,
      "epoch": 0.36363636363636365,
      "grad_norm": 0.3617284297943115,
      "learning_rate": 0.0001552999316315497,
      "loss": 0.8972,
      "mean_token_accuracy": 0.8209393374621868,
      "num_tokens": 19196135.0,
      "step": 2500
    },
    {
      "entropy": 0.9223854258656502,
      "epoch": 0.3650909090909091,
      "grad_norm": 0.4283069968223572,
      "learning_rate": 0.00015488880364933986,
      "loss": 0.9079,
      "mean_token_accuracy": 0.8193068109452725,
      "num_tokens": 19271772.0,
      "step": 2510
    },
    {
      "entropy": 0.9177697144448758,
      "epoch": 0.36654545454545456,
      "grad_norm": 0.3647020757198334,
      "learning_rate": 0.00015447634390561673,
      "loss": 0.8898,
      "mean_token_accuracy": 0.8221660308539868,
      "num_tokens": 19345709.0,
      "step": 2520
    },
    {
      "entropy": 0.9341088324785233,
      "epoch": 0.368,
      "grad_norm": 0.4360695779323578,
      "learning_rate": 0.00015406256241697397,
      "loss": 0.9136,
      "mean_token_accuracy": 0.8160908088088036,
      "num_tokens": 19422973.0,
      "step": 2530
    },
    {
      "entropy": 0.9377870224416256,
      "epoch": 0.3694545454545455,
      "grad_norm": 0.3926893472671509,
      "learning_rate": 0.00015364746923210383,
      "loss": 0.9092,
      "mean_token_accuracy": 0.8166688092052936,
      "num_tokens": 19499742.0,
      "step": 2540
    },
    {
      "entropy": 0.934078136831522,
      "epoch": 0.3709090909090909,
      "grad_norm": 0.40978193283081055,
      "learning_rate": 0.00015323107443155308,
      "loss": 0.9122,
      "mean_token_accuracy": 0.8181058809161186,
      "num_tokens": 19576335.0,
      "step": 2550
    },
    {
      "entropy": 0.9231787696480751,
      "epoch": 0.37236363636363634,
      "grad_norm": 0.711661696434021,
      "learning_rate": 0.00015281338812747836,
      "loss": 0.9058,
      "mean_token_accuracy": 0.8205353997647762,
      "num_tokens": 19651253.0,
      "step": 2560
    },
    {
      "entropy": 0.9429825611412526,
      "epoch": 0.3738181818181818,
      "grad_norm": 0.5801389217376709,
      "learning_rate": 0.00015239442046340043,
      "loss": 0.9243,
      "mean_token_accuracy": 0.8164154589176178,
      "num_tokens": 19727856.0,
      "step": 2570
    },
    {
      "entropy": 0.9396571643650532,
      "epoch": 0.37527272727272726,
      "grad_norm": 0.4592108726501465,
      "learning_rate": 0.0001519741816139579,
      "loss": 0.9048,
      "mean_token_accuracy": 0.8183368548750878,
      "num_tokens": 19803736.0,
      "step": 2580
    },
    {
      "entropy": 0.9264605455100536,
      "epoch": 0.37672727272727274,
      "grad_norm": 0.41448894143104553,
      "learning_rate": 0.00015155268178466028,
      "loss": 0.9131,
      "mean_token_accuracy": 0.8172147408127785,
      "num_tokens": 19880288.0,
      "step": 2590
    },
    {
      "entropy": 0.9423249386250973,
      "epoch": 0.3781818181818182,
      "grad_norm": 0.36910226941108704,
      "learning_rate": 0.00015112993121163974,
      "loss": 0.911,
      "mean_token_accuracy": 0.8178658500313759,
      "num_tokens": 19958616.0,
      "step": 2600
    },
    {
      "entropy": 0.9392407588660717,
      "epoch": 0.37963636363636366,
      "grad_norm": 0.4865938723087311,
      "learning_rate": 0.00015070594016140308,
      "loss": 0.9265,
      "mean_token_accuracy": 0.8160700887441635,
      "num_tokens": 20036860.0,
      "step": 2610
    },
    {
      "entropy": 0.9442516446113587,
      "epoch": 0.3810909090909091,
      "grad_norm": 0.4857174754142761,
      "learning_rate": 0.00015028071893058185,
      "loss": 0.9325,
      "mean_token_accuracy": 0.8172270834445954,
      "num_tokens": 20113912.0,
      "step": 2620
    },
    {
      "entropy": 0.9298271015286446,
      "epoch": 0.3825454545454545,
      "grad_norm": 0.3504846394062042,
      "learning_rate": 0.00014985427784568273,
      "loss": 0.9085,
      "mean_token_accuracy": 0.8181630812585354,
      "num_tokens": 20190831.0,
      "step": 2630
    },
    {
      "entropy": 0.9338633924722671,
      "epoch": 0.384,
      "grad_norm": 0.5028424859046936,
      "learning_rate": 0.0001494266272628366,
      "loss": 0.9165,
      "mean_token_accuracy": 0.8184519417583942,
      "num_tokens": 20267706.0,
      "step": 2640
    },
    {
      "entropy": 0.9261907540261746,
      "epoch": 0.38545454545454544,
      "grad_norm": 0.39501842856407166,
      "learning_rate": 0.00014899777756754691,
      "loss": 0.9128,
      "mean_token_accuracy": 0.8201067417860031,
      "num_tokens": 20342990.0,
      "step": 2650
    },
    {
      "entropy": 0.9520396128296852,
      "epoch": 0.3869090909090909,
      "grad_norm": 0.8775481581687927,
      "learning_rate": 0.00014856773917443762,
      "loss": 0.9312,
      "mean_token_accuracy": 0.8147456295788288,
      "num_tokens": 20419842.0,
      "step": 2660
    },
    {
      "entropy": 0.926639711856842,
      "epoch": 0.38836363636363636,
      "grad_norm": 0.43765246868133545,
      "learning_rate": 0.0001481365225270003,
      "loss": 0.9036,
      "mean_token_accuracy": 0.8191298067569732,
      "num_tokens": 20496226.0,
      "step": 2670
    },
    {
      "entropy": 0.9401526026427746,
      "epoch": 0.38981818181818184,
      "grad_norm": 0.4963286221027374,
      "learning_rate": 0.00014770413809734041,
      "loss": 0.9218,
      "mean_token_accuracy": 0.8168404638767243,
      "num_tokens": 20573037.0,
      "step": 2680
    },
    {
      "entropy": 0.9355879001319408,
      "epoch": 0.39127272727272727,
      "grad_norm": 0.35393384099006653,
      "learning_rate": 0.00014727059638592295,
      "loss": 0.9185,
      "mean_token_accuracy": 0.8172001279890537,
      "num_tokens": 20650509.0,
      "step": 2690
    },
    {
      "entropy": 0.9356629192829132,
      "epoch": 0.3927272727272727,
      "grad_norm": 0.3640415668487549,
      "learning_rate": 0.00014683590792131766,
      "loss": 0.9027,
      "mean_token_accuracy": 0.8211157515645027,
      "num_tokens": 20724944.0,
      "step": 2700
    },
    {
      "entropy": 0.9365813709795475,
      "epoch": 0.3941818181818182,
      "grad_norm": 0.4473666548728943,
      "learning_rate": 0.00014640008325994316,
      "loss": 0.9141,
      "mean_token_accuracy": 0.8174393519759178,
      "num_tokens": 20800950.0,
      "step": 2710
    },
    {
      "entropy": 0.9238623924553394,
      "epoch": 0.3956363636363636,
      "grad_norm": 0.37241610884666443,
      "learning_rate": 0.00014596313298581058,
      "loss": 0.8932,
      "mean_token_accuracy": 0.8217294663190842,
      "num_tokens": 20875390.0,
      "step": 2720
    },
    {
      "entropy": 0.9539158314466476,
      "epoch": 0.3970909090909091,
      "grad_norm": 0.41429755091667175,
      "learning_rate": 0.00014552506771026667,
      "loss": 0.9324,
      "mean_token_accuracy": 0.8137750118970871,
      "num_tokens": 20954271.0,
      "step": 2730
    },
    {
      "entropy": 0.9392914563417435,
      "epoch": 0.39854545454545454,
      "grad_norm": 0.36372101306915283,
      "learning_rate": 0.00014508589807173594,
      "loss": 0.9199,
      "mean_token_accuracy": 0.8174258455634117,
      "num_tokens": 21031738.0,
      "step": 2740
    },
    {
      "entropy": 0.9314709268510342,
      "epoch": 0.4,
      "grad_norm": 0.6674671769142151,
      "learning_rate": 0.00014464563473546245,
      "loss": 0.9073,
      "mean_token_accuracy": 0.8193281888961792,
      "num_tokens": 21106326.0,
      "step": 2750
    },
    {
      "entropy": 0.9411804787814617,
      "epoch": 0.40145454545454545,
      "grad_norm": 0.46535301208496094,
      "learning_rate": 0.00014420428839325057,
      "loss": 0.9291,
      "mean_token_accuracy": 0.8160315252840519,
      "num_tokens": 21186853.0,
      "step": 2760
    },
    {
      "entropy": 0.9423932202160359,
      "epoch": 0.4029090909090909,
      "grad_norm": 0.37130892276763916,
      "learning_rate": 0.00014376186976320575,
      "loss": 0.9168,
      "mean_token_accuracy": 0.8181857027113437,
      "num_tokens": 21265098.0,
      "step": 2770
    },
    {
      "entropy": 0.9356568641960621,
      "epoch": 0.40436363636363637,
      "grad_norm": 0.3903407156467438,
      "learning_rate": 0.00014331838958947374,
      "loss": 0.9191,
      "mean_token_accuracy": 0.8169067747890949,
      "num_tokens": 21342699.0,
      "step": 2780
    },
    {
      "entropy": 0.9259685985744,
      "epoch": 0.4058181818181818,
      "grad_norm": 0.7272932529449463,
      "learning_rate": 0.00014287385864198007,
      "loss": 0.9111,
      "mean_token_accuracy": 0.8178241029381752,
      "num_tokens": 21417249.0,
      "step": 2790
    },
    {
      "entropy": 0.9134186536073685,
      "epoch": 0.4072727272727273,
      "grad_norm": 0.6104690432548523,
      "learning_rate": 0.0001424282877161682,
      "loss": 0.8962,
      "mean_token_accuracy": 0.8213642582297325,
      "num_tokens": 21491979.0,
      "step": 2800
    },
    {
      "entropy": 0.9297884210944176,
      "epoch": 0.4087272727272727,
      "grad_norm": 0.3607460856437683,
      "learning_rate": 0.00014198168763273752,
      "loss": 0.9043,
      "mean_token_accuracy": 0.817758695781231,
      "num_tokens": 21569581.0,
      "step": 2810
    },
    {
      "entropy": 0.921866150945425,
      "epoch": 0.4101818181818182,
      "grad_norm": 0.3911571502685547,
      "learning_rate": 0.0001415340692373806,
      "loss": 0.9039,
      "mean_token_accuracy": 0.8202633537352085,
      "num_tokens": 21645313.0,
      "step": 2820
    },
    {
      "entropy": 0.949428279697895,
      "epoch": 0.41163636363636363,
      "grad_norm": 0.4143660068511963,
      "learning_rate": 0.0001410854434005197,
      "loss": 0.9279,
      "mean_token_accuracy": 0.8163035720586777,
      "num_tokens": 21722390.0,
      "step": 2830
    },
    {
      "entropy": 0.9189708463847637,
      "epoch": 0.41309090909090906,
      "grad_norm": 0.6003644466400146,
      "learning_rate": 0.0001406358210170428,
      "loss": 0.8956,
      "mean_token_accuracy": 0.8213484719395637,
      "num_tokens": 21800558.0,
      "step": 2840
    },
    {
      "entropy": 0.9367615543305874,
      "epoch": 0.41454545454545455,
      "grad_norm": 0.402132511138916,
      "learning_rate": 0.00014018521300603905,
      "loss": 0.9122,
      "mean_token_accuracy": 0.8187360368669033,
      "num_tokens": 21877227.0,
      "step": 2850
    },
    {
      "entropy": 0.9265430241823196,
      "epoch": 0.416,
      "grad_norm": 0.39140671491622925,
      "learning_rate": 0.0001397336303105336,
      "loss": 0.9063,
      "mean_token_accuracy": 0.8227818325161934,
      "num_tokens": 21951033.0,
      "step": 2860
    },
    {
      "entropy": 0.9349323078989983,
      "epoch": 0.41745454545454547,
      "grad_norm": 0.417484313249588,
      "learning_rate": 0.00013928108389722177,
      "loss": 0.9161,
      "mean_token_accuracy": 0.8168958708643913,
      "num_tokens": 22029456.0,
      "step": 2870
    },
    {
      "entropy": 0.9182321138679981,
      "epoch": 0.4189090909090909,
      "grad_norm": 0.9603578448295593,
      "learning_rate": 0.00013882758475620285,
      "loss": 0.8906,
      "mean_token_accuracy": 0.8226393982768059,
      "num_tokens": 22103768.0,
      "step": 2880
    },
    {
      "entropy": 0.9337234787642956,
      "epoch": 0.4203636363636364,
      "grad_norm": 0.9330223798751831,
      "learning_rate": 0.0001383731439007132,
      "loss": 0.9237,
      "mean_token_accuracy": 0.814931558072567,
      "num_tokens": 22182926.0,
      "step": 2890
    },
    {
      "entropy": 0.9424798399209976,
      "epoch": 0.4218181818181818,
      "grad_norm": 0.34551888704299927,
      "learning_rate": 0.00013791777236685855,
      "loss": 0.9136,
      "mean_token_accuracy": 0.8205113738775254,
      "num_tokens": 22259520.0,
      "step": 2900
    },
    {
      "entropy": 0.9198446124792099,
      "epoch": 0.4232727272727273,
      "grad_norm": 0.4358190894126892,
      "learning_rate": 0.00013746148121334642,
      "loss": 0.8901,
      "mean_token_accuracy": 0.8215723693370819,
      "num_tokens": 22334318.0,
      "step": 2910
    },
    {
      "entropy": 0.9174405969679356,
      "epoch": 0.42472727272727273,
      "grad_norm": 0.4261787235736847,
      "learning_rate": 0.00013700428152121701,
      "loss": 0.8959,
      "mean_token_accuracy": 0.8221362054347991,
      "num_tokens": 22410832.0,
      "step": 2920
    },
    {
      "entropy": 0.9318659543991089,
      "epoch": 0.42618181818181816,
      "grad_norm": 0.5009744763374329,
      "learning_rate": 0.0001365461843935747,
      "loss": 0.9019,
      "mean_token_accuracy": 0.8202016733586788,
      "num_tokens": 22486296.0,
      "step": 2930
    },
    {
      "entropy": 0.9342516735196114,
      "epoch": 0.42763636363636365,
      "grad_norm": 0.3965824544429779,
      "learning_rate": 0.00013608720095531794,
      "loss": 0.9133,
      "mean_token_accuracy": 0.8176328152418136,
      "num_tokens": 22563289.0,
      "step": 2940
    },
    {
      "entropy": 0.942034150660038,
      "epoch": 0.4290909090909091,
      "grad_norm": 0.4476518929004669,
      "learning_rate": 0.00013562734235286937,
      "loss": 0.9263,
      "mean_token_accuracy": 0.819250100851059,
      "num_tokens": 22638561.0,
      "step": 2950
    },
    {
      "entropy": 0.9392903119325637,
      "epoch": 0.43054545454545456,
      "grad_norm": 0.5829343795776367,
      "learning_rate": 0.00013516661975390476,
      "loss": 0.9163,
      "mean_token_accuracy": 0.8194300256669521,
      "num_tokens": 22714187.0,
      "step": 2960
    },
    {
      "entropy": 0.9254804484546184,
      "epoch": 0.432,
      "grad_norm": 0.573879063129425,
      "learning_rate": 0.0001347050443470824,
      "loss": 0.9169,
      "mean_token_accuracy": 0.8167784102261066,
      "num_tokens": 22792256.0,
      "step": 2970
    },
    {
      "entropy": 0.9662094607949256,
      "epoch": 0.4334545454545455,
      "grad_norm": 0.47143566608428955,
      "learning_rate": 0.00013424262734177078,
      "loss": 0.9346,
      "mean_token_accuracy": 0.8130164839327335,
      "num_tokens": 22868956.0,
      "step": 2980
    },
    {
      "entropy": 0.9308706648647785,
      "epoch": 0.4349090909090909,
      "grad_norm": 0.36211660504341125,
      "learning_rate": 0.00013377937996777666,
      "loss": 0.9079,
      "mean_token_accuracy": 0.8199804499745369,
      "num_tokens": 22944002.0,
      "step": 2990
    },
    {
      "entropy": 0.9251275800168515,
      "epoch": 0.43636363636363634,
      "grad_norm": 0.3628769814968109,
      "learning_rate": 0.0001333153134750725,
      "loss": 0.9148,
      "mean_token_accuracy": 0.8186253055930137,
      "num_tokens": 23022304.0,
      "step": 3000
    },
    {
      "epoch": 0.43636363636363634,
      "eval_entropy": 0.947712210559845,
      "eval_loss": 0.9226884841918945,
      "eval_mean_token_accuracy": 0.8106017510414123,
      "eval_num_tokens": 23022304.0,
      "eval_runtime": 78.0625,
      "eval_samples_per_second": 64.051,
      "eval_steps_per_second": 8.006,
      "step": 3000
    },
    {
      "entropy": 0.9396435245871544,
      "epoch": 0.43781818181818183,
      "grad_norm": 0.347475528717041,
      "learning_rate": 0.0001328504391335229,
      "loss": 0.9084,
      "mean_token_accuracy": 0.8182823576033116,
      "num_tokens": 23097660.0,
      "step": 3010
    },
    {
      "entropy": 0.916470754146576,
      "epoch": 0.43927272727272726,
      "grad_norm": 0.39701876044273376,
      "learning_rate": 0.0001323847682326111,
      "loss": 0.8897,
      "mean_token_accuracy": 0.822086288779974,
      "num_tokens": 23171356.0,
      "step": 3020
    },
    {
      "entropy": 0.9525364696979522,
      "epoch": 0.44072727272727275,
      "grad_norm": 0.39641302824020386,
      "learning_rate": 0.00013191831208116486,
      "loss": 0.9539,
      "mean_token_accuracy": 0.8152362138032914,
      "num_tokens": 23248906.0,
      "step": 3030
    },
    {
      "entropy": 0.9237325944006443,
      "epoch": 0.4421818181818182,
      "grad_norm": 0.5871150493621826,
      "learning_rate": 0.00013145108200708175,
      "loss": 0.8914,
      "mean_token_accuracy": 0.8203249268233777,
      "num_tokens": 23322555.0,
      "step": 3040
    },
    {
      "entropy": 0.9378106243908405,
      "epoch": 0.44363636363636366,
      "grad_norm": 0.3605862259864807,
      "learning_rate": 0.00013098308935705403,
      "loss": 0.9241,
      "mean_token_accuracy": 0.8165755428373813,
      "num_tokens": 23400467.0,
      "step": 3050
    },
    {
      "entropy": 0.9633555516600609,
      "epoch": 0.4450909090909091,
      "grad_norm": 0.4816967248916626,
      "learning_rate": 0.00013051434549629317,
      "loss": 0.9773,
      "mean_token_accuracy": 0.8138452783226967,
      "num_tokens": 23478934.0,
      "step": 3060
    },
    {
      "entropy": 0.9115117900073528,
      "epoch": 0.4465454545454545,
      "grad_norm": 0.36916565895080566,
      "learning_rate": 0.00013004486180825383,
      "loss": 0.8961,
      "mean_token_accuracy": 0.823471300303936,
      "num_tokens": 23552888.0,
      "step": 3070
    },
    {
      "entropy": 0.9228253945708275,
      "epoch": 0.448,
      "grad_norm": 0.6769089698791504,
      "learning_rate": 0.00012957464969435734,
      "loss": 0.8953,
      "mean_token_accuracy": 0.8215885289013386,
      "num_tokens": 23626705.0,
      "step": 3080
    },
    {
      "entropy": 0.9263963654637337,
      "epoch": 0.44945454545454544,
      "grad_norm": 0.4560895562171936,
      "learning_rate": 0.00012910372057371491,
      "loss": 0.9051,
      "mean_token_accuracy": 0.8180726014077664,
      "num_tokens": 23703595.0,
      "step": 3090
    },
    {
      "entropy": 0.9448229886591435,
      "epoch": 0.4509090909090909,
      "grad_norm": 0.5193542242050171,
      "learning_rate": 0.0001286320858828502,
      "loss": 0.9195,
      "mean_token_accuracy": 0.8177183635532856,
      "num_tokens": 23779080.0,
      "step": 3100
    },
    {
      "entropy": 0.9222862467169761,
      "epoch": 0.45236363636363636,
      "grad_norm": 0.6392302513122559,
      "learning_rate": 0.0001281597570754218,
      "loss": 0.9057,
      "mean_token_accuracy": 0.8185005567967891,
      "num_tokens": 23854814.0,
      "step": 3110
    },
    {
      "entropy": 0.9366964548826218,
      "epoch": 0.45381818181818184,
      "grad_norm": 0.544901430606842,
      "learning_rate": 0.00012768674562194476,
      "loss": 0.9275,
      "mean_token_accuracy": 0.8163784302771091,
      "num_tokens": 23933307.0,
      "step": 3120
    },
    {
      "entropy": 0.9299830220639705,
      "epoch": 0.4552727272727273,
      "grad_norm": 0.4428144693374634,
      "learning_rate": 0.0001272130630095123,
      "loss": 0.9112,
      "mean_token_accuracy": 0.8183266878128052,
      "num_tokens": 24009025.0,
      "step": 3130
    },
    {
      "entropy": 0.9151601083576679,
      "epoch": 0.4567272727272727,
      "grad_norm": 0.38302376866340637,
      "learning_rate": 0.00012673872074151683,
      "loss": 0.8858,
      "mean_token_accuracy": 0.8227717474102973,
      "num_tokens": 24084460.0,
      "step": 3140
    },
    {
      "entropy": 0.9553165949881077,
      "epoch": 0.4581818181818182,
      "grad_norm": 0.34640029072761536,
      "learning_rate": 0.00012626373033737033,
      "loss": 0.9359,
      "mean_token_accuracy": 0.8156644940376282,
      "num_tokens": 24160235.0,
      "step": 3150
    },
    {
      "entropy": 0.9485594473779202,
      "epoch": 0.4596363636363636,
      "grad_norm": 0.730147659778595,
      "learning_rate": 0.00012578810333222488,
      "loss": 0.9349,
      "mean_token_accuracy": 0.8165864571928978,
      "num_tokens": 24238608.0,
      "step": 3160
    },
    {
      "entropy": 0.9255790121853351,
      "epoch": 0.4610909090909091,
      "grad_norm": 0.42998138070106506,
      "learning_rate": 0.00012531185127669244,
      "loss": 0.9021,
      "mean_token_accuracy": 0.8210115492343902,
      "num_tokens": 24315473.0,
      "step": 3170
    },
    {
      "entropy": 0.9337904252111912,
      "epoch": 0.46254545454545454,
      "grad_norm": 0.41894665360450745,
      "learning_rate": 0.00012483498573656436,
      "loss": 0.911,
      "mean_token_accuracy": 0.818168356269598,
      "num_tokens": 24393906.0,
      "step": 3180
    },
    {
      "entropy": 0.9694912165403367,
      "epoch": 0.464,
      "grad_norm": 0.3748018741607666,
      "learning_rate": 0.0001243575182925303,
      "loss": 1.0177,
      "mean_token_accuracy": 0.810963698849082,
      "num_tokens": 24473099.0,
      "step": 3190
    },
    {
      "entropy": 0.9436316221952439,
      "epoch": 0.46545454545454545,
      "grad_norm": 0.792568027973175,
      "learning_rate": 0.0001238794605398974,
      "loss": 0.9198,
      "mean_token_accuracy": 0.8177410371601581,
      "num_tokens": 24549264.0,
      "step": 3200
    },
    {
      "entropy": 0.9587245568633079,
      "epoch": 0.4669090909090909,
      "grad_norm": 0.39849016070365906,
      "learning_rate": 0.00012340082408830835,
      "loss": 0.9658,
      "mean_token_accuracy": 0.814987201988697,
      "num_tokens": 24625693.0,
      "step": 3210
    },
    {
      "entropy": 0.948624425381422,
      "epoch": 0.46836363636363637,
      "grad_norm": 0.3882438540458679,
      "learning_rate": 0.0001229216205614595,
      "loss": 0.9192,
      "mean_token_accuracy": 0.8177899941802025,
      "num_tokens": 24701883.0,
      "step": 3220
    },
    {
      "entropy": 0.9479258358478546,
      "epoch": 0.4698181818181818,
      "grad_norm": 0.5587784647941589,
      "learning_rate": 0.00012244186159681873,
      "loss": 0.9434,
      "mean_token_accuracy": 0.8147072538733482,
      "num_tokens": 24780137.0,
      "step": 3230
    },
    {
      "entropy": 0.9296126678586006,
      "epoch": 0.4712727272727273,
      "grad_norm": 0.35228222608566284,
      "learning_rate": 0.00012196155884534269,
      "loss": 0.9097,
      "mean_token_accuracy": 0.8193468026816845,
      "num_tokens": 24856002.0,
      "step": 3240
    },
    {
      "entropy": 0.9226066932082176,
      "epoch": 0.4727272727272727,
      "grad_norm": 0.6554537415504456,
      "learning_rate": 0.00012148072397119394,
      "loss": 0.891,
      "mean_token_accuracy": 0.8198705278337002,
      "num_tokens": 24931747.0,
      "step": 3250
    },
    {
      "entropy": 0.9482523567974568,
      "epoch": 0.4741818181818182,
      "grad_norm": 0.6361430287361145,
      "learning_rate": 0.00012099936865145763,
      "loss": 0.9299,
      "mean_token_accuracy": 0.8141291528940201,
      "num_tokens": 25008782.0,
      "step": 3260
    },
    {
      "entropy": 0.944165738672018,
      "epoch": 0.47563636363636363,
      "grad_norm": 0.41412845253944397,
      "learning_rate": 0.00012051750457585789,
      "loss": 0.9122,
      "mean_token_accuracy": 0.8184866182506084,
      "num_tokens": 25085940.0,
      "step": 3270
    },
    {
      "entropy": 0.9486552193760872,
      "epoch": 0.47709090909090907,
      "grad_norm": 0.6368948221206665,
      "learning_rate": 0.00012003514344647413,
      "loss": 0.9258,
      "mean_token_accuracy": 0.8154248282313347,
      "num_tokens": 25163594.0,
      "step": 3280
    },
    {
      "entropy": 0.9159012153744698,
      "epoch": 0.47854545454545455,
      "grad_norm": 0.5390453338623047,
      "learning_rate": 0.00011955229697745655,
      "loss": 0.8938,
      "mean_token_accuracy": 0.8222610250115394,
      "num_tokens": 25238709.0,
      "step": 3290
    },
    {
      "entropy": 0.9439261749386787,
      "epoch": 0.48,
      "grad_norm": 0.43710118532180786,
      "learning_rate": 0.00011906897689474199,
      "loss": 0.9254,
      "mean_token_accuracy": 0.818860150873661,
      "num_tokens": 25314417.0,
      "step": 3300
    },
    {
      "entropy": 0.9147123090922833,
      "epoch": 0.48145454545454547,
      "grad_norm": 0.39578351378440857,
      "learning_rate": 0.00011858519493576893,
      "loss": 0.8903,
      "mean_token_accuracy": 0.8215675607323647,
      "num_tokens": 25390270.0,
      "step": 3310
    },
    {
      "entropy": 0.9170261770486832,
      "epoch": 0.4829090909090909,
      "grad_norm": 0.3433591425418854,
      "learning_rate": 0.00011810096284919252,
      "loss": 0.9035,
      "mean_token_accuracy": 0.8198836460709572,
      "num_tokens": 25466771.0,
      "step": 3320
    },
    {
      "entropy": 0.9297826215624809,
      "epoch": 0.4843636363636364,
      "grad_norm": 0.3432069718837738,
      "learning_rate": 0.00011761629239459926,
      "loss": 0.9082,
      "mean_token_accuracy": 0.8195572912693023,
      "num_tokens": 25543759.0,
      "step": 3330
    },
    {
      "entropy": 0.9283205479383468,
      "epoch": 0.4858181818181818,
      "grad_norm": 0.4968811869621277,
      "learning_rate": 0.00011713119534222152,
      "loss": 0.9147,
      "mean_token_accuracy": 0.8183020256459713,
      "num_tokens": 25621348.0,
      "step": 3340
    },
    {
      "entropy": 0.9285181269049645,
      "epoch": 0.48727272727272725,
      "grad_norm": 1.168716311454773,
      "learning_rate": 0.00011664568347265146,
      "loss": 0.906,
      "mean_token_accuracy": 0.8207002833485604,
      "num_tokens": 25697445.0,
      "step": 3350
    },
    {
      "entropy": 0.9199300065636635,
      "epoch": 0.48872727272727273,
      "grad_norm": 0.3925750255584717,
      "learning_rate": 0.00011615976857655522,
      "loss": 0.9024,
      "mean_token_accuracy": 0.8196565262973309,
      "num_tokens": 25774629.0,
      "step": 3360
    },
    {
      "entropy": 0.917039792984724,
      "epoch": 0.49018181818181816,
      "grad_norm": 0.39110034704208374,
      "learning_rate": 0.00011567346245438632,
      "loss": 0.8925,
      "mean_token_accuracy": 0.8205363370478154,
      "num_tokens": 25851159.0,
      "step": 3370
    },
    {
      "entropy": 0.9277261577546596,
      "epoch": 0.49163636363636365,
      "grad_norm": 0.3983883261680603,
      "learning_rate": 0.0001151867769160994,
      "loss": 0.89,
      "mean_token_accuracy": 0.8212614722549916,
      "num_tokens": 25927101.0,
      "step": 3380
    },
    {
      "entropy": 0.925445581227541,
      "epoch": 0.4930909090909091,
      "grad_norm": 0.4415944516658783,
      "learning_rate": 0.00011469972378086302,
      "loss": 0.9093,
      "mean_token_accuracy": 0.8164033927023411,
      "num_tokens": 26004710.0,
      "step": 3390
    },
    {
      "entropy": 0.9090649954974651,
      "epoch": 0.49454545454545457,
      "grad_norm": 0.36825454235076904,
      "learning_rate": 0.00011421231487677293,
      "loss": 0.8923,
      "mean_token_accuracy": 0.8208632886409759,
      "num_tokens": 26081021.0,
      "step": 3400
    },
    {
      "entropy": 0.9246860764920711,
      "epoch": 0.496,
      "grad_norm": 0.34179189801216125,
      "learning_rate": 0.0001137245620405648,
      "loss": 0.895,
      "mean_token_accuracy": 0.8194523319602013,
      "num_tokens": 26156724.0,
      "step": 3410
    },
    {
      "entropy": 0.9491905942559242,
      "epoch": 0.4974545454545454,
      "grad_norm": 0.37097907066345215,
      "learning_rate": 0.00011323647711732652,
      "loss": 0.9263,
      "mean_token_accuracy": 0.8174832746386528,
      "num_tokens": 26235045.0,
      "step": 3420
    },
    {
      "entropy": 0.9204013898968697,
      "epoch": 0.4989090909090909,
      "grad_norm": 1.1383200883865356,
      "learning_rate": 0.00011274807196021095,
      "loss": 0.9073,
      "mean_token_accuracy": 0.8202569909393788,
      "num_tokens": 26310893.0,
      "step": 3430
    },
    {
      "entropy": 0.9221264578402042,
      "epoch": 0.5003636363636363,
      "grad_norm": 0.9400169253349304,
      "learning_rate": 0.0001122593584301477,
      "loss": 0.8983,
      "mean_token_accuracy": 0.8210197605192662,
      "num_tokens": 26385660.0,
      "step": 3440
    },
    {
      "entropy": 0.9211826339364052,
      "epoch": 0.5018181818181818,
      "grad_norm": 0.47897812724113464,
      "learning_rate": 0.00011177034839555519,
      "loss": 0.903,
      "mean_token_accuracy": 0.8199583508074284,
      "num_tokens": 26462236.0,
      "step": 3450
    },
    {
      "entropy": 0.9359012946486474,
      "epoch": 0.5032727272727273,
      "grad_norm": 0.35076892375946045,
      "learning_rate": 0.00011128105373205256,
      "loss": 0.9084,
      "mean_token_accuracy": 0.8181154794991017,
      "num_tokens": 26539077.0,
      "step": 3460
    },
    {
      "entropy": 0.9265558190643788,
      "epoch": 0.5047272727272727,
      "grad_norm": 0.3212345242500305,
      "learning_rate": 0.0001107914863221711,
      "loss": 0.9012,
      "mean_token_accuracy": 0.8197455100715161,
      "num_tokens": 26616327.0,
      "step": 3470
    },
    {
      "entropy": 0.9231369532644749,
      "epoch": 0.5061818181818182,
      "grad_norm": 0.7979731559753418,
      "learning_rate": 0.0001103016580550658,
      "loss": 0.902,
      "mean_token_accuracy": 0.8190245240926742,
      "num_tokens": 26692145.0,
      "step": 3480
    },
    {
      "entropy": 0.9156353622674942,
      "epoch": 0.5076363636363637,
      "grad_norm": 0.3245525062084198,
      "learning_rate": 0.00010981158082622653,
      "loss": 0.894,
      "mean_token_accuracy": 0.8221806295216083,
      "num_tokens": 26767988.0,
      "step": 3490
    },
    {
      "entropy": 0.9295620888471603,
      "epoch": 0.509090909090909,
      "grad_norm": 0.36039286851882935,
      "learning_rate": 0.00010932126653718919,
      "loss": 0.9055,
      "mean_token_accuracy": 0.8197094917297363,
      "num_tokens": 26843488.0,
      "step": 3500
    },
    {
      "entropy": 0.9358175002038479,
      "epoch": 0.5105454545454545,
      "grad_norm": 0.4395211935043335,
      "learning_rate": 0.0001088307270952468,
      "loss": 0.93,
      "mean_token_accuracy": 0.8159300364553929,
      "num_tokens": 26921791.0,
      "step": 3510
    },
    {
      "entropy": 0.9184973612427711,
      "epoch": 0.512,
      "grad_norm": 0.4433652460575104,
      "learning_rate": 0.00010833997441316004,
      "loss": 0.8887,
      "mean_token_accuracy": 0.8236219830811023,
      "num_tokens": 26996743.0,
      "step": 3520
    },
    {
      "entropy": 0.9166963383555412,
      "epoch": 0.5134545454545455,
      "grad_norm": 0.401826947927475,
      "learning_rate": 0.00010784902040886832,
      "loss": 0.8952,
      "mean_token_accuracy": 0.8214032381772995,
      "num_tokens": 27073260.0,
      "step": 3530
    },
    {
      "entropy": 0.9150002479553223,
      "epoch": 0.5149090909090909,
      "grad_norm": 0.4118078947067261,
      "learning_rate": 0.00010735787700520004,
      "loss": 0.8916,
      "mean_token_accuracy": 0.8216412000358104,
      "num_tokens": 27148611.0,
      "step": 3540
    },
    {
      "entropy": 0.9454425044357777,
      "epoch": 0.5163636363636364,
      "grad_norm": 0.7244296073913574,
      "learning_rate": 0.00010686655612958325,
      "loss": 0.9173,
      "mean_token_accuracy": 0.816456987708807,
      "num_tokens": 27230353.0,
      "step": 3550
    },
    {
      "entropy": 0.9069771558046341,
      "epoch": 0.5178181818181818,
      "grad_norm": 0.32180866599082947,
      "learning_rate": 0.00010637506971375577,
      "loss": 0.898,
      "mean_token_accuracy": 0.8206721089780331,
      "num_tokens": 27308270.0,
      "step": 3560
    },
    {
      "entropy": 0.9277030676603317,
      "epoch": 0.5192727272727272,
      "grad_norm": 0.577333390712738,
      "learning_rate": 0.00010588342969347578,
      "loss": 0.8913,
      "mean_token_accuracy": 0.8211231052875518,
      "num_tokens": 27385138.0,
      "step": 3570
    },
    {
      "entropy": 0.9140991292893886,
      "epoch": 0.5207272727272727,
      "grad_norm": 0.4405682682991028,
      "learning_rate": 0.00010539164800823155,
      "loss": 0.9013,
      "mean_token_accuracy": 0.8207581497728824,
      "num_tokens": 27460730.0,
      "step": 3580
    },
    {
      "entropy": 0.9315586917102336,
      "epoch": 0.5221818181818182,
      "grad_norm": 0.8212738633155823,
      "learning_rate": 0.00010489973660095182,
      "loss": 0.9001,
      "mean_token_accuracy": 0.8187609612941742,
      "num_tokens": 27538332.0,
      "step": 3590
    },
    {
      "entropy": 0.9198866352438927,
      "epoch": 0.5236363636363637,
      "grad_norm": 1.2040369510650635,
      "learning_rate": 0.00010440770741771553,
      "loss": 0.8999,
      "mean_token_accuracy": 0.8207184448838234,
      "num_tokens": 27614736.0,
      "step": 3600
    },
    {
      "entropy": 0.9046032093465328,
      "epoch": 0.525090909090909,
      "grad_norm": 0.6865532398223877,
      "learning_rate": 0.00010391557240746192,
      "loss": 0.8698,
      "mean_token_accuracy": 0.8253046363592148,
      "num_tokens": 27688293.0,
      "step": 3610
    },
    {
      "entropy": 0.9542017750442028,
      "epoch": 0.5265454545454545,
      "grad_norm": 0.4730345904827118,
      "learning_rate": 0.00010342334352170016,
      "loss": 0.939,
      "mean_token_accuracy": 0.8143124967813492,
      "num_tokens": 27768383.0,
      "step": 3620
    },
    {
      "entropy": 0.9306189320981503,
      "epoch": 0.528,
      "grad_norm": 0.376300573348999,
      "learning_rate": 0.00010293103271421916,
      "loss": 0.9123,
      "mean_token_accuracy": 0.8206489957869053,
      "num_tokens": 27842512.0,
      "step": 3630
    },
    {
      "entropy": 0.9370056606829167,
      "epoch": 0.5294545454545454,
      "grad_norm": 0.4288793206214905,
      "learning_rate": 0.00010243865194079745,
      "loss": 0.9279,
      "mean_token_accuracy": 0.8167462557554245,
      "num_tokens": 27921618.0,
      "step": 3640
    },
    {
      "entropy": 0.921815349906683,
      "epoch": 0.5309090909090909,
      "grad_norm": 0.5747948884963989,
      "learning_rate": 0.00010194621315891253,
      "loss": 0.8987,
      "mean_token_accuracy": 0.8218463979661464,
      "num_tokens": 27998037.0,
      "step": 3650
    },
    {
      "entropy": 0.9359241731464862,
      "epoch": 0.5323636363636364,
      "grad_norm": 0.4183412492275238,
      "learning_rate": 0.0001014537283274507,
      "loss": 0.9061,
      "mean_token_accuracy": 0.8193797774612903,
      "num_tokens": 28075384.0,
      "step": 3660
    },
    {
      "entropy": 0.9195265114307404,
      "epoch": 0.5338181818181819,
      "grad_norm": 0.7315105199813843,
      "learning_rate": 0.00010096120940641656,
      "loss": 0.8984,
      "mean_token_accuracy": 0.8199678152799607,
      "num_tokens": 28151833.0,
      "step": 3670
    },
    {
      "entropy": 0.9044848591089248,
      "epoch": 0.5352727272727272,
      "grad_norm": 0.8692927360534668,
      "learning_rate": 0.00010046866835664261,
      "loss": 0.886,
      "mean_token_accuracy": 0.8231448031961918,
      "num_tokens": 28227799.0,
      "step": 3680
    },
    {
      "entropy": 0.9234697446227074,
      "epoch": 0.5367272727272727,
      "grad_norm": 0.6376194357872009,
      "learning_rate": 9.997611713949872e-05,
      "loss": 0.9013,
      "mean_token_accuracy": 0.819769125431776,
      "num_tokens": 28304947.0,
      "step": 3690
    },
    {
      "entropy": 0.9290397010743618,
      "epoch": 0.5381818181818182,
      "grad_norm": 0.5948243737220764,
      "learning_rate": 9.948356771660168e-05,
      "loss": 0.9035,
      "mean_token_accuracy": 0.8185986518859864,
      "num_tokens": 28382467.0,
      "step": 3700
    },
    {
      "entropy": 0.9379419051110744,
      "epoch": 0.5396363636363637,
      "grad_norm": 2.2813308238983154,
      "learning_rate": 9.899103204952468e-05,
      "loss": 0.9169,
      "mean_token_accuracy": 0.815712084621191,
      "num_tokens": 28461716.0,
      "step": 3710
    },
    {
      "entropy": 0.9275504149496555,
      "epoch": 0.5410909090909091,
      "grad_norm": 0.34508106112480164,
      "learning_rate": 9.849852209950689e-05,
      "loss": 0.9048,
      "mean_token_accuracy": 0.8185636311769485,
      "num_tokens": 28538563.0,
      "step": 3720
    },
    {
      "entropy": 0.9352700494229793,
      "epoch": 0.5425454545454546,
      "grad_norm": 0.37615418434143066,
      "learning_rate": 9.80060498271629e-05,
      "loss": 0.9155,
      "mean_token_accuracy": 0.8164530210196972,
      "num_tokens": 28619086.0,
      "step": 3730
    },
    {
      "entropy": 0.9283501178026199,
      "epoch": 0.544,
      "grad_norm": 0.7256191968917847,
      "learning_rate": 9.751362719219235e-05,
      "loss": 0.9183,
      "mean_token_accuracy": 0.8186564229428768,
      "num_tokens": 28695671.0,
      "step": 3740
    },
    {
      "entropy": 0.9386952750384807,
      "epoch": 0.5454545454545454,
      "grad_norm": 0.6169796586036682,
      "learning_rate": 9.702126615308941e-05,
      "loss": 0.9019,
      "mean_token_accuracy": 0.8199047222733498,
      "num_tokens": 28771736.0,
      "step": 3750
    },
    {
      "entropy": 0.9412241369485855,
      "epoch": 0.5469090909090909,
      "grad_norm": 0.7588170766830444,
      "learning_rate": 9.65289786668524e-05,
      "loss": 0.9407,
      "mean_token_accuracy": 0.8170472726225853,
      "num_tokens": 28849338.0,
      "step": 3760
    },
    {
      "entropy": 0.9540625616908074,
      "epoch": 0.5483636363636364,
      "grad_norm": 0.38552454113960266,
      "learning_rate": 9.603677668869331e-05,
      "loss": 0.9463,
      "mean_token_accuracy": 0.8130639024078846,
      "num_tokens": 28930968.0,
      "step": 3770
    },
    {
      "entropy": 0.9177017278969288,
      "epoch": 0.5498181818181819,
      "grad_norm": 0.4654134511947632,
      "learning_rate": 9.554467217174777e-05,
      "loss": 0.8947,
      "mean_token_accuracy": 0.8225653968751431,
      "num_tokens": 29006513.0,
      "step": 3780
    },
    {
      "entropy": 0.9458075389266014,
      "epoch": 0.5512727272727272,
      "grad_norm": 0.8493314981460571,
      "learning_rate": 9.505267706678436e-05,
      "loss": 0.9878,
      "mean_token_accuracy": 0.8140935592353344,
      "num_tokens": 29083506.0,
      "step": 3790
    },
    {
      "entropy": 0.9165011703968048,
      "epoch": 0.5527272727272727,
      "grad_norm": 0.4751242399215698,
      "learning_rate": 9.45608033219147e-05,
      "loss": 0.8935,
      "mean_token_accuracy": 0.8196017131209373,
      "num_tokens": 29160892.0,
      "step": 3800
    },
    {
      "entropy": 0.9129310235381126,
      "epoch": 0.5541818181818182,
      "grad_norm": 0.3513168394565582,
      "learning_rate": 9.406906288230316e-05,
      "loss": 0.899,
      "mean_token_accuracy": 0.821053197979927,
      "num_tokens": 29235237.0,
      "step": 3810
    },
    {
      "entropy": 0.9064737364649773,
      "epoch": 0.5556363636363636,
      "grad_norm": 0.38477998971939087,
      "learning_rate": 9.357746768987676e-05,
      "loss": 0.8776,
      "mean_token_accuracy": 0.824255996197462,
      "num_tokens": 29309459.0,
      "step": 3820
    },
    {
      "entropy": 0.9401685081422329,
      "epoch": 0.5570909090909091,
      "grad_norm": 0.3995867967605591,
      "learning_rate": 9.308602968303524e-05,
      "loss": 0.9129,
      "mean_token_accuracy": 0.8195973120629787,
      "num_tokens": 29386095.0,
      "step": 3830
    },
    {
      "entropy": 0.9213009759783745,
      "epoch": 0.5585454545454546,
      "grad_norm": 0.643458902835846,
      "learning_rate": 9.2594760796361e-05,
      "loss": 0.8994,
      "mean_token_accuracy": 0.8200056202709675,
      "num_tokens": 29462603.0,
      "step": 3840
    },
    {
      "entropy": 0.918635991960764,
      "epoch": 0.56,
      "grad_norm": 0.8464357852935791,
      "learning_rate": 9.210367296032943e-05,
      "loss": 0.9001,
      "mean_token_accuracy": 0.8222006931900978,
      "num_tokens": 29537771.0,
      "step": 3850
    },
    {
      "entropy": 0.9163896553218365,
      "epoch": 0.5614545454545454,
      "grad_norm": 0.37169432640075684,
      "learning_rate": 9.161277810101906e-05,
      "loss": 0.895,
      "mean_token_accuracy": 0.8202171288430691,
      "num_tokens": 29615868.0,
      "step": 3860
    },
    {
      "entropy": 0.9250077605247498,
      "epoch": 0.5629090909090909,
      "grad_norm": 1.156219720840454,
      "learning_rate": 9.112208813982203e-05,
      "loss": 0.906,
      "mean_token_accuracy": 0.8188555419445038,
      "num_tokens": 29693840.0,
      "step": 3870
    },
    {
      "entropy": 0.9499965131282806,
      "epoch": 0.5643636363636364,
      "grad_norm": 1.324691891670227,
      "learning_rate": 9.063161499315442e-05,
      "loss": 0.9182,
      "mean_token_accuracy": 0.8170868992805481,
      "num_tokens": 29772620.0,
      "step": 3880
    },
    {
      "entropy": 0.9008949503302575,
      "epoch": 0.5658181818181818,
      "grad_norm": 0.43073347210884094,
      "learning_rate": 9.014137057216711e-05,
      "loss": 0.8852,
      "mean_token_accuracy": 0.8227784849703312,
      "num_tokens": 29848445.0,
      "step": 3890
    },
    {
      "entropy": 0.8970596104860306,
      "epoch": 0.5672727272727273,
      "grad_norm": 0.45119038224220276,
      "learning_rate": 8.965136678245631e-05,
      "loss": 0.8933,
      "mean_token_accuracy": 0.8247206792235374,
      "num_tokens": 29922761.0,
      "step": 3900
    },
    {
      "entropy": 0.9342648290097714,
      "epoch": 0.5687272727272727,
      "grad_norm": 0.9876037240028381,
      "learning_rate": 8.916161552377436e-05,
      "loss": 0.9123,
      "mean_token_accuracy": 0.8185786202549934,
      "num_tokens": 30000359.0,
      "step": 3910
    },
    {
      "entropy": 0.9663207672536374,
      "epoch": 0.5701818181818182,
      "grad_norm": 0.6844055652618408,
      "learning_rate": 8.867212868974107e-05,
      "loss": 0.9551,
      "mean_token_accuracy": 0.8155784040689469,
      "num_tokens": 30077474.0,
      "step": 3920
    },
    {
      "entropy": 0.9191053174436092,
      "epoch": 0.5716363636363636,
      "grad_norm": 0.5892969369888306,
      "learning_rate": 8.81829181675546e-05,
      "loss": 0.8949,
      "mean_token_accuracy": 0.8220530681312084,
      "num_tokens": 30151340.0,
      "step": 3930
    },
    {
      "entropy": 0.9044974833726883,
      "epoch": 0.5730909090909091,
      "grad_norm": 0.778037428855896,
      "learning_rate": 8.769399583770292e-05,
      "loss": 0.8892,
      "mean_token_accuracy": 0.8229080185294151,
      "num_tokens": 30228713.0,
      "step": 3940
    },
    {
      "entropy": 0.9285637088119983,
      "epoch": 0.5745454545454546,
      "grad_norm": 0.3809204697608948,
      "learning_rate": 8.72053735736752e-05,
      "loss": 0.9108,
      "mean_token_accuracy": 0.8179547920823097,
      "num_tokens": 30307555.0,
      "step": 3950
    },
    {
      "entropy": 0.9175647012889385,
      "epoch": 0.576,
      "grad_norm": 1.3078035116195679,
      "learning_rate": 8.671706324167353e-05,
      "loss": 0.8917,
      "mean_token_accuracy": 0.8220877289772034,
      "num_tokens": 30383413.0,
      "step": 3960
    },
    {
      "entropy": 0.9246532939374447,
      "epoch": 0.5774545454545454,
      "grad_norm": 0.5710193514823914,
      "learning_rate": 8.622907670032464e-05,
      "loss": 0.9018,
      "mean_token_accuracy": 0.8195373341441154,
      "num_tokens": 30461770.0,
      "step": 3970
    },
    {
      "entropy": 0.9507615320384503,
      "epoch": 0.5789090909090909,
      "grad_norm": 0.36621034145355225,
      "learning_rate": 8.574142580039214e-05,
      "loss": 0.9635,
      "mean_token_accuracy": 0.8138333059847355,
      "num_tokens": 30541665.0,
      "step": 3980
    },
    {
      "entropy": 0.9386632166802883,
      "epoch": 0.5803636363636364,
      "grad_norm": 0.4421805143356323,
      "learning_rate": 8.525412238448848e-05,
      "loss": 0.9079,
      "mean_token_accuracy": 0.8193704970180988,
      "num_tokens": 30619217.0,
      "step": 3990
    },
    {
      "entropy": 0.9030759565532207,
      "epoch": 0.5818181818181818,
      "grad_norm": 0.4840448498725891,
      "learning_rate": 8.47671782867875e-05,
      "loss": 0.8797,
      "mean_token_accuracy": 0.8235183075070381,
      "num_tokens": 30693435.0,
      "step": 4000
    },
    {
      "epoch": 0.5818181818181818,
      "eval_entropy": 0.9226345249176026,
      "eval_loss": 0.9059992432594299,
      "eval_mean_token_accuracy": 0.8117552498817444,
      "eval_num_tokens": 30693435.0,
      "eval_runtime": 78.4875,
      "eval_samples_per_second": 63.704,
      "eval_steps_per_second": 7.963,
      "step": 4000
    },
    {
      "entropy": 0.9224749609827996,
      "epoch": 0.5832727272727273,
      "grad_norm": 0.3205111622810364,
      "learning_rate": 8.428060533273695e-05,
      "loss": 0.8999,
      "mean_token_accuracy": 0.8211710818111897,
      "num_tokens": 30768893.0,
      "step": 4010
    },
    {
      "entropy": 0.9314353063702583,
      "epoch": 0.5847272727272728,
      "grad_norm": 0.35186830163002014,
      "learning_rate": 8.37944153387714e-05,
      "loss": 0.9154,
      "mean_token_accuracy": 0.8171250179409981,
      "num_tokens": 30845555.0,
      "step": 4020
    },
    {
      "entropy": 0.9353597678244114,
      "epoch": 0.5861818181818181,
      "grad_norm": 0.6230372786521912,
      "learning_rate": 8.330862011202519e-05,
      "loss": 0.9066,
      "mean_token_accuracy": 0.8200264059007167,
      "num_tokens": 30921718.0,
      "step": 4030
    },
    {
      "entropy": 0.9002460077404976,
      "epoch": 0.5876363636363636,
      "grad_norm": 0.3467552959918976,
      "learning_rate": 8.282323145004571e-05,
      "loss": 0.8692,
      "mean_token_accuracy": 0.8239211171865464,
      "num_tokens": 30997508.0,
      "step": 4040
    },
    {
      "entropy": 0.9211130246520043,
      "epoch": 0.5890909090909091,
      "grad_norm": 2.3690154552459717,
      "learning_rate": 8.233826114050696e-05,
      "loss": 0.9048,
      "mean_token_accuracy": 0.8180987447500229,
      "num_tokens": 31075376.0,
      "step": 4050
    },
    {
      "entropy": 0.9173007644712925,
      "epoch": 0.5905454545454546,
      "grad_norm": 0.3382624685764313,
      "learning_rate": 8.185372096092322e-05,
      "loss": 0.885,
      "mean_token_accuracy": 0.8216130308806896,
      "num_tokens": 31151185.0,
      "step": 4060
    },
    {
      "entropy": 0.9371884010732174,
      "epoch": 0.592,
      "grad_norm": 0.9030754566192627,
      "learning_rate": 8.13696226783631e-05,
      "loss": 0.9301,
      "mean_token_accuracy": 0.8147789672017097,
      "num_tokens": 31229056.0,
      "step": 4070
    },
    {
      "entropy": 0.9133959837257862,
      "epoch": 0.5934545454545455,
      "grad_norm": 0.35741662979125977,
      "learning_rate": 8.08859780491636e-05,
      "loss": 0.8917,
      "mean_token_accuracy": 0.8221455074846744,
      "num_tokens": 31304335.0,
      "step": 4080
    },
    {
      "entropy": 0.9080164395272732,
      "epoch": 0.5949090909090909,
      "grad_norm": 0.49753931164741516,
      "learning_rate": 8.040279881864492e-05,
      "loss": 0.887,
      "mean_token_accuracy": 0.8221993774175644,
      "num_tokens": 31379536.0,
      "step": 4090
    },
    {
      "entropy": 0.9248534195125103,
      "epoch": 0.5963636363636363,
      "grad_norm": 0.35766759514808655,
      "learning_rate": 7.992009672082495e-05,
      "loss": 0.9019,
      "mean_token_accuracy": 0.8192827641963959,
      "num_tokens": 31456220.0,
      "step": 4100
    },
    {
      "entropy": 0.9308187551796436,
      "epoch": 0.5978181818181818,
      "grad_norm": 0.6227002739906311,
      "learning_rate": 7.943788347813432e-05,
      "loss": 0.9086,
      "mean_token_accuracy": 0.8193796806037426,
      "num_tokens": 31533446.0,
      "step": 4110
    },
    {
      "entropy": 0.9144325762987137,
      "epoch": 0.5992727272727273,
      "grad_norm": 1.2459226846694946,
      "learning_rate": 7.895617080113195e-05,
      "loss": 0.9024,
      "mean_token_accuracy": 0.8201359301805496,
      "num_tokens": 31611084.0,
      "step": 4120
    },
    {
      "entropy": 0.9166338913142681,
      "epoch": 0.6007272727272728,
      "grad_norm": 0.3561236560344696,
      "learning_rate": 7.847497038822044e-05,
      "loss": 0.8981,
      "mean_token_accuracy": 0.8198485419154167,
      "num_tokens": 31688912.0,
      "step": 4130
    },
    {
      "entropy": 0.9203705497086048,
      "epoch": 0.6021818181818182,
      "grad_norm": 0.6236023306846619,
      "learning_rate": 7.799429392536209e-05,
      "loss": 0.8902,
      "mean_token_accuracy": 0.8216739609837532,
      "num_tokens": 31764534.0,
      "step": 4140
    },
    {
      "entropy": 0.9371412299573422,
      "epoch": 0.6036363636363636,
      "grad_norm": 0.8278318047523499,
      "learning_rate": 7.751415308579494e-05,
      "loss": 0.923,
      "mean_token_accuracy": 0.8161679744720459,
      "num_tokens": 31841612.0,
      "step": 4150
    },
    {
      "entropy": 0.9218975573778152,
      "epoch": 0.6050909090909091,
      "grad_norm": 0.7787909507751465,
      "learning_rate": 7.70345595297496e-05,
      "loss": 0.9358,
      "mean_token_accuracy": 0.8203811950981617,
      "num_tokens": 31919103.0,
      "step": 4160
    },
    {
      "entropy": 0.9152127832174302,
      "epoch": 0.6065454545454545,
      "grad_norm": 0.3186095058917999,
      "learning_rate": 7.655552490416571e-05,
      "loss": 0.8983,
      "mean_token_accuracy": 0.8220426417887211,
      "num_tokens": 31997642.0,
      "step": 4170
    },
    {
      "entropy": 0.9264437086880207,
      "epoch": 0.608,
      "grad_norm": 1.0923210382461548,
      "learning_rate": 7.607706084240931e-05,
      "loss": 0.8971,
      "mean_token_accuracy": 0.8194640293717385,
      "num_tokens": 32074441.0,
      "step": 4180
    },
    {
      "entropy": 0.9173982113599777,
      "epoch": 0.6094545454545455,
      "grad_norm": 0.45009058713912964,
      "learning_rate": 7.559917896399043e-05,
      "loss": 0.8987,
      "mean_token_accuracy": 0.8222583763301372,
      "num_tokens": 32150958.0,
      "step": 4190
    },
    {
      "entropy": 0.9083267413079739,
      "epoch": 0.610909090909091,
      "grad_norm": 0.6000900864601135,
      "learning_rate": 7.512189087428057e-05,
      "loss": 0.8963,
      "mean_token_accuracy": 0.8209093257784843,
      "num_tokens": 32226361.0,
      "step": 4200
    },
    {
      "entropy": 0.9476249985396862,
      "epoch": 0.6123636363636363,
      "grad_norm": 0.34333670139312744,
      "learning_rate": 7.464520816423117e-05,
      "loss": 0.9171,
      "mean_token_accuracy": 0.8171191327273846,
      "num_tokens": 32305274.0,
      "step": 4210
    },
    {
      "entropy": 0.9213205456733704,
      "epoch": 0.6138181818181818,
      "grad_norm": 0.323318213224411,
      "learning_rate": 7.416914241009198e-05,
      "loss": 0.9058,
      "mean_token_accuracy": 0.820676113665104,
      "num_tokens": 32379985.0,
      "step": 4220
    },
    {
      "entropy": 0.9409652277827263,
      "epoch": 0.6152727272727273,
      "grad_norm": 0.5688571929931641,
      "learning_rate": 7.369370517312999e-05,
      "loss": 0.9373,
      "mean_token_accuracy": 0.8150431983172893,
      "num_tokens": 32461118.0,
      "step": 4230
    },
    {
      "entropy": 0.8937712088227272,
      "epoch": 0.6167272727272727,
      "grad_norm": 0.3120442032814026,
      "learning_rate": 7.321890799934859e-05,
      "loss": 0.8769,
      "mean_token_accuracy": 0.8248605616390705,
      "num_tokens": 32536163.0,
      "step": 4240
    },
    {
      "entropy": 0.9199181906878948,
      "epoch": 0.6181818181818182,
      "grad_norm": 0.3446398973464966,
      "learning_rate": 7.274476241920722e-05,
      "loss": 0.8886,
      "mean_token_accuracy": 0.8221151031553745,
      "num_tokens": 32610946.0,
      "step": 4250
    },
    {
      "entropy": 0.9408010371029377,
      "epoch": 0.6196363636363637,
      "grad_norm": 0.4573688507080078,
      "learning_rate": 7.227127994734143e-05,
      "loss": 0.927,
      "mean_token_accuracy": 0.815653482824564,
      "num_tokens": 32688544.0,
      "step": 4260
    },
    {
      "entropy": 0.9303324475884438,
      "epoch": 0.6210909090909091,
      "grad_norm": 0.351698637008667,
      "learning_rate": 7.17984720822831e-05,
      "loss": 0.9094,
      "mean_token_accuracy": 0.8197804532945157,
      "num_tokens": 32764950.0,
      "step": 4270
    },
    {
      "entropy": 0.9053874962031842,
      "epoch": 0.6225454545454545,
      "grad_norm": 0.35766610503196716,
      "learning_rate": 7.13263503061813e-05,
      "loss": 0.8853,
      "mean_token_accuracy": 0.8218501470983028,
      "num_tokens": 32842151.0,
      "step": 4280
    },
    {
      "entropy": 0.9122004747390747,
      "epoch": 0.624,
      "grad_norm": 0.3826545774936676,
      "learning_rate": 7.085492608452338e-05,
      "loss": 0.8946,
      "mean_token_accuracy": 0.8221177138388157,
      "num_tokens": 32916904.0,
      "step": 4290
    },
    {
      "entropy": 0.9165596097707749,
      "epoch": 0.6254545454545455,
      "grad_norm": 1.2036513090133667,
      "learning_rate": 7.038421086585666e-05,
      "loss": 0.8885,
      "mean_token_accuracy": 0.8209186360239983,
      "num_tokens": 32993609.0,
      "step": 4300
    },
    {
      "entropy": 0.9260142505168915,
      "epoch": 0.6269090909090909,
      "grad_norm": 1.3860843181610107,
      "learning_rate": 6.991421608151017e-05,
      "loss": 0.8996,
      "mean_token_accuracy": 0.8197734817862511,
      "num_tokens": 33071167.0,
      "step": 4310
    },
    {
      "entropy": 0.9168666034936905,
      "epoch": 0.6283636363636363,
      "grad_norm": 2.340118885040283,
      "learning_rate": 6.944495314531727e-05,
      "loss": 0.8975,
      "mean_token_accuracy": 0.8201201424002648,
      "num_tokens": 33147789.0,
      "step": 4320
    },
    {
      "entropy": 0.9065008342266083,
      "epoch": 0.6298181818181818,
      "grad_norm": 0.6106637716293335,
      "learning_rate": 6.897643345333834e-05,
      "loss": 0.8967,
      "mean_token_accuracy": 0.8197059437632561,
      "num_tokens": 33225832.0,
      "step": 4330
    },
    {
      "entropy": 0.9281016632914543,
      "epoch": 0.6312727272727273,
      "grad_norm": 0.9546725153923035,
      "learning_rate": 6.850866838358409e-05,
      "loss": 0.906,
      "mean_token_accuracy": 0.8203367196023464,
      "num_tokens": 33301739.0,
      "step": 4340
    },
    {
      "entropy": 0.9266930267214775,
      "epoch": 0.6327272727272727,
      "grad_norm": 0.3478836417198181,
      "learning_rate": 6.804166929573917e-05,
      "loss": 0.8965,
      "mean_token_accuracy": 0.8206172339618206,
      "num_tokens": 33377878.0,
      "step": 4350
    },
    {
      "entropy": 0.9037258170545102,
      "epoch": 0.6341818181818182,
      "grad_norm": 0.3244507312774658,
      "learning_rate": 6.75754475308864e-05,
      "loss": 0.885,
      "mean_token_accuracy": 0.8226259790360928,
      "num_tokens": 33454196.0,
      "step": 4360
    },
    {
      "entropy": 0.9108875006437301,
      "epoch": 0.6356363636363637,
      "grad_norm": 0.3166865408420563,
      "learning_rate": 6.711001441123121e-05,
      "loss": 0.8991,
      "mean_token_accuracy": 0.8241880655288696,
      "num_tokens": 33529129.0,
      "step": 4370
    },
    {
      "entropy": 0.9191383279860019,
      "epoch": 0.637090909090909,
      "grad_norm": 0.351441890001297,
      "learning_rate": 6.664538123982683e-05,
      "loss": 0.8989,
      "mean_token_accuracy": 0.8207916006445884,
      "num_tokens": 33607830.0,
      "step": 4380
    },
    {
      "entropy": 0.9276580266654492,
      "epoch": 0.6385454545454545,
      "grad_norm": 0.34574443101882935,
      "learning_rate": 6.618155930029973e-05,
      "loss": 0.9048,
      "mean_token_accuracy": 0.8198268927633763,
      "num_tokens": 33685645.0,
      "step": 4390
    },
    {
      "entropy": 0.9335759162902832,
      "epoch": 0.64,
      "grad_norm": 0.5790361762046814,
      "learning_rate": 6.571855985657552e-05,
      "loss": 0.917,
      "mean_token_accuracy": 0.8164054766297341,
      "num_tokens": 33763192.0,
      "step": 4400
    },
    {
      "entropy": 0.9067804291844368,
      "epoch": 0.6414545454545455,
      "grad_norm": 0.3304104208946228,
      "learning_rate": 6.525639415260563e-05,
      "loss": 0.8866,
      "mean_token_accuracy": 0.8219463184475899,
      "num_tokens": 33838644.0,
      "step": 4410
    },
    {
      "entropy": 0.9119504414498806,
      "epoch": 0.6429090909090909,
      "grad_norm": 0.8898224234580994,
      "learning_rate": 6.479507341209393e-05,
      "loss": 0.8943,
      "mean_token_accuracy": 0.8221902497112751,
      "num_tokens": 33914910.0,
      "step": 4420
    },
    {
      "entropy": 0.9117854446172714,
      "epoch": 0.6443636363636364,
      "grad_norm": 0.8432343006134033,
      "learning_rate": 6.433460883822452e-05,
      "loss": 0.8844,
      "mean_token_accuracy": 0.821457427740097,
      "num_tokens": 33990914.0,
      "step": 4430
    },
    {
      "entropy": 0.9375110886991024,
      "epoch": 0.6458181818181818,
      "grad_norm": 0.5810782313346863,
      "learning_rate": 6.387501161338928e-05,
      "loss": 0.9054,
      "mean_token_accuracy": 0.8187261156737804,
      "num_tokens": 34069434.0,
      "step": 4440
    },
    {
      "entropy": 0.937576549500227,
      "epoch": 0.6472727272727272,
      "grad_norm": 0.35709822177886963,
      "learning_rate": 6.341629289891659e-05,
      "loss": 0.9263,
      "mean_token_accuracy": 0.8171633973717689,
      "num_tokens": 34148749.0,
      "step": 4450
    },
    {
      "entropy": 0.9383350938558579,
      "epoch": 0.6487272727272727,
      "grad_norm": 0.5527284741401672,
      "learning_rate": 6.295846383480026e-05,
      "loss": 0.9314,
      "mean_token_accuracy": 0.8169618725776673,
      "num_tokens": 34225306.0,
      "step": 4460
    },
    {
      "entropy": 0.9290240414440631,
      "epoch": 0.6501818181818182,
      "grad_norm": 0.7125141024589539,
      "learning_rate": 6.250153553942878e-05,
      "loss": 0.8952,
      "mean_token_accuracy": 0.820123678445816,
      "num_tokens": 34301885.0,
      "step": 4470
    },
    {
      "entropy": 0.9101897843182087,
      "epoch": 0.6516363636363637,
      "grad_norm": 0.7149432897567749,
      "learning_rate": 6.204551910931557e-05,
      "loss": 0.893,
      "mean_token_accuracy": 0.8224738091230392,
      "num_tokens": 34378346.0,
      "step": 4480
    },
    {
      "entropy": 0.8988116376101971,
      "epoch": 0.653090909090909,
      "grad_norm": 1.1322219371795654,
      "learning_rate": 6.159042561882934e-05,
      "loss": 0.8757,
      "mean_token_accuracy": 0.8231632746756077,
      "num_tokens": 34454099.0,
      "step": 4490
    },
    {
      "entropy": 0.9239438407123088,
      "epoch": 0.6545454545454545,
      "grad_norm": 0.34981048107147217,
      "learning_rate": 6.113626611992519e-05,
      "loss": 0.9089,
      "mean_token_accuracy": 0.8181489415466785,
      "num_tokens": 34531037.0,
      "step": 4500
    },
    {
      "entropy": 0.9121604941785335,
      "epoch": 0.656,
      "grad_norm": 0.8667944669723511,
      "learning_rate": 6.0683051641876175e-05,
      "loss": 0.8849,
      "mean_token_accuracy": 0.8227492026984692,
      "num_tokens": 34606413.0,
      "step": 4510
    },
    {
      "entropy": 0.9448349975049496,
      "epoch": 0.6574545454545454,
      "grad_norm": 0.4684516191482544,
      "learning_rate": 6.023079319100555e-05,
      "loss": 0.9271,
      "mean_token_accuracy": 0.8161519937217235,
      "num_tokens": 34682859.0,
      "step": 4520
    },
    {
      "entropy": 0.9321636602282524,
      "epoch": 0.6589090909090909,
      "grad_norm": 0.3283658027648926,
      "learning_rate": 5.977950175041943e-05,
      "loss": 0.9136,
      "mean_token_accuracy": 0.8175563149154186,
      "num_tokens": 34761381.0,
      "step": 4530
    },
    {
      "entropy": 0.9159560807049274,
      "epoch": 0.6603636363636364,
      "grad_norm": 1.8966891765594482,
      "learning_rate": 5.932918827974003e-05,
      "loss": 0.8926,
      "mean_token_accuracy": 0.8220967844128608,
      "num_tokens": 34836927.0,
      "step": 4540
    },
    {
      "entropy": 0.9200945071876049,
      "epoch": 0.6618181818181819,
      "grad_norm": 0.3431169390678406,
      "learning_rate": 5.887986371483962e-05,
      "loss": 0.9019,
      "mean_token_accuracy": 0.8226698577404022,
      "num_tokens": 34913494.0,
      "step": 4550
    },
    {
      "entropy": 0.8993129044771194,
      "epoch": 0.6632727272727272,
      "grad_norm": 0.3218921720981598,
      "learning_rate": 5.843153896757475e-05,
      "loss": 0.8754,
      "mean_token_accuracy": 0.8246046505868435,
      "num_tokens": 34989288.0,
      "step": 4560
    },
    {
      "entropy": 0.9213286116719246,
      "epoch": 0.6647272727272727,
      "grad_norm": 0.44739851355552673,
      "learning_rate": 5.798422492552151e-05,
      "loss": 0.8968,
      "mean_token_accuracy": 0.8216261245310307,
      "num_tokens": 35065311.0,
      "step": 4570
    },
    {
      "entropy": 0.9399213455617428,
      "epoch": 0.6661818181818182,
      "grad_norm": 0.3798374533653259,
      "learning_rate": 5.7537932451710864e-05,
      "loss": 0.9269,
      "mean_token_accuracy": 0.816948776692152,
      "num_tokens": 35143191.0,
      "step": 4580
    },
    {
      "entropy": 0.9026488415896893,
      "epoch": 0.6676363636363636,
      "grad_norm": 0.3327493369579315,
      "learning_rate": 5.7092672384365045e-05,
      "loss": 0.8772,
      "mean_token_accuracy": 0.8241571098566055,
      "num_tokens": 35219781.0,
      "step": 4590
    },
    {
      "entropy": 0.9310063399374485,
      "epoch": 0.6690909090909091,
      "grad_norm": 0.5970094799995422,
      "learning_rate": 5.664845553663427e-05,
      "loss": 0.9159,
      "mean_token_accuracy": 0.8162389121949672,
      "num_tokens": 35301018.0,
      "step": 4600
    },
    {
      "entropy": 0.9052704386413097,
      "epoch": 0.6705454545454546,
      "grad_norm": 1.2368245124816895,
      "learning_rate": 5.620529269633419e-05,
      "loss": 0.8824,
      "mean_token_accuracy": 0.8242752820253372,
      "num_tokens": 35375305.0,
      "step": 4610
    },
    {
      "entropy": 0.9087167643010616,
      "epoch": 0.672,
      "grad_norm": 0.5536038875579834,
      "learning_rate": 5.5763194625683735e-05,
      "loss": 0.8892,
      "mean_token_accuracy": 0.8252256274223327,
      "num_tokens": 35449523.0,
      "step": 4620
    },
    {
      "entropy": 0.931006995588541,
      "epoch": 0.6734545454545454,
      "grad_norm": 0.7629865407943726,
      "learning_rate": 5.532217206104401e-05,
      "loss": 0.9011,
      "mean_token_accuracy": 0.8168572470545769,
      "num_tokens": 35525855.0,
      "step": 4630
    },
    {
      "entropy": 0.9062797881662845,
      "epoch": 0.6749090909090909,
      "grad_norm": 0.34807315468788147,
      "learning_rate": 5.488223571265733e-05,
      "loss": 0.8869,
      "mean_token_accuracy": 0.8224447555840015,
      "num_tokens": 35600766.0,
      "step": 4640
    },
    {
      "entropy": 0.9221105262637138,
      "epoch": 0.6763636363636364,
      "grad_norm": 0.3406118154525757,
      "learning_rate": 5.4443396264387304e-05,
      "loss": 0.9019,
      "mean_token_accuracy": 0.819121991097927,
      "num_tokens": 35677614.0,
      "step": 4650
    },
    {
      "entropy": 0.9147805757820606,
      "epoch": 0.6778181818181818,
      "grad_norm": 1.066024899482727,
      "learning_rate": 5.4005664373459285e-05,
      "loss": 0.8992,
      "mean_token_accuracy": 0.8201844908297062,
      "num_tokens": 35754102.0,
      "step": 4660
    },
    {
      "entropy": 0.8988436676561833,
      "epoch": 0.6792727272727273,
      "grad_norm": 0.47898566722869873,
      "learning_rate": 5.3569050670201516e-05,
      "loss": 0.8794,
      "mean_token_accuracy": 0.8236523412168026,
      "num_tokens": 35828877.0,
      "step": 4670
    },
    {
      "entropy": 0.917811781913042,
      "epoch": 0.6807272727272727,
      "grad_norm": 1.123764991760254,
      "learning_rate": 5.313356575778708e-05,
      "loss": 0.8936,
      "mean_token_accuracy": 0.8228101126849652,
      "num_tokens": 35903785.0,
      "step": 4680
    },
    {
      "entropy": 0.9052754618227482,
      "epoch": 0.6821818181818182,
      "grad_norm": 0.34496447443962097,
      "learning_rate": 5.269922021197635e-05,
      "loss": 0.8813,
      "mean_token_accuracy": 0.8240257889032364,
      "num_tokens": 35978342.0,
      "step": 4690
    },
    {
      "entropy": 0.9212935179471969,
      "epoch": 0.6836363636363636,
      "grad_norm": 0.34346017241477966,
      "learning_rate": 5.22660245808602e-05,
      "loss": 0.9071,
      "mean_token_accuracy": 0.8209236077964306,
      "num_tokens": 36055844.0,
      "step": 4700
    },
    {
      "entropy": 0.9109322354197502,
      "epoch": 0.6850909090909091,
      "grad_norm": 0.8080201148986816,
      "learning_rate": 5.1833989384603674e-05,
      "loss": 0.8938,
      "mean_token_accuracy": 0.8222977101802826,
      "num_tokens": 36133038.0,
      "step": 4710
    },
    {
      "entropy": 0.9155547618865967,
      "epoch": 0.6865454545454546,
      "grad_norm": 0.43798285722732544,
      "learning_rate": 5.140312511519072e-05,
      "loss": 0.8869,
      "mean_token_accuracy": 0.8229060113430023,
      "num_tokens": 36208973.0,
      "step": 4720
    },
    {
      "entropy": 0.9071539863944054,
      "epoch": 0.688,
      "grad_norm": 0.6105244755744934,
      "learning_rate": 5.097344223616927e-05,
      "loss": 0.8864,
      "mean_token_accuracy": 0.8228571750223637,
      "num_tokens": 36286284.0,
      "step": 4730
    },
    {
      "entropy": 0.9173211127519607,
      "epoch": 0.6894545454545454,
      "grad_norm": 0.4710405766963959,
      "learning_rate": 5.054495118239724e-05,
      "loss": 0.8989,
      "mean_token_accuracy": 0.8202463157474995,
      "num_tokens": 36364109.0,
      "step": 4740
    },
    {
      "entropy": 0.9191012769937515,
      "epoch": 0.6909090909090909,
      "grad_norm": 0.39346182346343994,
      "learning_rate": 5.011766235978892e-05,
      "loss": 0.9027,
      "mean_token_accuracy": 0.8204705208539963,
      "num_tokens": 36440260.0,
      "step": 4750
    },
    {
      "entropy": 0.9279646016657352,
      "epoch": 0.6923636363636364,
      "grad_norm": 0.46738266944885254,
      "learning_rate": 4.969158614506252e-05,
      "loss": 0.9008,
      "mean_token_accuracy": 0.8226397335529327,
      "num_tokens": 36515689.0,
      "step": 4760
    },
    {
      "entropy": 0.921799573302269,
      "epoch": 0.6938181818181818,
      "grad_norm": 2.155463457107544,
      "learning_rate": 4.9266732885487926e-05,
      "loss": 0.8976,
      "mean_token_accuracy": 0.8203616641461849,
      "num_tokens": 36592038.0,
      "step": 4770
    },
    {
      "entropy": 0.9005033083260059,
      "epoch": 0.6952727272727273,
      "grad_norm": 0.3272842466831207,
      "learning_rate": 4.88431128986356e-05,
      "loss": 0.8826,
      "mean_token_accuracy": 0.8243346132338047,
      "num_tokens": 36670238.0,
      "step": 4780
    },
    {
      "entropy": 0.9036756217479706,
      "epoch": 0.6967272727272728,
      "grad_norm": 0.3420988917350769,
      "learning_rate": 4.8420736472125994e-05,
      "loss": 0.9006,
      "mean_token_accuracy": 0.8209830693900585,
      "num_tokens": 36745652.0,
      "step": 4790
    },
    {
      "entropy": 0.9203361749649048,
      "epoch": 0.6981818181818182,
      "grad_norm": 1.3303462266921997,
      "learning_rate": 4.7999613863379513e-05,
      "loss": 0.9036,
      "mean_token_accuracy": 0.8206400662660599,
      "num_tokens": 36824621.0,
      "step": 4800
    },
    {
      "entropy": 0.9381427116692066,
      "epoch": 0.6996363636363636,
      "grad_norm": 0.4559713900089264,
      "learning_rate": 4.7579755299367713e-05,
      "loss": 0.9342,
      "mean_token_accuracy": 0.8164194501936436,
      "num_tokens": 36901731.0,
      "step": 4810
    },
    {
      "entropy": 0.9082010470330715,
      "epoch": 0.7010909090909091,
      "grad_norm": 0.31959229707717896,
      "learning_rate": 4.716117097636477e-05,
      "loss": 0.8837,
      "mean_token_accuracy": 0.8240106962621212,
      "num_tokens": 36979208.0,
      "step": 4820
    },
    {
      "entropy": 0.9182284452021122,
      "epoch": 0.7025454545454546,
      "grad_norm": 0.5974369645118713,
      "learning_rate": 4.67438710596999e-05,
      "loss": 0.8948,
      "mean_token_accuracy": 0.8215501293540001,
      "num_tokens": 37055307.0,
      "step": 4830
    },
    {
      "entropy": 0.9068014062941074,
      "epoch": 0.704,
      "grad_norm": 0.3271491825580597,
      "learning_rate": 4.6327865683510396e-05,
      "loss": 0.8898,
      "mean_token_accuracy": 0.8260469518601894,
      "num_tokens": 37129546.0,
      "step": 4840
    },
    {
      "entropy": 0.9046911887824536,
      "epoch": 0.7054545454545454,
      "grad_norm": 0.3307289481163025,
      "learning_rate": 4.591316495049567e-05,
      "loss": 0.8944,
      "mean_token_accuracy": 0.8209660328924656,
      "num_tokens": 37208382.0,
      "step": 4850
    },
    {
      "entropy": 0.906683124601841,
      "epoch": 0.7069090909090909,
      "grad_norm": 0.31656327843666077,
      "learning_rate": 4.549977893167188e-05,
      "loss": 0.8929,
      "mean_token_accuracy": 0.8215794719755649,
      "num_tokens": 37283806.0,
      "step": 4860
    },
    {
      "entropy": 0.9107134640216827,
      "epoch": 0.7083636363636364,
      "grad_norm": 0.33436915278434753,
      "learning_rate": 4.508771766612727e-05,
      "loss": 0.8912,
      "mean_token_accuracy": 0.8219263046979904,
      "num_tokens": 37362060.0,
      "step": 4870
    },
    {
      "entropy": 0.938027660548687,
      "epoch": 0.7098181818181818,
      "grad_norm": 0.3237261474132538,
      "learning_rate": 4.467699116077847e-05,
      "loss": 0.9307,
      "mean_token_accuracy": 0.8167957946658134,
      "num_tokens": 37439998.0,
      "step": 4880
    },
    {
      "entropy": 0.9301370844244957,
      "epoch": 0.7112727272727273,
      "grad_norm": 0.3891702890396118,
      "learning_rate": 4.4267609390127344e-05,
      "loss": 0.922,
      "mean_token_accuracy": 0.8177198253571987,
      "num_tokens": 37515912.0,
      "step": 4890
    },
    {
      "entropy": 0.911509207636118,
      "epoch": 0.7127272727272728,
      "grad_norm": 0.2914981544017792,
      "learning_rate": 4.385958229601898e-05,
      "loss": 0.8769,
      "mean_token_accuracy": 0.8222219668328762,
      "num_tokens": 37592700.0,
      "step": 4900
    },
    {
      "entropy": 0.9143662184476853,
      "epoch": 0.7141818181818181,
      "grad_norm": 0.3326908051967621,
      "learning_rate": 4.345291978739995e-05,
      "loss": 0.8977,
      "mean_token_accuracy": 0.8202255286276341,
      "num_tokens": 37669405.0,
      "step": 4910
    },
    {
      "entropy": 0.9253424435853959,
      "epoch": 0.7156363636363636,
      "grad_norm": 0.3408204913139343,
      "learning_rate": 4.304763174007798e-05,
      "loss": 0.8967,
      "mean_token_accuracy": 0.8195987194776535,
      "num_tokens": 37749117.0,
      "step": 4920
    },
    {
      "entropy": 0.9295443311333657,
      "epoch": 0.7170909090909091,
      "grad_norm": 0.3627084195613861,
      "learning_rate": 4.264372799648193e-05,
      "loss": 0.9048,
      "mean_token_accuracy": 0.8182410605251789,
      "num_tokens": 37828392.0,
      "step": 4930
    },
    {
      "entropy": 0.9167982846498489,
      "epoch": 0.7185454545454546,
      "grad_norm": 0.40901535749435425,
      "learning_rate": 4.224121836542277e-05,
      "loss": 0.9011,
      "mean_token_accuracy": 0.8199195042252541,
      "num_tokens": 37905217.0,
      "step": 4940
    },
    {
      "entropy": 0.9198883838951588,
      "epoch": 0.72,
      "grad_norm": 0.7457308173179626,
      "learning_rate": 4.1840112621855465e-05,
      "loss": 0.9056,
      "mean_token_accuracy": 0.8203414805233479,
      "num_tokens": 37982324.0,
      "step": 4950
    },
    {
      "entropy": 0.9170742996037007,
      "epoch": 0.7214545454545455,
      "grad_norm": 0.5554471015930176,
      "learning_rate": 4.144042050664154e-05,
      "loss": 0.8942,
      "mean_token_accuracy": 0.8203064344823361,
      "num_tokens": 38061849.0,
      "step": 4960
    },
    {
      "entropy": 0.9040765814483166,
      "epoch": 0.722909090909091,
      "grad_norm": 1.6444945335388184,
      "learning_rate": 4.104215172631255e-05,
      "loss": 0.8776,
      "mean_token_accuracy": 0.8247710056602955,
      "num_tokens": 38137324.0,
      "step": 4970
    },
    {
      "entropy": 0.9116667337715626,
      "epoch": 0.7243636363636363,
      "grad_norm": 0.5622697472572327,
      "learning_rate": 4.0645315952834264e-05,
      "loss": 0.9033,
      "mean_token_accuracy": 0.819647303968668,
      "num_tokens": 38214376.0,
      "step": 4980
    },
    {
      "entropy": 0.9142092511057853,
      "epoch": 0.7258181818181818,
      "grad_norm": 2.6932389736175537,
      "learning_rate": 4.024992282337184e-05,
      "loss": 0.884,
      "mean_token_accuracy": 0.8248827517032623,
      "num_tokens": 38287505.0,
      "step": 4990
    },
    {
      "entropy": 0.9225122049450875,
      "epoch": 0.7272727272727273,
      "grad_norm": 1.9417043924331665,
      "learning_rate": 3.9855981940055895e-05,
      "loss": 0.8931,
      "mean_token_accuracy": 0.8227336160838604,
      "num_tokens": 38364587.0,
      "step": 5000
    },
    {
      "epoch": 0.7272727272727273,
      "eval_entropy": 0.9167415462493896,
      "eval_loss": 0.944766104221344,
      "eval_mean_token_accuracy": 0.8129959221839905,
      "eval_num_tokens": 38364587.0,
      "eval_runtime": 78.076,
      "eval_samples_per_second": 64.04,
      "eval_steps_per_second": 8.005,
      "step": 5000
    },
    {
      "entropy": 0.9145618513226509,
      "epoch": 0.7287272727272728,
      "grad_norm": 1.0130982398986816,
      "learning_rate": 3.946350286974914e-05,
      "loss": 0.889,
      "mean_token_accuracy": 0.8210049398243427,
      "num_tokens": 38442118.0,
      "step": 5010
    },
    {
      "entropy": 0.9164961226284504,
      "epoch": 0.7301818181818182,
      "grad_norm": 0.3274339735507965,
      "learning_rate": 3.907249514381419e-05,
      "loss": 0.8902,
      "mean_token_accuracy": 0.821975689381361,
      "num_tokens": 38519059.0,
      "step": 5020
    },
    {
      "entropy": 0.903435967862606,
      "epoch": 0.7316363636363636,
      "grad_norm": 2.559694290161133,
      "learning_rate": 3.8682968257881954e-05,
      "loss": 0.8971,
      "mean_token_accuracy": 0.8234760813415051,
      "num_tokens": 38595562.0,
      "step": 5030
    },
    {
      "entropy": 0.8993699155747891,
      "epoch": 0.7330909090909091,
      "grad_norm": 0.8827573657035828,
      "learning_rate": 3.82949316716211e-05,
      "loss": 0.8912,
      "mean_token_accuracy": 0.8228749997913838,
      "num_tokens": 38668505.0,
      "step": 5040
    },
    {
      "entropy": 0.9492502108216285,
      "epoch": 0.7345454545454545,
      "grad_norm": 0.3277486264705658,
      "learning_rate": 3.7908394808508427e-05,
      "loss": 0.9327,
      "mean_token_accuracy": 0.8169902831315994,
      "num_tokens": 38748156.0,
      "step": 5050
    },
    {
      "entropy": 0.9224891781806945,
      "epoch": 0.736,
      "grad_norm": 3.9432334899902344,
      "learning_rate": 3.752336705559988e-05,
      "loss": 0.8988,
      "mean_token_accuracy": 0.8203701607882976,
      "num_tokens": 38825253.0,
      "step": 5060
    },
    {
      "entropy": 0.9180438406765461,
      "epoch": 0.7374545454545455,
      "grad_norm": 2.536931276321411,
      "learning_rate": 3.7139857763302665e-05,
      "loss": 0.8999,
      "mean_token_accuracy": 0.819855236262083,
      "num_tokens": 38904647.0,
      "step": 5070
    },
    {
      "entropy": 0.918037424236536,
      "epoch": 0.738909090909091,
      "grad_norm": 0.34127771854400635,
      "learning_rate": 3.675787624514802e-05,
      "loss": 0.9006,
      "mean_token_accuracy": 0.8193914517760277,
      "num_tokens": 38982580.0,
      "step": 5080
    },
    {
      "entropy": 0.9227818869054317,
      "epoch": 0.7403636363636363,
      "grad_norm": 0.3767578601837158,
      "learning_rate": 3.637743177756529e-05,
      "loss": 0.8943,
      "mean_token_accuracy": 0.8203182458877564,
      "num_tokens": 39058467.0,
      "step": 5090
    },
    {
      "entropy": 0.9119371868669987,
      "epoch": 0.7418181818181818,
      "grad_norm": 1.933066964149475,
      "learning_rate": 3.5998533599656495e-05,
      "loss": 0.8961,
      "mean_token_accuracy": 0.8217470467090606,
      "num_tokens": 39133672.0,
      "step": 5100
    },
    {
      "entropy": 0.9073677368462085,
      "epoch": 0.7432727272727273,
      "grad_norm": 0.3380196690559387,
      "learning_rate": 3.5621190912971904e-05,
      "loss": 0.8874,
      "mean_token_accuracy": 0.8223125129938126,
      "num_tokens": 39207784.0,
      "step": 5110
    },
    {
      "entropy": 0.9009414814412594,
      "epoch": 0.7447272727272727,
      "grad_norm": 3.2998690605163574,
      "learning_rate": 3.524541288128667e-05,
      "loss": 0.8741,
      "mean_token_accuracy": 0.8245840072631836,
      "num_tokens": 39281502.0,
      "step": 5120
    },
    {
      "entropy": 0.9070143595337867,
      "epoch": 0.7461818181818182,
      "grad_norm": 0.2989789843559265,
      "learning_rate": 3.4871208630378305e-05,
      "loss": 0.8772,
      "mean_token_accuracy": 0.8259104959666729,
      "num_tokens": 39354545.0,
      "step": 5130
    },
    {
      "entropy": 0.9156953662633895,
      "epoch": 0.7476363636363637,
      "grad_norm": 0.3448035418987274,
      "learning_rate": 3.449858724780502e-05,
      "loss": 0.9086,
      "mean_token_accuracy": 0.8190042689442635,
      "num_tokens": 39434017.0,
      "step": 5140
    },
    {
      "entropy": 0.9149294622242451,
      "epoch": 0.7490909090909091,
      "grad_norm": 0.3348819315433502,
      "learning_rate": 3.412755778268505e-05,
      "loss": 0.8929,
      "mean_token_accuracy": 0.8223293416202069,
      "num_tokens": 39510040.0,
      "step": 5150
    },
    {
      "entropy": 0.92013523504138,
      "epoch": 0.7505454545454545,
      "grad_norm": 0.7382948398590088,
      "learning_rate": 3.3758129245476766e-05,
      "loss": 0.9099,
      "mean_token_accuracy": 0.8201128125190735,
      "num_tokens": 39587287.0,
      "step": 5160
    },
    {
      "entropy": 0.9161408551037311,
      "epoch": 0.752,
      "grad_norm": 0.6711187958717346,
      "learning_rate": 3.339031060776014e-05,
      "loss": 0.9098,
      "mean_token_accuracy": 0.8177258402109147,
      "num_tokens": 39664088.0,
      "step": 5170
    },
    {
      "entropy": 0.9200255520641804,
      "epoch": 0.7534545454545455,
      "grad_norm": 0.34548547863960266,
      "learning_rate": 3.3024110802018535e-05,
      "loss": 0.8878,
      "mean_token_accuracy": 0.8237753942608833,
      "num_tokens": 39737765.0,
      "step": 5180
    },
    {
      "entropy": 0.9344703152775764,
      "epoch": 0.7549090909090909,
      "grad_norm": 0.3410409986972809,
      "learning_rate": 3.265953872142206e-05,
      "loss": 0.9205,
      "mean_token_accuracy": 0.8173702888190746,
      "num_tokens": 39816946.0,
      "step": 5190
    },
    {
      "entropy": 0.9066876843571663,
      "epoch": 0.7563636363636363,
      "grad_norm": 0.39815202355384827,
      "learning_rate": 3.229660321961147e-05,
      "loss": 0.8959,
      "mean_token_accuracy": 0.8212850585579872,
      "num_tokens": 39895255.0,
      "step": 5200
    },
    {
      "entropy": 0.9421879507601261,
      "epoch": 0.7578181818181818,
      "grad_norm": 1.315977931022644,
      "learning_rate": 3.193531311048311e-05,
      "loss": 0.9193,
      "mean_token_accuracy": 0.8162787385284901,
      "num_tokens": 39972246.0,
      "step": 5210
    },
    {
      "entropy": 0.9184886425733566,
      "epoch": 0.7592727272727273,
      "grad_norm": 0.32920485734939575,
      "learning_rate": 3.157567716797498e-05,
      "loss": 0.893,
      "mean_token_accuracy": 0.822785185277462,
      "num_tokens": 40049383.0,
      "step": 5220
    },
    {
      "entropy": 0.8912613660097122,
      "epoch": 0.7607272727272727,
      "grad_norm": 0.33051466941833496,
      "learning_rate": 3.121770412585364e-05,
      "loss": 0.8607,
      "mean_token_accuracy": 0.8278714805841446,
      "num_tokens": 40122483.0,
      "step": 5230
    },
    {
      "entropy": 0.9078203670680522,
      "epoch": 0.7621818181818182,
      "grad_norm": 0.32255810499191284,
      "learning_rate": 3.086140267750195e-05,
      "loss": 0.883,
      "mean_token_accuracy": 0.8227630846202374,
      "num_tokens": 40200942.0,
      "step": 5240
    },
    {
      "entropy": 0.9007733002305031,
      "epoch": 0.7636363636363637,
      "grad_norm": 0.3562650978565216,
      "learning_rate": 3.0506781475708234e-05,
      "loss": 0.8855,
      "mean_token_accuracy": 0.8229335404932498,
      "num_tokens": 40277363.0,
      "step": 5250
    },
    {
      "entropy": 0.9164358474314213,
      "epoch": 0.765090909090909,
      "grad_norm": 0.35381633043289185,
      "learning_rate": 3.0153849132455868e-05,
      "loss": 0.8919,
      "mean_token_accuracy": 0.8217478588223457,
      "num_tokens": 40354782.0,
      "step": 5260
    },
    {
      "entropy": 0.913616917282343,
      "epoch": 0.7665454545454545,
      "grad_norm": 0.35221511125564575,
      "learning_rate": 2.9802614218714342e-05,
      "loss": 0.8862,
      "mean_token_accuracy": 0.8239908576011657,
      "num_tokens": 40430145.0,
      "step": 5270
    },
    {
      "entropy": 0.9246010832488537,
      "epoch": 0.768,
      "grad_norm": 0.32586926221847534,
      "learning_rate": 2.9453085264231023e-05,
      "loss": 0.9003,
      "mean_token_accuracy": 0.8196610607206821,
      "num_tokens": 40507149.0,
      "step": 5280
    },
    {
      "entropy": 0.9094504900276661,
      "epoch": 0.7694545454545455,
      "grad_norm": 1.4361320734024048,
      "learning_rate": 2.910527075732401e-05,
      "loss": 0.8834,
      "mean_token_accuracy": 0.8229763463139534,
      "num_tokens": 40584650.0,
      "step": 5290
    },
    {
      "entropy": 0.9310476154088974,
      "epoch": 0.7709090909090909,
      "grad_norm": 0.33833417296409607,
      "learning_rate": 2.875917914467598e-05,
      "loss": 0.9038,
      "mean_token_accuracy": 0.8182396002113819,
      "num_tokens": 40662417.0,
      "step": 5300
    },
    {
      "entropy": 0.9101719483733177,
      "epoch": 0.7723636363636364,
      "grad_norm": 0.3172771632671356,
      "learning_rate": 2.8414818831129098e-05,
      "loss": 0.8864,
      "mean_token_accuracy": 0.821593638509512,
      "num_tokens": 40739831.0,
      "step": 5310
    },
    {
      "entropy": 0.9114699877798558,
      "epoch": 0.7738181818181818,
      "grad_norm": 2.5748441219329834,
      "learning_rate": 2.8072198179480908e-05,
      "loss": 0.8998,
      "mean_token_accuracy": 0.8192075364291668,
      "num_tokens": 40818006.0,
      "step": 5320
    },
    {
      "entropy": 0.9052566841244698,
      "epoch": 0.7752727272727272,
      "grad_norm": 0.8072139620780945,
      "learning_rate": 2.7731325510281252e-05,
      "loss": 0.8905,
      "mean_token_accuracy": 0.8216906942427158,
      "num_tokens": 40895087.0,
      "step": 5330
    },
    {
      "entropy": 0.9593810178339481,
      "epoch": 0.7767272727272727,
      "grad_norm": 1.659779667854309,
      "learning_rate": 2.7392209101630157e-05,
      "loss": 0.9655,
      "mean_token_accuracy": 0.8138592358678579,
      "num_tokens": 40972952.0,
      "step": 5340
    },
    {
      "entropy": 0.9060130745172501,
      "epoch": 0.7781818181818182,
      "grad_norm": 0.5863545536994934,
      "learning_rate": 2.7054857188976784e-05,
      "loss": 0.8805,
      "mean_token_accuracy": 0.8246444381773472,
      "num_tokens": 41049635.0,
      "step": 5350
    },
    {
      "entropy": 0.931407830119133,
      "epoch": 0.7796363636363637,
      "grad_norm": 0.34131190180778503,
      "learning_rate": 2.6719277964919527e-05,
      "loss": 0.9083,
      "mean_token_accuracy": 0.8182107910513878,
      "num_tokens": 41126152.0,
      "step": 5360
    },
    {
      "entropy": 0.9074273183941841,
      "epoch": 0.7810909090909091,
      "grad_norm": 0.35227325558662415,
      "learning_rate": 2.6385479579007012e-05,
      "loss": 0.8771,
      "mean_token_accuracy": 0.8249678060412406,
      "num_tokens": 41200196.0,
      "step": 5370
    },
    {
      "entropy": 0.9232403814792634,
      "epoch": 0.7825454545454545,
      "grad_norm": 2.6211049556732178,
      "learning_rate": 2.6053470137540077e-05,
      "loss": 0.895,
      "mean_token_accuracy": 0.8201713971793652,
      "num_tokens": 41276170.0,
      "step": 5380
    },
    {
      "entropy": 0.9107597105205059,
      "epoch": 0.784,
      "grad_norm": 0.9971133470535278,
      "learning_rate": 2.5723257703375147e-05,
      "loss": 0.8952,
      "mean_token_accuracy": 0.8221598446369172,
      "num_tokens": 41354544.0,
      "step": 5390
    },
    {
      "entropy": 0.9056881733238697,
      "epoch": 0.7854545454545454,
      "grad_norm": 0.4628162384033203,
      "learning_rate": 2.5394850295728146e-05,
      "loss": 0.8938,
      "mean_token_accuracy": 0.8212578587234021,
      "num_tokens": 41430868.0,
      "step": 5400
    },
    {
      "entropy": 0.9165485695004463,
      "epoch": 0.7869090909090909,
      "grad_norm": 0.3809981346130371,
      "learning_rate": 2.5068255889979976e-05,
      "loss": 0.8876,
      "mean_token_accuracy": 0.8206007726490497,
      "num_tokens": 41508162.0,
      "step": 5410
    },
    {
      "entropy": 0.9290221504867077,
      "epoch": 0.7883636363636364,
      "grad_norm": 0.32021892070770264,
      "learning_rate": 2.4743482417482735e-05,
      "loss": 0.9133,
      "mean_token_accuracy": 0.8185132816433907,
      "num_tokens": 41588151.0,
      "step": 5420
    },
    {
      "entropy": 0.9080618299543858,
      "epoch": 0.7898181818181819,
      "grad_norm": 0.32850199937820435,
      "learning_rate": 2.4420537765367142e-05,
      "loss": 0.8793,
      "mean_token_accuracy": 0.8246118858456611,
      "num_tokens": 41663786.0,
      "step": 5430
    },
    {
      "entropy": 0.9262031398713588,
      "epoch": 0.7912727272727272,
      "grad_norm": 2.3897554874420166,
      "learning_rate": 2.409942977635091e-05,
      "loss": 0.9047,
      "mean_token_accuracy": 0.8178823798894882,
      "num_tokens": 41743108.0,
      "step": 5440
    },
    {
      "entropy": 0.9042971536517144,
      "epoch": 0.7927272727272727,
      "grad_norm": 0.3209560811519623,
      "learning_rate": 2.3780166248548364e-05,
      "loss": 0.8868,
      "mean_token_accuracy": 0.8218206010758877,
      "num_tokens": 41819645.0,
      "step": 5450
    },
    {
      "entropy": 0.9179877236485481,
      "epoch": 0.7941818181818182,
      "grad_norm": 0.3202578127384186,
      "learning_rate": 2.346275493528109e-05,
      "loss": 0.9153,
      "mean_token_accuracy": 0.820369104295969,
      "num_tokens": 41896535.0,
      "step": 5460
    },
    {
      "entropy": 0.9163044169545174,
      "epoch": 0.7956363636363636,
      "grad_norm": 0.817571759223938,
      "learning_rate": 2.314720354488959e-05,
      "loss": 0.8894,
      "mean_token_accuracy": 0.819698441028595,
      "num_tokens": 41973115.0,
      "step": 5470
    },
    {
      "entropy": 0.9083671689033508,
      "epoch": 0.7970909090909091,
      "grad_norm": 1.006474494934082,
      "learning_rate": 2.2833519740546103e-05,
      "loss": 0.8804,
      "mean_token_accuracy": 0.8232364565134048,
      "num_tokens": 42050256.0,
      "step": 5480
    },
    {
      "entropy": 0.9098470561206341,
      "epoch": 0.7985454545454546,
      "grad_norm": 0.35074856877326965,
      "learning_rate": 2.252171114006847e-05,
      "loss": 0.9051,
      "mean_token_accuracy": 0.8213311359286308,
      "num_tokens": 42125229.0,
      "step": 5490
    },
    {
      "entropy": 0.9336720332503319,
      "epoch": 0.8,
      "grad_norm": 0.3655327260494232,
      "learning_rate": 2.2211785315735213e-05,
      "loss": 0.919,
      "mean_token_accuracy": 0.8171964854001998,
      "num_tokens": 42202853.0,
      "step": 5500
    },
    {
      "entropy": 0.9251384161412716,
      "epoch": 0.8014545454545454,
      "grad_norm": 0.3506443202495575,
      "learning_rate": 2.1903749794101522e-05,
      "loss": 0.9048,
      "mean_token_accuracy": 0.8205279417335987,
      "num_tokens": 42279546.0,
      "step": 5510
    },
    {
      "entropy": 0.9413960322737693,
      "epoch": 0.8029090909090909,
      "grad_norm": 0.3986915051937103,
      "learning_rate": 2.159761205581664e-05,
      "loss": 0.9188,
      "mean_token_accuracy": 0.8179226972162723,
      "num_tokens": 42356932.0,
      "step": 5520
    },
    {
      "entropy": 0.9280008278787136,
      "epoch": 0.8043636363636364,
      "grad_norm": 1.7976831197738647,
      "learning_rate": 2.129337953544203e-05,
      "loss": 0.9192,
      "mean_token_accuracy": 0.817876136302948,
      "num_tokens": 42434945.0,
      "step": 5530
    },
    {
      "entropy": 0.917744517326355,
      "epoch": 0.8058181818181818,
      "grad_norm": 0.670444130897522,
      "learning_rate": 2.0991059621270915e-05,
      "loss": 0.8856,
      "mean_token_accuracy": 0.8219639703631401,
      "num_tokens": 42512155.0,
      "step": 5540
    },
    {
      "entropy": 0.9068499661982059,
      "epoch": 0.8072727272727273,
      "grad_norm": 0.5417830348014832,
      "learning_rate": 2.0690659655148857e-05,
      "loss": 0.8847,
      "mean_token_accuracy": 0.8228611163794994,
      "num_tokens": 42589538.0,
      "step": 5550
    },
    {
      "entropy": 0.8938118264079093,
      "epoch": 0.8087272727272727,
      "grad_norm": 0.8048520088195801,
      "learning_rate": 2.039218693229545e-05,
      "loss": 0.8767,
      "mean_token_accuracy": 0.825986260175705,
      "num_tokens": 42663055.0,
      "step": 5560
    },
    {
      "entropy": 0.9236326240003109,
      "epoch": 0.8101818181818182,
      "grad_norm": 1.3545376062393188,
      "learning_rate": 2.009564870112709e-05,
      "loss": 0.9078,
      "mean_token_accuracy": 0.8180381782352925,
      "num_tokens": 42740723.0,
      "step": 5570
    },
    {
      "entropy": 0.8940872102975845,
      "epoch": 0.8116363636363636,
      "grad_norm": 0.5655390024185181,
      "learning_rate": 1.9801052163081006e-05,
      "loss": 0.8713,
      "mean_token_accuracy": 0.8249186836183071,
      "num_tokens": 42816045.0,
      "step": 5580
    },
    {
      "entropy": 0.914474855363369,
      "epoch": 0.8130909090909091,
      "grad_norm": 0.35351067781448364,
      "learning_rate": 1.9508404472440418e-05,
      "loss": 0.8969,
      "mean_token_accuracy": 0.8209439516067505,
      "num_tokens": 42893348.0,
      "step": 5590
    },
    {
      "entropy": 0.9080572672188282,
      "epoch": 0.8145454545454546,
      "grad_norm": 0.4727298617362976,
      "learning_rate": 1.9217712736160705e-05,
      "loss": 0.8863,
      "mean_token_accuracy": 0.8232846416532993,
      "num_tokens": 42969891.0,
      "step": 5600
    },
    {
      "entropy": 0.9193252146244049,
      "epoch": 0.816,
      "grad_norm": 0.3219492435455322,
      "learning_rate": 1.8928984013696927e-05,
      "loss": 0.8887,
      "mean_token_accuracy": 0.8216960929334164,
      "num_tokens": 43048711.0,
      "step": 5610
    },
    {
      "entropy": 0.9039079166948796,
      "epoch": 0.8174545454545454,
      "grad_norm": 0.41121330857276917,
      "learning_rate": 1.8642225316832167e-05,
      "loss": 0.8823,
      "mean_token_accuracy": 0.8243647992610932,
      "num_tokens": 43124290.0,
      "step": 5620
    },
    {
      "entropy": 0.9116926275193691,
      "epoch": 0.8189090909090909,
      "grad_norm": 0.4170006215572357,
      "learning_rate": 1.8357443609507547e-05,
      "loss": 0.8885,
      "mean_token_accuracy": 0.8218486055731773,
      "num_tokens": 43200155.0,
      "step": 5630
    },
    {
      "entropy": 0.8914009846746922,
      "epoch": 0.8203636363636364,
      "grad_norm": 0.3774668574333191,
      "learning_rate": 1.8074645807652825e-05,
      "loss": 0.8614,
      "mean_token_accuracy": 0.8262716047465801,
      "num_tokens": 43274762.0,
      "step": 5640
    },
    {
      "entropy": 0.9105676539242268,
      "epoch": 0.8218181818181818,
      "grad_norm": 0.3389342427253723,
      "learning_rate": 1.7793838779018674e-05,
      "loss": 0.9006,
      "mean_token_accuracy": 0.821910735219717,
      "num_tokens": 43352595.0,
      "step": 5650
    },
    {
      "entropy": 0.9219238504767417,
      "epoch": 0.8232727272727273,
      "grad_norm": 0.45230501890182495,
      "learning_rate": 1.7515029343009777e-05,
      "loss": 0.9283,
      "mean_token_accuracy": 0.8188553206622601,
      "num_tokens": 43429221.0,
      "step": 5660
    },
    {
      "entropy": 0.9218499854207038,
      "epoch": 0.8247272727272728,
      "grad_norm": 0.4920508861541748,
      "learning_rate": 1.7238224270519165e-05,
      "loss": 0.9065,
      "mean_token_accuracy": 0.8194623336195945,
      "num_tokens": 43505245.0,
      "step": 5670
    },
    {
      "entropy": 0.8820628426969052,
      "epoch": 0.8261818181818181,
      "grad_norm": 0.4515336751937866,
      "learning_rate": 1.696343028376394e-05,
      "loss": 0.8686,
      "mean_token_accuracy": 0.8276709161698819,
      "num_tokens": 43579860.0,
      "step": 5680
    },
    {
      "entropy": 0.9073334261775017,
      "epoch": 0.8276363636363636,
      "grad_norm": 0.3067150115966797,
      "learning_rate": 1.669065405612193e-05,
      "loss": 0.8829,
      "mean_token_accuracy": 0.8239874728024006,
      "num_tokens": 43655093.0,
      "step": 5690
    },
    {
      "entropy": 0.9091994293034077,
      "epoch": 0.8290909090909091,
      "grad_norm": 0.3159719705581665,
      "learning_rate": 1.6419902211969643e-05,
      "loss": 0.8874,
      "mean_token_accuracy": 0.8220887213945389,
      "num_tokens": 43731458.0,
      "step": 5700
    },
    {
      "entropy": 0.9184407569468022,
      "epoch": 0.8305454545454546,
      "grad_norm": 0.5727894902229309,
      "learning_rate": 1.6151181326521332e-05,
      "loss": 0.9082,
      "mean_token_accuracy": 0.8187850311398506,
      "num_tokens": 43807665.0,
      "step": 5710
    },
    {
      "entropy": 0.9286717921495438,
      "epoch": 0.832,
      "grad_norm": 0.7335017323493958,
      "learning_rate": 1.5884497925669418e-05,
      "loss": 0.9085,
      "mean_token_accuracy": 0.8177166223526001,
      "num_tokens": 43885791.0,
      "step": 5720
    },
    {
      "entropy": 0.9017570033669472,
      "epoch": 0.8334545454545454,
      "grad_norm": 0.30094099044799805,
      "learning_rate": 1.5619858485825967e-05,
      "loss": 0.8766,
      "mean_token_accuracy": 0.823994617909193,
      "num_tokens": 43963234.0,
      "step": 5730
    },
    {
      "entropy": 0.9207366243004799,
      "epoch": 0.8349090909090909,
      "grad_norm": 0.3066195845603943,
      "learning_rate": 1.535726943376543e-05,
      "loss": 0.9014,
      "mean_token_accuracy": 0.8193856805562973,
      "num_tokens": 44041938.0,
      "step": 5740
    },
    {
      "entropy": 0.9334449142217636,
      "epoch": 0.8363636363636363,
      "grad_norm": 0.3157121539115906,
      "learning_rate": 1.5096737146468522e-05,
      "loss": 0.9157,
      "mean_token_accuracy": 0.8169816955924034,
      "num_tokens": 44119391.0,
      "step": 5750
    },
    {
      "entropy": 0.9535537295043468,
      "epoch": 0.8378181818181818,
      "grad_norm": 0.3890473544597626,
      "learning_rate": 1.4838267950967356e-05,
      "loss": 0.9345,
      "mean_token_accuracy": 0.8143138892948627,
      "num_tokens": 44197991.0,
      "step": 5760
    },
    {
      "entropy": 0.9177822828292846,
      "epoch": 0.8392727272727273,
      "grad_norm": 0.3412627577781677,
      "learning_rate": 1.458186812419187e-05,
      "loss": 0.8999,
      "mean_token_accuracy": 0.8185959592461586,
      "num_tokens": 44276841.0,
      "step": 5770
    },
    {
      "entropy": 0.912161386013031,
      "epoch": 0.8407272727272728,
      "grad_norm": 0.36407360434532166,
      "learning_rate": 1.4327543892817288e-05,
      "loss": 0.8862,
      "mean_token_accuracy": 0.8223249286413192,
      "num_tokens": 44353553.0,
      "step": 5780
    },
    {
      "entropy": 0.8996298290789128,
      "epoch": 0.8421818181818181,
      "grad_norm": 0.3385894000530243,
      "learning_rate": 1.407530143311299e-05,
      "loss": 0.8744,
      "mean_token_accuracy": 0.8245961762964725,
      "num_tokens": 44430454.0,
      "step": 5790
    },
    {
      "entropy": 0.8858574166893959,
      "epoch": 0.8436363636363636,
      "grad_norm": 0.3109019100666046,
      "learning_rate": 1.3825146870792487e-05,
      "loss": 0.8609,
      "mean_token_accuracy": 0.8287350215017796,
      "num_tokens": 44504637.0,
      "step": 5800
    },
    {
      "entropy": 0.905403683334589,
      "epoch": 0.8450909090909091,
      "grad_norm": 0.33812251687049866,
      "learning_rate": 1.3577086280864632e-05,
      "loss": 0.8917,
      "mean_token_accuracy": 0.8217367753386497,
      "num_tokens": 44582467.0,
      "step": 5810
    },
    {
      "entropy": 0.8935080736875534,
      "epoch": 0.8465454545454546,
      "grad_norm": 0.4578856825828552,
      "learning_rate": 1.3331125687486126e-05,
      "loss": 0.8703,
      "mean_token_accuracy": 0.8245539911091327,
      "num_tokens": 44659005.0,
      "step": 5820
    },
    {
      "entropy": 0.9361538954079152,
      "epoch": 0.848,
      "grad_norm": 0.31013602018356323,
      "learning_rate": 1.3087271063815244e-05,
      "loss": 0.923,
      "mean_token_accuracy": 0.8165486060082913,
      "num_tokens": 44739227.0,
      "step": 5830
    },
    {
      "entropy": 0.9088335067033768,
      "epoch": 0.8494545454545455,
      "grad_norm": 0.32604363560676575,
      "learning_rate": 1.2845528331866671e-05,
      "loss": 0.8863,
      "mean_token_accuracy": 0.8229722075164319,
      "num_tokens": 44816692.0,
      "step": 5840
    },
    {
      "entropy": 0.8995483376085758,
      "epoch": 0.850909090909091,
      "grad_norm": 0.3095364570617676,
      "learning_rate": 1.2605903362367824e-05,
      "loss": 0.875,
      "mean_token_accuracy": 0.8238675244152546,
      "num_tokens": 44893443.0,
      "step": 5850
    },
    {
      "entropy": 0.9014830864965916,
      "epoch": 0.8523636363636363,
      "grad_norm": 0.49402737617492676,
      "learning_rate": 1.2368401974616153e-05,
      "loss": 0.8729,
      "mean_token_accuracy": 0.8236888878047466,
      "num_tokens": 44970393.0,
      "step": 5860
    },
    {
      "entropy": 0.9028633058071136,
      "epoch": 0.8538181818181818,
      "grad_norm": 0.30887338519096375,
      "learning_rate": 1.213302993633793e-05,
      "loss": 0.8759,
      "mean_token_accuracy": 0.8243600711226463,
      "num_tokens": 45046091.0,
      "step": 5870
    },
    {
      "entropy": 0.9114751204848289,
      "epoch": 0.8552727272727273,
      "grad_norm": 0.9053654670715332,
      "learning_rate": 1.1899792963548117e-05,
      "loss": 0.8926,
      "mean_token_accuracy": 0.8210045278072358,
      "num_tokens": 45124224.0,
      "step": 5880
    },
    {
      "entropy": 0.9158106587827206,
      "epoch": 0.8567272727272728,
      "grad_norm": 0.3252062499523163,
      "learning_rate": 1.1668696720411572e-05,
      "loss": 0.8935,
      "mean_token_accuracy": 0.8214658476412297,
      "num_tokens": 45200375.0,
      "step": 5890
    },
    {
      "entropy": 0.9093938983976841,
      "epoch": 0.8581818181818182,
      "grad_norm": 0.36540377140045166,
      "learning_rate": 1.143974681910547e-05,
      "loss": 0.9017,
      "mean_token_accuracy": 0.8203778505325318,
      "num_tokens": 45276728.0,
      "step": 5900
    },
    {
      "entropy": 0.9019094638526439,
      "epoch": 0.8596363636363636,
      "grad_norm": 0.9092381596565247,
      "learning_rate": 1.1212948819682999e-05,
      "loss": 0.8854,
      "mean_token_accuracy": 0.8230563633143901,
      "num_tokens": 45351853.0,
      "step": 5910
    },
    {
      "entropy": 0.9108839437365532,
      "epoch": 0.8610909090909091,
      "grad_norm": 0.3634931147098541,
      "learning_rate": 1.098830822993843e-05,
      "loss": 0.8772,
      "mean_token_accuracy": 0.8236395262181759,
      "num_tokens": 45427231.0,
      "step": 5920
    },
    {
      "entropy": 0.9179142944514751,
      "epoch": 0.8625454545454545,
      "grad_norm": 0.32144612073898315,
      "learning_rate": 1.0765830505273273e-05,
      "loss": 0.8924,
      "mean_token_accuracy": 0.8206780150532722,
      "num_tokens": 45503521.0,
      "step": 5930
    },
    {
      "entropy": 0.9086673229932785,
      "epoch": 0.864,
      "grad_norm": 0.3346118927001953,
      "learning_rate": 1.0545521048563785e-05,
      "loss": 0.8869,
      "mean_token_accuracy": 0.8204560600221157,
      "num_tokens": 45581502.0,
      "step": 5940
    },
    {
      "entropy": 0.9081209503114224,
      "epoch": 0.8654545454545455,
      "grad_norm": 0.6537035703659058,
      "learning_rate": 1.0327385210029801e-05,
      "loss": 0.8852,
      "mean_token_accuracy": 0.8215902402997017,
      "num_tokens": 45659228.0,
      "step": 5950
    },
    {
      "entropy": 0.9263642437756061,
      "epoch": 0.866909090909091,
      "grad_norm": 0.3149193823337555,
      "learning_rate": 1.0111428287104829e-05,
      "loss": 0.9329,
      "mean_token_accuracy": 0.8180480360984802,
      "num_tokens": 45737235.0,
      "step": 5960
    },
    {
      "entropy": 0.9080370858311653,
      "epoch": 0.8683636363636363,
      "grad_norm": 0.33345770835876465,
      "learning_rate": 9.897655524307342e-06,
      "loss": 0.8877,
      "mean_token_accuracy": 0.8218265272676945,
      "num_tokens": 45814371.0,
      "step": 5970
    },
    {
      "entropy": 0.899022338539362,
      "epoch": 0.8698181818181818,
      "grad_norm": 1.97817862033844,
      "learning_rate": 9.686072113113407e-06,
      "loss": 0.8803,
      "mean_token_accuracy": 0.8241935893893242,
      "num_tokens": 45890122.0,
      "step": 5980
    },
    {
      "entropy": 0.9251417450606823,
      "epoch": 0.8712727272727273,
      "grad_norm": 0.3278456926345825,
      "learning_rate": 9.476683191830666e-06,
      "loss": 0.8925,
      "mean_token_accuracy": 0.8199293427169323,
      "num_tokens": 45969291.0,
      "step": 5990
    },
    {
      "entropy": 0.899966686218977,
      "epoch": 0.8727272727272727,
      "grad_norm": 0.324158251285553,
      "learning_rate": 9.269493845473523e-06,
      "loss": 0.8847,
      "mean_token_accuracy": 0.8225129701197147,
      "num_tokens": 46044659.0,
      "step": 6000
    },
    {
      "epoch": 0.8727272727272727,
      "eval_entropy": 0.9082655986785889,
      "eval_loss": 0.9296354651451111,
      "eval_mean_token_accuracy": 0.8134225456237792,
      "eval_num_tokens": 46044659.0,
      "eval_runtime": 78.1885,
      "eval_samples_per_second": 63.948,
      "eval_steps_per_second": 7.994,
      "step": 6000
    },
    {
      "entropy": 0.8993717812001705,
      "epoch": 0.8741818181818182,
      "grad_norm": 0.33104297518730164,
      "learning_rate": 9.06450910563968e-06,
      "loss": 0.8686,
      "mean_token_accuracy": 0.8254872515797615,
      "num_tokens": 46119591.0,
      "step": 6010
    },
    {
      "entropy": 0.9017520911991597,
      "epoch": 0.8756363636363637,
      "grad_norm": 0.3194045424461365,
      "learning_rate": 8.861733950387915e-06,
      "loss": 0.8883,
      "mean_token_accuracy": 0.8212776839733124,
      "num_tokens": 46197101.0,
      "step": 6020
    },
    {
      "entropy": 0.9101865872740745,
      "epoch": 0.8770909090909091,
      "grad_norm": 0.30542537569999695,
      "learning_rate": 8.661173304117178e-06,
      "loss": 0.889,
      "mean_token_accuracy": 0.8223033159971237,
      "num_tokens": 46273776.0,
      "step": 6030
    },
    {
      "entropy": 0.891840324550867,
      "epoch": 0.8785454545454545,
      "grad_norm": 0.6645185351371765,
      "learning_rate": 8.462832037446993e-06,
      "loss": 0.8706,
      "mean_token_accuracy": 0.8264122419059277,
      "num_tokens": 46347359.0,
      "step": 6040
    },
    {
      "entropy": 0.9125166840851306,
      "epoch": 0.88,
      "grad_norm": 0.3202907145023346,
      "learning_rate": 8.266714967099286e-06,
      "loss": 0.8944,
      "mean_token_accuracy": 0.8212385781109333,
      "num_tokens": 46424205.0,
      "step": 6050
    },
    {
      "entropy": 0.9032726883888245,
      "epoch": 0.8814545454545455,
      "grad_norm": 0.8547842502593994,
      "learning_rate": 8.072826855781257e-06,
      "loss": 0.8836,
      "mean_token_accuracy": 0.8238522745668888,
      "num_tokens": 46501005.0,
      "step": 6060
    },
    {
      "entropy": 0.9115559689700603,
      "epoch": 0.8829090909090909,
      "grad_norm": 0.2994537949562073,
      "learning_rate": 7.88117241206985e-06,
      "loss": 0.8964,
      "mean_token_accuracy": 0.8225395679473877,
      "num_tokens": 46577416.0,
      "step": 6070
    },
    {
      "entropy": 0.9119720533490181,
      "epoch": 0.8843636363636364,
      "grad_norm": 0.31298911571502686,
      "learning_rate": 7.691756290297259e-06,
      "loss": 0.879,
      "mean_token_accuracy": 0.8213375210762024,
      "num_tokens": 46654800.0,
      "step": 6080
    },
    {
      "entropy": 0.8955303199589253,
      "epoch": 0.8858181818181818,
      "grad_norm": 6.103427886962891,
      "learning_rate": 7.504583090438062e-06,
      "loss": 0.8648,
      "mean_token_accuracy": 0.8251885525882244,
      "num_tokens": 46730256.0,
      "step": 6090
    },
    {
      "entropy": 0.8955484725534916,
      "epoch": 0.8872727272727273,
      "grad_norm": 0.3136664032936096,
      "learning_rate": 7.319657357997413e-06,
      "loss": 0.8783,
      "mean_token_accuracy": 0.8237858466804028,
      "num_tokens": 46805882.0,
      "step": 6100
    },
    {
      "entropy": 0.9244707800447941,
      "epoch": 0.8887272727272727,
      "grad_norm": 0.34006479382514954,
      "learning_rate": 7.136983583900615e-06,
      "loss": 0.9219,
      "mean_token_accuracy": 0.8182192638516426,
      "num_tokens": 46883478.0,
      "step": 6110
    },
    {
      "entropy": 0.9033333383500576,
      "epoch": 0.8901818181818182,
      "grad_norm": 0.34446612000465393,
      "learning_rate": 6.956566204384199e-06,
      "loss": 0.8934,
      "mean_token_accuracy": 0.8230147622525692,
      "num_tokens": 46958821.0,
      "step": 6120
    },
    {
      "entropy": 0.8889298714697361,
      "epoch": 0.8916363636363637,
      "grad_norm": 0.31922370195388794,
      "learning_rate": 6.778409600888029e-06,
      "loss": 0.8707,
      "mean_token_accuracy": 0.8255295574665069,
      "num_tokens": 47034208.0,
      "step": 6130
    },
    {
      "entropy": 0.9161468952894211,
      "epoch": 0.893090909090909,
      "grad_norm": 3.375911235809326,
      "learning_rate": 6.6025180999490305e-06,
      "loss": 0.9118,
      "mean_token_accuracy": 0.8201681829988956,
      "num_tokens": 47111500.0,
      "step": 6140
    },
    {
      "entropy": 0.9138515323400498,
      "epoch": 0.8945454545454545,
      "grad_norm": 0.3383484184741974,
      "learning_rate": 6.428895973096056e-06,
      "loss": 0.8851,
      "mean_token_accuracy": 0.8209197491407394,
      "num_tokens": 47188237.0,
      "step": 6150
    },
    {
      "entropy": 0.9050987802445889,
      "epoch": 0.896,
      "grad_norm": 0.36527344584465027,
      "learning_rate": 6.257547436746141e-06,
      "loss": 0.8905,
      "mean_token_accuracy": 0.8225465290248394,
      "num_tokens": 47264603.0,
      "step": 6160
    },
    {
      "entropy": 0.9179876156151294,
      "epoch": 0.8974545454545455,
      "grad_norm": 0.3297809064388275,
      "learning_rate": 6.088476652102145e-06,
      "loss": 0.8951,
      "mean_token_accuracy": 0.8211302883923054,
      "num_tokens": 47340993.0,
      "step": 6170
    },
    {
      "entropy": 0.9083369344472885,
      "epoch": 0.8989090909090909,
      "grad_norm": 1.8338009119033813,
      "learning_rate": 5.921687725051632e-06,
      "loss": 0.8852,
      "mean_token_accuracy": 0.82315573990345,
      "num_tokens": 47415150.0,
      "step": 6180
    },
    {
      "entropy": 0.8889398947358131,
      "epoch": 0.9003636363636364,
      "grad_norm": 0.3286149501800537,
      "learning_rate": 5.757184706067277e-06,
      "loss": 0.869,
      "mean_token_accuracy": 0.8250431105494499,
      "num_tokens": 47491474.0,
      "step": 6190
    },
    {
      "entropy": 0.905424141138792,
      "epoch": 0.9018181818181819,
      "grad_norm": 0.3469639718532562,
      "learning_rate": 5.594971590108399e-06,
      "loss": 0.9018,
      "mean_token_accuracy": 0.8211894407868385,
      "num_tokens": 47567827.0,
      "step": 6200
    },
    {
      "entropy": 0.9114717729389668,
      "epoch": 0.9032727272727272,
      "grad_norm": 1.8720717430114746,
      "learning_rate": 5.4350523165239645e-06,
      "loss": 0.8947,
      "mean_token_accuracy": 0.820029516518116,
      "num_tokens": 47647385.0,
      "step": 6210
    },
    {
      "entropy": 0.8960614062845706,
      "epoch": 0.9047272727272727,
      "grad_norm": 0.32637879252433777,
      "learning_rate": 5.277430768956922e-06,
      "loss": 0.8671,
      "mean_token_accuracy": 0.8259123407304287,
      "num_tokens": 47720917.0,
      "step": 6220
    },
    {
      "entropy": 0.9236697740852833,
      "epoch": 0.9061818181818182,
      "grad_norm": 0.3289095163345337,
      "learning_rate": 5.122110775249927e-06,
      "loss": 0.8918,
      "mean_token_accuracy": 0.8198946259915829,
      "num_tokens": 47798165.0,
      "step": 6230
    },
    {
      "entropy": 0.9044543921947479,
      "epoch": 0.9076363636363637,
      "grad_norm": 0.3405933976173401,
      "learning_rate": 4.969096107352316e-06,
      "loss": 0.8831,
      "mean_token_accuracy": 0.8225955821573734,
      "num_tokens": 47873847.0,
      "step": 6240
    },
    {
      "entropy": 0.9037906058132649,
      "epoch": 0.9090909090909091,
      "grad_norm": 0.32288992404937744,
      "learning_rate": 4.818390481228547e-06,
      "loss": 0.8831,
      "mean_token_accuracy": 0.8237175151705742,
      "num_tokens": 47950263.0,
      "step": 6250
    },
    {
      "entropy": 0.8919797301292419,
      "epoch": 0.9105454545454545,
      "grad_norm": 0.3142983913421631,
      "learning_rate": 4.669997556767988e-06,
      "loss": 0.8745,
      "mean_token_accuracy": 0.8249924443662167,
      "num_tokens": 48025710.0,
      "step": 6260
    },
    {
      "entropy": 0.8958311684429645,
      "epoch": 0.912,
      "grad_norm": 0.301633358001709,
      "learning_rate": 4.52392093769593e-06,
      "loss": 0.882,
      "mean_token_accuracy": 0.8238031409680844,
      "num_tokens": 48101876.0,
      "step": 6270
    },
    {
      "entropy": 0.9080388635396958,
      "epoch": 0.9134545454545454,
      "grad_norm": 0.3057876229286194,
      "learning_rate": 4.380164171486198e-06,
      "loss": 0.8974,
      "mean_token_accuracy": 0.8207934826612473,
      "num_tokens": 48181398.0,
      "step": 6280
    },
    {
      "entropy": 0.9020656362175942,
      "epoch": 0.9149090909090909,
      "grad_norm": 0.32301098108291626,
      "learning_rate": 4.238730749274913e-06,
      "loss": 0.884,
      "mean_token_accuracy": 0.8238483086228371,
      "num_tokens": 48257041.0,
      "step": 6290
    },
    {
      "entropy": 0.9208923012018204,
      "epoch": 0.9163636363636364,
      "grad_norm": 0.3088277578353882,
      "learning_rate": 4.099624105775764e-06,
      "loss": 0.8969,
      "mean_token_accuracy": 0.8195697702467442,
      "num_tokens": 48336009.0,
      "step": 6300
    },
    {
      "entropy": 0.9096413090825081,
      "epoch": 0.9178181818181819,
      "grad_norm": 0.3320503234863281,
      "learning_rate": 3.962847619196496e-06,
      "loss": 0.8889,
      "mean_token_accuracy": 0.8219702959060669,
      "num_tokens": 48413520.0,
      "step": 6310
    },
    {
      "entropy": 0.9034752696752548,
      "epoch": 0.9192727272727272,
      "grad_norm": 0.38523438572883606,
      "learning_rate": 3.82840461115702e-06,
      "loss": 0.8728,
      "mean_token_accuracy": 0.8240837588906288,
      "num_tokens": 48488903.0,
      "step": 6320
    },
    {
      "entropy": 0.8971277795732021,
      "epoch": 0.9207272727272727,
      "grad_norm": 2.067976236343384,
      "learning_rate": 3.6962983466086684e-06,
      "loss": 0.8661,
      "mean_token_accuracy": 0.8253998480737209,
      "num_tokens": 48564337.0,
      "step": 6330
    },
    {
      "entropy": 0.9000528261065484,
      "epoch": 0.9221818181818182,
      "grad_norm": 0.33974602818489075,
      "learning_rate": 3.5665320337548666e-06,
      "loss": 0.8667,
      "mean_token_accuracy": 0.8244885370135308,
      "num_tokens": 48641523.0,
      "step": 6340
    },
    {
      "entropy": 0.9108474776148796,
      "epoch": 0.9236363636363636,
      "grad_norm": 0.3082718253135681,
      "learning_rate": 3.43910882397328e-06,
      "loss": 0.8816,
      "mean_token_accuracy": 0.821860957890749,
      "num_tokens": 48717572.0,
      "step": 6350
    },
    {
      "entropy": 0.91778749153018,
      "epoch": 0.9250909090909091,
      "grad_norm": 0.7094966173171997,
      "learning_rate": 3.314031811739268e-06,
      "loss": 0.9008,
      "mean_token_accuracy": 0.820485782623291,
      "num_tokens": 48794160.0,
      "step": 6360
    },
    {
      "entropy": 0.9082001946866513,
      "epoch": 0.9265454545454546,
      "grad_norm": 0.8361312747001648,
      "learning_rate": 3.1913040345507593e-06,
      "loss": 0.8926,
      "mean_token_accuracy": 0.822116008400917,
      "num_tokens": 48869605.0,
      "step": 6370
    },
    {
      "entropy": 0.9257027067244052,
      "epoch": 0.928,
      "grad_norm": 0.33025625348091125,
      "learning_rate": 3.0709284728544047e-06,
      "loss": 0.9251,
      "mean_token_accuracy": 0.818819022923708,
      "num_tokens": 48944924.0,
      "step": 6380
    },
    {
      "entropy": 0.9270994037389755,
      "epoch": 0.9294545454545454,
      "grad_norm": 1.8861573934555054,
      "learning_rate": 2.9529080499732917e-06,
      "loss": 0.9123,
      "mean_token_accuracy": 0.8182081528007984,
      "num_tokens": 49023270.0,
      "step": 6390
    },
    {
      "entropy": 0.9081657931208611,
      "epoch": 0.9309090909090909,
      "grad_norm": 2.497610330581665,
      "learning_rate": 2.837245632035893e-06,
      "loss": 0.8942,
      "mean_token_accuracy": 0.8210309714078903,
      "num_tokens": 49099764.0,
      "step": 6400
    },
    {
      "entropy": 0.9176856569945813,
      "epoch": 0.9323636363636364,
      "grad_norm": 2.121029853820801,
      "learning_rate": 2.723944027906503e-06,
      "loss": 0.9129,
      "mean_token_accuracy": 0.820872713625431,
      "num_tokens": 49177635.0,
      "step": 6410
    },
    {
      "entropy": 0.9278782814741134,
      "epoch": 0.9338181818181818,
      "grad_norm": 0.3344612121582031,
      "learning_rate": 2.6130059891169797e-06,
      "loss": 0.9445,
      "mean_token_accuracy": 0.8178577579557895,
      "num_tokens": 49257160.0,
      "step": 6420
    },
    {
      "entropy": 0.8968269281089306,
      "epoch": 0.9352727272727273,
      "grad_norm": 0.32175830006599426,
      "learning_rate": 2.504434209799966e-06,
      "loss": 0.88,
      "mean_token_accuracy": 0.8237360425293445,
      "num_tokens": 49333387.0,
      "step": 6430
    },
    {
      "entropy": 0.9058445796370507,
      "epoch": 0.9367272727272727,
      "grad_norm": 0.3254484534263611,
      "learning_rate": 2.3982313266234193e-06,
      "loss": 0.8815,
      "mean_token_accuracy": 0.8237637273967267,
      "num_tokens": 49408486.0,
      "step": 6440
    },
    {
      "entropy": 0.9265412963926792,
      "epoch": 0.9381818181818182,
      "grad_norm": 0.331331342458725,
      "learning_rate": 2.294399918726618e-06,
      "loss": 0.9117,
      "mean_token_accuracy": 0.8172571666538715,
      "num_tokens": 49484948.0,
      "step": 6450
    },
    {
      "entropy": 0.9107875868678093,
      "epoch": 0.9396363636363636,
      "grad_norm": 0.45179134607315063,
      "learning_rate": 2.1929425076575067e-06,
      "loss": 0.8875,
      "mean_token_accuracy": 0.8206865198910236,
      "num_tokens": 49564787.0,
      "step": 6460
    },
    {
      "entropy": 0.9484403192996979,
      "epoch": 0.9410909090909091,
      "grad_norm": 0.32299381494522095,
      "learning_rate": 2.093861557311479e-06,
      "loss": 0.9599,
      "mean_token_accuracy": 0.8149694137275219,
      "num_tokens": 49644908.0,
      "step": 6470
    },
    {
      "entropy": 0.9207295253872871,
      "epoch": 0.9425454545454546,
      "grad_norm": 0.3640280067920685,
      "learning_rate": 1.997159473871544e-06,
      "loss": 0.9052,
      "mean_token_accuracy": 0.8190280869603157,
      "num_tokens": 49723356.0,
      "step": 6480
    },
    {
      "entropy": 0.9006929636001587,
      "epoch": 0.944,
      "grad_norm": 0.31023108959198,
      "learning_rate": 1.9028386057498202e-06,
      "loss": 0.8671,
      "mean_token_accuracy": 0.8262156061828136,
      "num_tokens": 49797029.0,
      "step": 6490
    },
    {
      "entropy": 0.935797930508852,
      "epoch": 0.9454545454545454,
      "grad_norm": 0.33681347966194153,
      "learning_rate": 1.810901243530616e-06,
      "loss": 0.915,
      "mean_token_accuracy": 0.8161994099617005,
      "num_tokens": 49876080.0,
      "step": 6500
    },
    {
      "entropy": 0.9388311050832272,
      "epoch": 0.9469090909090909,
      "grad_norm": 2.5916805267333984,
      "learning_rate": 1.7213496199147198e-06,
      "loss": 0.924,
      "mean_token_accuracy": 0.8182121217250824,
      "num_tokens": 49955757.0,
      "step": 6510
    },
    {
      "entropy": 0.924860893189907,
      "epoch": 0.9483636363636364,
      "grad_norm": 0.3176397383213043,
      "learning_rate": 1.6341859096652162e-06,
      "loss": 0.9189,
      "mean_token_accuracy": 0.8182754583656788,
      "num_tokens": 50034921.0,
      "step": 6520
    },
    {
      "entropy": 0.9215592637658119,
      "epoch": 0.9498181818181818,
      "grad_norm": 0.3312931954860687,
      "learning_rate": 1.5494122295546778e-06,
      "loss": 0.897,
      "mean_token_accuracy": 0.8201039671897888,
      "num_tokens": 50111618.0,
      "step": 6530
    },
    {
      "entropy": 0.9161364257335662,
      "epoch": 0.9512727272727273,
      "grad_norm": 0.3159554600715637,
      "learning_rate": 1.4670306383137089e-06,
      "loss": 0.8933,
      "mean_token_accuracy": 0.8197489641606808,
      "num_tokens": 50188800.0,
      "step": 6540
    },
    {
      "entropy": 0.8958120048046112,
      "epoch": 0.9527272727272728,
      "grad_norm": 0.3147352933883667,
      "learning_rate": 1.38704313658101e-06,
      "loss": 0.8729,
      "mean_token_accuracy": 0.8254260867834091,
      "num_tokens": 50264579.0,
      "step": 6550
    },
    {
      "entropy": 0.903196832537651,
      "epoch": 0.9541818181818181,
      "grad_norm": 0.31137725710868835,
      "learning_rate": 1.3094516668547762e-06,
      "loss": 0.8806,
      "mean_token_accuracy": 0.8244303494691849,
      "num_tokens": 50340059.0,
      "step": 6560
    },
    {
      "entropy": 0.8903951339423657,
      "epoch": 0.9556363636363636,
      "grad_norm": 2.127527952194214,
      "learning_rate": 1.2342581134455008e-06,
      "loss": 0.8679,
      "mean_token_accuracy": 0.8251245476305484,
      "num_tokens": 50415334.0,
      "step": 6570
    },
    {
      "entropy": 0.9220633573830128,
      "epoch": 0.9570909090909091,
      "grad_norm": 0.3190872073173523,
      "learning_rate": 1.161464302430214e-06,
      "loss": 0.9037,
      "mean_token_accuracy": 0.8186452768743038,
      "num_tokens": 50495609.0,
      "step": 6580
    },
    {
      "entropy": 0.9037490114569664,
      "epoch": 0.9585454545454546,
      "grad_norm": 0.35063058137893677,
      "learning_rate": 1.0910720016081627e-06,
      "loss": 0.8781,
      "mean_token_accuracy": 0.8231127128005028,
      "num_tokens": 50571158.0,
      "step": 6590
    },
    {
      "entropy": 0.898887837678194,
      "epoch": 0.96,
      "grad_norm": 0.3321762979030609,
      "learning_rate": 1.0230829204578769e-06,
      "loss": 0.8647,
      "mean_token_accuracy": 0.8250281892716884,
      "num_tokens": 50646197.0,
      "step": 6600
    },
    {
      "entropy": 0.908264522254467,
      "epoch": 0.9614545454545455,
      "grad_norm": 0.3521866500377655,
      "learning_rate": 9.574987100956352e-07,
      "loss": 0.8764,
      "mean_token_accuracy": 0.8238757200539112,
      "num_tokens": 50724334.0,
      "step": 6610
    },
    {
      "entropy": 0.9157738953828811,
      "epoch": 0.9629090909090909,
      "grad_norm": 0.3183552026748657,
      "learning_rate": 8.943209632354054e-07,
      "loss": 0.8934,
      "mean_token_accuracy": 0.8203241616487503,
      "num_tokens": 50803191.0,
      "step": 6620
    },
    {
      "entropy": 0.9221185706555843,
      "epoch": 0.9643636363636363,
      "grad_norm": 0.32127588987350464,
      "learning_rate": 8.335512141500846e-07,
      "loss": 0.8943,
      "mean_token_accuracy": 0.8205978877842426,
      "num_tokens": 50881278.0,
      "step": 6630
    },
    {
      "entropy": 0.91321404799819,
      "epoch": 0.9658181818181818,
      "grad_norm": 0.3270174264907837,
      "learning_rate": 7.751909386343242e-07,
      "loss": 0.8821,
      "mean_token_accuracy": 0.8214927159249783,
      "num_tokens": 50956132.0,
      "step": 6640
    },
    {
      "entropy": 0.9102289408445359,
      "epoch": 0.9672727272727273,
      "grad_norm": 1.5056520700454712,
      "learning_rate": 7.192415539686665e-07,
      "loss": 0.891,
      "mean_token_accuracy": 0.8208358325064182,
      "num_tokens": 51034586.0,
      "step": 6650
    },
    {
      "entropy": 0.8969598315656185,
      "epoch": 0.9687272727272728,
      "grad_norm": 0.30618613958358765,
      "learning_rate": 6.65704418885111e-07,
      "loss": 0.8673,
      "mean_token_accuracy": 0.8258568048477173,
      "num_tokens": 51112025.0,
      "step": 6660
    },
    {
      "entropy": 0.9234667181968689,
      "epoch": 0.9701818181818181,
      "grad_norm": 0.3337618112564087,
      "learning_rate": 6.145808335341243e-07,
      "loss": 0.8963,
      "mean_token_accuracy": 0.8207845896482467,
      "num_tokens": 51189368.0,
      "step": 6670
    },
    {
      "entropy": 0.9242277570068836,
      "epoch": 0.9716363636363636,
      "grad_norm": 0.32344672083854675,
      "learning_rate": 5.658720394530478e-07,
      "loss": 0.8976,
      "mean_token_accuracy": 0.8184162557125092,
      "num_tokens": 51268418.0,
      "step": 6680
    },
    {
      "entropy": 0.9331931956112385,
      "epoch": 0.9730909090909091,
      "grad_norm": 0.36297062039375305,
      "learning_rate": 5.195792195359924e-07,
      "loss": 0.9198,
      "mean_token_accuracy": 0.8186567753553391,
      "num_tokens": 51347561.0,
      "step": 6690
    },
    {
      "entropy": 0.9222750753164292,
      "epoch": 0.9745454545454545,
      "grad_norm": 0.3219527006149292,
      "learning_rate": 4.7570349800507546e-07,
      "loss": 0.8938,
      "mean_token_accuracy": 0.8191801868379116,
      "num_tokens": 51425804.0,
      "step": 6700
    },
    {
      "entropy": 0.904650616645813,
      "epoch": 0.976,
      "grad_norm": 0.3058840334415436,
      "learning_rate": 4.342459403831242e-07,
      "loss": 0.8824,
      "mean_token_accuracy": 0.8226540274918079,
      "num_tokens": 51503722.0,
      "step": 6710
    },
    {
      "entropy": 0.9009366378188133,
      "epoch": 0.9774545454545455,
      "grad_norm": 0.32405465841293335,
      "learning_rate": 3.952075534678079e-07,
      "loss": 0.8749,
      "mean_token_accuracy": 0.8230279020965099,
      "num_tokens": 51578996.0,
      "step": 6720
    },
    {
      "entropy": 0.9061621010303498,
      "epoch": 0.978909090909091,
      "grad_norm": 2.5726563930511475,
      "learning_rate": 3.585892853071928e-07,
      "loss": 0.8912,
      "mean_token_accuracy": 0.8214048929512501,
      "num_tokens": 51657221.0,
      "step": 6730
    },
    {
      "entropy": 0.9052648693323135,
      "epoch": 0.9803636363636363,
      "grad_norm": 0.2956988513469696,
      "learning_rate": 3.2439202517668286e-07,
      "loss": 0.8914,
      "mean_token_accuracy": 0.8212430521845817,
      "num_tokens": 51735276.0,
      "step": 6740
    },
    {
      "entropy": 0.9187554679811001,
      "epoch": 0.9818181818181818,
      "grad_norm": 0.3320491909980774,
      "learning_rate": 2.926166035574923e-07,
      "loss": 0.9299,
      "mean_token_accuracy": 0.8186180919408799,
      "num_tokens": 51813941.0,
      "step": 6750
    },
    {
      "entropy": 0.9081330545246601,
      "epoch": 0.9832727272727273,
      "grad_norm": 0.30943775177001953,
      "learning_rate": 2.632637921163831e-07,
      "loss": 0.8868,
      "mean_token_accuracy": 0.8217815391719341,
      "num_tokens": 51890812.0,
      "step": 6760
    },
    {
      "entropy": 0.8877025507390499,
      "epoch": 0.9847272727272727,
      "grad_norm": 0.3373887240886688,
      "learning_rate": 2.3633430368702262e-07,
      "loss": 0.855,
      "mean_token_accuracy": 0.8284515894949436,
      "num_tokens": 51962819.0,
      "step": 6770
    },
    {
      "entropy": 0.9245131060481071,
      "epoch": 0.9861818181818182,
      "grad_norm": 1.2123104333877563,
      "learning_rate": 2.118287922526059e-07,
      "loss": 0.8985,
      "mean_token_accuracy": 0.8212319493293763,
      "num_tokens": 52041518.0,
      "step": 6780
    },
    {
      "entropy": 0.9151236355304718,
      "epoch": 0.9876363636363636,
      "grad_norm": 0.3329769968986511,
      "learning_rate": 1.8974785292999898e-07,
      "loss": 0.8917,
      "mean_token_accuracy": 0.8217348761856555,
      "num_tokens": 52120146.0,
      "step": 6790
    },
    {
      "entropy": 0.9201492570340634,
      "epoch": 0.9890909090909091,
      "grad_norm": 0.334018349647522,
      "learning_rate": 1.7009202195529057e-07,
      "loss": 0.9226,
      "mean_token_accuracy": 0.8178925760090351,
      "num_tokens": 52198078.0,
      "step": 6800
    },
    {
      "entropy": 0.9055890217423439,
      "epoch": 0.9905454545454545,
      "grad_norm": 0.8450275659561157,
      "learning_rate": 1.5286177667074262e-07,
      "loss": 0.8797,
      "mean_token_accuracy": 0.8226368710398674,
      "num_tokens": 52273502.0,
      "step": 6810
    },
    {
      "entropy": 0.9188882060348987,
      "epoch": 0.992,
      "grad_norm": 0.3331072926521301,
      "learning_rate": 1.3805753551323864e-07,
      "loss": 0.9063,
      "mean_token_accuracy": 0.8196715794503688,
      "num_tokens": 52350145.0,
      "step": 6820
    },
    {
      "entropy": 0.9134249776601792,
      "epoch": 0.9934545454545455,
      "grad_norm": 0.3241958022117615,
      "learning_rate": 1.2567965800407473e-07,
      "loss": 0.8883,
      "mean_token_accuracy": 0.8222455978393555,
      "num_tokens": 52429758.0,
      "step": 6830
    },
    {
      "entropy": 0.8976367473602295,
      "epoch": 0.9949090909090909,
      "grad_norm": 1.5700055360794067,
      "learning_rate": 1.157284447402819e-07,
      "loss": 0.8828,
      "mean_token_accuracy": 0.8237587973475456,
      "num_tokens": 52504858.0,
      "step": 6840
    },
    {
      "entropy": 0.9396997675299644,
      "epoch": 0.9963636363636363,
      "grad_norm": 0.32852768898010254,
      "learning_rate": 1.0820413738729128e-07,
      "loss": 0.9255,
      "mean_token_accuracy": 0.8178542837500572,
      "num_tokens": 52582462.0,
      "step": 6850
    },
    {
      "entropy": 0.904629698395729,
      "epoch": 0.9978181818181818,
      "grad_norm": 0.6703066229820251,
      "learning_rate": 1.031069186730529e-07,
      "loss": 0.8881,
      "mean_token_accuracy": 0.8228366106748581,
      "num_tokens": 52658068.0,
      "step": 6860
    },
    {
      "entropy": 0.9132826693356038,
      "epoch": 0.9992727272727273,
      "grad_norm": 0.3042033910751343,
      "learning_rate": 1.0043691238364123e-07,
      "loss": 0.8982,
      "mean_token_accuracy": 0.8215601086616516,
      "num_tokens": 52733067.0,
      "step": 6870
    }
  ],
  "logging_steps": 10,
  "max_steps": 6875,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 1000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 5.34572363346346e+18,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}