{
  "best_global_step": 6000,
  "best_metric": 0.06929752975702286,
  "best_model_checkpoint": "runs/iter3_2epoch_r256_aggressive/checkpoint-6000",
  "epoch": 1.5671183673469389,
  "eval_steps": 1000,
  "global_step": 6000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "entropy": 0.9600491151213646,
      "epoch": 0.00026122448979591835,
      "grad_norm": 4.874749183654785,
      "learning_rate": 0.0,
      "loss": 1.5426,
      "mean_token_accuracy": 0.5253331288695335,
      "num_tokens": 92658.0,
      "step": 1
    },
    {
      "entropy": 1.0668207285355549,
      "epoch": 0.013061224489795919,
      "grad_norm": 0.3741045296192169,
      "learning_rate": 3.181818181818182e-05,
      "loss": 0.8201,
      "mean_token_accuracy": 0.713900758295643,
      "num_tokens": 4695822.0,
      "step": 50
    },
    {
      "entropy": 0.7840494087338448,
      "epoch": 0.026122448979591838,
      "grad_norm": 1.6816201210021973,
      "learning_rate": 6.428571428571429e-05,
      "loss": 0.5652,
      "mean_token_accuracy": 0.8121247433125973,
      "num_tokens": 9384030.0,
      "step": 100
    },
    {
      "entropy": 0.6429135516285897,
      "epoch": 0.03918367346938775,
      "grad_norm": 1.8123995065689087,
      "learning_rate": 9.675324675324677e-05,
      "loss": 0.4768,
      "mean_token_accuracy": 0.8343716338276863,
      "num_tokens": 14068659.0,
      "step": 150
    },
    {
      "entropy": 0.6180185662209987,
      "epoch": 0.052244897959183675,
      "grad_norm": 0.228230819106102,
      "learning_rate": 9.99911270857259e-05,
      "loss": 0.4195,
      "mean_token_accuracy": 0.8369381006062031,
      "num_tokens": 18749150.0,
      "step": 200
    },
    {
      "entropy": 0.6392524507641792,
      "epoch": 0.0653061224489796,
      "grad_norm": 0.3391082286834717,
      "learning_rate": 9.99604593263008e-05,
      "loss": 0.3747,
      "mean_token_accuracy": 0.8521625240147114,
      "num_tokens": 23444738.0,
      "step": 250
    },
    {
      "entropy": 0.6395396235585212,
      "epoch": 0.0783673469387755,
      "grad_norm": 0.35865047574043274,
      "learning_rate": 9.990790061392936e-05,
      "loss": 0.4622,
      "mean_token_accuracy": 0.817176415771246,
      "num_tokens": 28146459.0,
      "step": 300
    },
    {
      "entropy": 0.6052160003781318,
      "epoch": 0.09142857142857143,
      "grad_norm": 0.461127907037735,
      "learning_rate": 9.983347397802953e-05,
      "loss": 0.3964,
      "mean_token_accuracy": 0.865231085717678,
      "num_tokens": 32829584.0,
      "step": 350
    },
    {
      "entropy": 0.6162980204820633,
      "epoch": 0.10448979591836735,
      "grad_norm": 0.6675931811332703,
      "learning_rate": 9.973721202979037e-05,
      "loss": 0.4523,
      "mean_token_accuracy": 0.8371678332984448,
      "num_tokens": 37519736.0,
      "step": 400
    },
    {
      "entropy": 0.6373377884924412,
      "epoch": 0.11755102040816326,
      "grad_norm": 0.7843639254570007,
      "learning_rate": 9.961915694788294e-05,
      "loss": 0.4246,
      "mean_token_accuracy": 0.8822347274422646,
      "num_tokens": 42214834.0,
      "step": 450
    },
    {
      "entropy": 0.6182832332700491,
      "epoch": 0.1306122448979592,
      "grad_norm": 0.12627503275871277,
      "learning_rate": 9.9479360459979e-05,
      "loss": 0.3963,
      "mean_token_accuracy": 0.8835959336161614,
      "num_tokens": 46896384.0,
      "step": 500
    },
    {
      "entropy": 0.604607280343771,
      "epoch": 0.1436734693877551,
      "grad_norm": 0.33763670921325684,
      "learning_rate": 9.931788382008588e-05,
      "loss": 0.3126,
      "mean_token_accuracy": 0.8891353133320808,
      "num_tokens": 51611949.0,
      "step": 550
    },
    {
      "entropy": 0.6003445901721716,
      "epoch": 0.156734693877551,
      "grad_norm": 0.4399724304676056,
      "learning_rate": 9.913479778170688e-05,
      "loss": 0.2951,
      "mean_token_accuracy": 0.89730602607131,
      "num_tokens": 56295022.0,
      "step": 600
    },
    {
      "entropy": 0.6263976456224918,
      "epoch": 0.16979591836734695,
      "grad_norm": 0.24783462285995483,
      "learning_rate": 9.893018256683967e-05,
      "loss": 0.345,
      "mean_token_accuracy": 0.8859454789757728,
      "num_tokens": 60989496.0,
      "step": 650
    },
    {
      "entropy": 0.6047938592731953,
      "epoch": 0.18285714285714286,
      "grad_norm": 0.6106021404266357,
      "learning_rate": 9.870412783082573e-05,
      "loss": 0.4187,
      "mean_token_accuracy": 0.8859497668594122,
      "num_tokens": 65706938.0,
      "step": 700
    },
    {
      "entropy": 0.6086559303104877,
      "epoch": 0.19591836734693877,
      "grad_norm": 0.6276345252990723,
      "learning_rate": 9.845673262306643e-05,
      "loss": 0.2703,
      "mean_token_accuracy": 0.9083889089524746,
      "num_tokens": 70440218.0,
      "step": 750
    },
    {
      "entropy": 0.5998903425782919,
      "epoch": 0.2089795918367347,
      "grad_norm": 0.46106088161468506,
      "learning_rate": 9.818810534362306e-05,
      "loss": 0.3502,
      "mean_token_accuracy": 0.8987785133719445,
      "num_tokens": 75125260.0,
      "step": 800
    },
    {
      "entropy": 0.5996488413214683,
      "epoch": 0.2220408163265306,
      "grad_norm": 0.2753312289714813,
      "learning_rate": 9.789836369571958e-05,
      "loss": 0.3382,
      "mean_token_accuracy": 0.9013783724606037,
      "num_tokens": 79822058.0,
      "step": 850
    },
    {
      "entropy": 0.6070075839012862,
      "epoch": 0.23510204081632652,
      "grad_norm": 0.44265103340148926,
      "learning_rate": 9.758763463416922e-05,
      "loss": 0.3303,
      "mean_token_accuracy": 0.9064150702953339,
      "num_tokens": 84511616.0,
      "step": 900
    },
    {
      "entropy": 0.6011689330637455,
      "epoch": 0.24816326530612245,
      "grad_norm": 0.2258819341659546,
      "learning_rate": 9.725605430974712e-05,
      "loss": 0.281,
      "mean_token_accuracy": 0.9128213880956173,
      "num_tokens": 89218784.0,
      "step": 950
    },
    {
      "entropy": 0.610101203173399,
      "epoch": 0.2612244897959184,
      "grad_norm": 0.5324413180351257,
      "learning_rate": 9.690376800953389e-05,
      "loss": 0.2765,
      "mean_token_accuracy": 0.9096981877088547,
      "num_tokens": 93934733.0,
      "step": 1000
    },
    {
      "epoch": 0.2612244897959184,
      "eval_entropy": 0.6072796992301941,
      "eval_loss": 0.22717319428920746,
      "eval_mean_token_accuracy": 0.9097091863632202,
      "eval_num_tokens": 93934733.0,
      "eval_runtime": 105.7447,
      "eval_samples_per_second": 189.135,
      "eval_steps_per_second": 5.91,
      "step": 1000
    },
    {
      "entropy": 0.6069025094807148,
      "epoch": 0.2742857142857143,
      "grad_norm": 0.32851845026016235,
      "learning_rate": 9.65309300932557e-05,
      "loss": 0.3301,
      "mean_token_accuracy": 0.9123546965420246,
      "num_tokens": 98646726.0,
      "step": 1050
    },
    {
      "entropy": 0.5923496520519257,
      "epoch": 0.2873469387755102,
      "grad_norm": 0.22054198384284973,
      "learning_rate": 9.61377039256492e-05,
      "loss": 0.2866,
      "mean_token_accuracy": 0.9174483200907707,
      "num_tokens": 103376112.0,
      "step": 1100
    },
    {
      "entropy": 0.5921645154803992,
      "epoch": 0.3004081632653061,
      "grad_norm": 0.8782572150230408,
      "learning_rate": 9.572426180488079e-05,
      "loss": 0.312,
      "mean_token_accuracy": 0.9164931161701679,
      "num_tokens": 108072462.0,
      "step": 1150
    },
    {
      "entropy": 0.6070084515213966,
      "epoch": 0.313469387755102,
      "grad_norm": 0.4552319049835205,
      "learning_rate": 9.529078488705135e-05,
      "loss": 0.3169,
      "mean_token_accuracy": 0.9169696615636349,
      "num_tokens": 112781107.0,
      "step": 1200
    },
    {
      "entropy": 0.5912807537615299,
      "epoch": 0.32653061224489793,
      "grad_norm": 0.3097328841686249,
      "learning_rate": 9.483746310682004e-05,
      "loss": 0.2641,
      "mean_token_accuracy": 0.9121362474560738,
      "num_tokens": 117459347.0,
      "step": 1250
    },
    {
      "entropy": 0.5872646420449018,
      "epoch": 0.3395918367346939,
      "grad_norm": 0.7857140898704529,
      "learning_rate": 9.43644950941813e-05,
      "loss": 0.3741,
      "mean_token_accuracy": 0.9139573207497597,
      "num_tokens": 122161796.0,
      "step": 1300
    },
    {
      "entropy": 0.5643932522833347,
      "epoch": 0.3526530612244898,
      "grad_norm": 0.4844133257865906,
      "learning_rate": 9.387208808743222e-05,
      "loss": 0.2447,
      "mean_token_accuracy": 0.9241362990438938,
      "num_tokens": 126862525.0,
      "step": 1350
    },
    {
      "entropy": 0.5853312531858683,
      "epoch": 0.3657142857142857,
      "grad_norm": 0.7309156656265259,
      "learning_rate": 9.336045784236764e-05,
      "loss": 0.2758,
      "mean_token_accuracy": 0.9223621267080307,
      "num_tokens": 131562166.0,
      "step": 1400
    },
    {
      "entropy": 0.5751851069927215,
      "epoch": 0.3787755102040816,
      "grad_norm": 0.6382029056549072,
      "learning_rate": 9.28298285377436e-05,
      "loss": 0.3572,
      "mean_token_accuracy": 0.9173358491063118,
      "num_tokens": 136280998.0,
      "step": 1450
    },
    {
      "entropy": 0.5755070626735688,
      "epoch": 0.39183673469387753,
      "grad_norm": 0.4856666326522827,
      "learning_rate": 9.228043267704974e-05,
      "loss": 0.2431,
      "mean_token_accuracy": 0.9318676285445691,
      "num_tokens": 140963483.0,
      "step": 1500
    },
    {
      "entropy": 0.5901833316683769,
      "epoch": 0.4048979591836735,
      "grad_norm": 0.23288387060165405,
      "learning_rate": 9.171251098663445e-05,
      "loss": 0.3199,
      "mean_token_accuracy": 0.9218292973935605,
      "num_tokens": 145663835.0,
      "step": 1550
    },
    {
      "entropy": 0.6068822190165519,
      "epoch": 0.4179591836734694,
      "grad_norm": 0.7570965886116028,
      "learning_rate": 9.112631231022669e-05,
      "loss": 0.2503,
      "mean_token_accuracy": 0.9219834499061108,
      "num_tokens": 150388643.0,
      "step": 1600
    },
    {
      "entropy": 0.6086373251676559,
      "epoch": 0.4310204081632653,
      "grad_norm": 0.40420103073120117,
      "learning_rate": 9.052209349990144e-05,
      "loss": 0.2979,
      "mean_token_accuracy": 0.9300994580984115,
      "num_tokens": 155072117.0,
      "step": 1650
    },
    {
      "entropy": 0.5892787943035365,
      "epoch": 0.4440816326530612,
      "grad_norm": 0.28975844383239746,
      "learning_rate": 8.990011930353579e-05,
      "loss": 0.2509,
      "mean_token_accuracy": 0.9298394879698754,
      "num_tokens": 159761671.0,
      "step": 1700
    },
    {
      "entropy": 0.5920831595361232,
      "epoch": 0.45714285714285713,
      "grad_norm": 0.3639383018016815,
      "learning_rate": 8.926066224880565e-05,
      "loss": 0.2603,
      "mean_token_accuracy": 0.9342290599644184,
      "num_tokens": 164475669.0,
      "step": 1750
    },
    {
      "entropy": 0.5960557947307825,
      "epoch": 0.47020408163265304,
      "grad_norm": 0.7768979668617249,
      "learning_rate": 8.86040025237734e-05,
      "loss": 0.3215,
      "mean_token_accuracy": 0.9318660141527653,
      "num_tokens": 169186115.0,
      "step": 1800
    },
    {
      "entropy": 0.596306427642703,
      "epoch": 0.483265306122449,
      "grad_norm": 0.2945772409439087,
      "learning_rate": 8.793042785411927e-05,
      "loss": 0.2908,
      "mean_token_accuracy": 0.9278665748238564,
      "num_tokens": 173898142.0,
      "step": 1850
    },
    {
      "entropy": 0.5824826972186565,
      "epoch": 0.4963265306122449,
      "grad_norm": 0.629438042640686,
      "learning_rate": 8.724023337706972e-05,
      "loss": 0.2442,
      "mean_token_accuracy": 0.9366743673384189,
      "num_tokens": 178579887.0,
      "step": 1900
    },
    {
      "entropy": 0.6178072960674763,
      "epoch": 0.5093877551020408,
      "grad_norm": 0.2474866807460785,
      "learning_rate": 8.653372151207855e-05,
      "loss": 0.2979,
      "mean_token_accuracy": 0.9314125381410122,
      "num_tokens": 183276089.0,
      "step": 1950
    },
    {
      "entropy": 0.6088858918845653,
      "epoch": 0.5224489795918368,
      "grad_norm": 0.4756447970867157,
      "learning_rate": 8.58112018283171e-05,
      "loss": 0.3085,
      "mean_token_accuracy": 0.933091163188219,
      "num_tokens": 187979857.0,
      "step": 2000
    },
    {
      "epoch": 0.5224489795918368,
      "eval_entropy": 0.5824218388557434,
      "eval_loss": 0.1687791496515274,
      "eval_mean_token_accuracy": 0.9382246346473694,
      "eval_num_tokens": 187979857.0,
      "eval_runtime": 105.81,
      "eval_samples_per_second": 189.018,
      "eval_steps_per_second": 5.907,
      "step": 2000
    },
    {
      "entropy": 0.6008229818940163,
      "epoch": 0.5355102040816326,
      "grad_norm": 0.968726634979248,
      "learning_rate": 8.507299090903164e-05,
      "loss": 0.217,
      "mean_token_accuracy": 0.9400185799598694,
      "num_tokens": 192675403.0,
      "step": 2050
    },
    {
      "entropy": 0.5939074163138867,
      "epoch": 0.5485714285714286,
      "grad_norm": 0.33287671208381653,
      "learning_rate": 8.43194122128276e-05,
      "loss": 0.2473,
      "mean_token_accuracy": 0.9392171530425548,
      "num_tokens": 197339702.0,
      "step": 2100
    },
    {
      "entropy": 0.5947689298540354,
      "epoch": 0.5616326530612245,
      "grad_norm": 0.5438774228096008,
      "learning_rate": 8.355079593194101e-05,
      "loss": 0.2401,
      "mean_token_accuracy": 0.9419188013672829,
      "num_tokens": 202042351.0,
      "step": 2150
    },
    {
      "entropy": 0.6115065851807594,
      "epoch": 0.5746938775510204,
      "grad_norm": 0.3915332555770874,
      "learning_rate": 8.276747884755978e-05,
      "loss": 0.2958,
      "mean_token_accuracy": 0.9387796081602573,
      "num_tokens": 206746259.0,
      "step": 2200
    },
    {
      "entropy": 0.5970733669400216,
      "epoch": 0.5877551020408164,
      "grad_norm": 0.7760431170463562,
      "learning_rate": 8.196980418225764e-05,
      "loss": 0.2355,
      "mean_token_accuracy": 0.94012371301651,
      "num_tokens": 211462162.0,
      "step": 2250
    },
    {
      "entropy": 0.5940753533691168,
      "epoch": 0.6008163265306122,
      "grad_norm": 0.2869822382926941,
      "learning_rate": 8.115812144960594e-05,
      "loss": 0.2469,
      "mean_token_accuracy": 0.9416494782269001,
      "num_tokens": 216190557.0,
      "step": 2300
    },
    {
      "entropy": 0.60329102024436,
      "epoch": 0.6138775510204082,
      "grad_norm": 0.6603459715843201,
      "learning_rate": 8.033278630102877e-05,
      "loss": 0.2189,
      "mean_token_accuracy": 0.9444279509782791,
      "num_tokens": 220880056.0,
      "step": 2350
    },
    {
      "entropy": 0.6052355971932412,
      "epoch": 0.626938775510204,
      "grad_norm": 0.3574765622615814,
      "learning_rate": 7.949416036996879e-05,
      "loss": 0.2728,
      "mean_token_accuracy": 0.9418448767066002,
      "num_tokens": 225589999.0,
      "step": 2400
    },
    {
      "entropy": 0.5736296055465937,
      "epoch": 0.64,
      "grad_norm": 0.31953155994415283,
      "learning_rate": 7.864261111343187e-05,
      "loss": 0.2291,
      "mean_token_accuracy": 0.9489433655142784,
      "num_tokens": 230305382.0,
      "step": 2450
    },
    {
      "entropy": 0.602155731767416,
      "epoch": 0.6530612244897959,
      "grad_norm": 0.6838288903236389,
      "learning_rate": 7.777851165098012e-05,
      "loss": 0.1958,
      "mean_token_accuracy": 0.946748715788126,
      "num_tokens": 235011703.0,
      "step": 2500
    },
    {
      "entropy": 0.5958992087841034,
      "epoch": 0.6661224489795918,
      "grad_norm": 0.22716443240642548,
      "learning_rate": 7.690224060124371e-05,
      "loss": 0.2536,
      "mean_token_accuracy": 0.9489493478089571,
      "num_tokens": 239706919.0,
      "step": 2550
    },
    {
      "entropy": 0.6019650375843049,
      "epoch": 0.6791836734693878,
      "grad_norm": 0.3493569493293762,
      "learning_rate": 7.601418191602328e-05,
      "loss": 0.2586,
      "mean_token_accuracy": 0.9482229042053223,
      "num_tokens": 244418186.0,
      "step": 2600
    },
    {
      "entropy": 0.586955601349473,
      "epoch": 0.6922448979591836,
      "grad_norm": 0.25364527106285095,
      "learning_rate": 7.511472471205541e-05,
      "loss": 0.2511,
      "mean_token_accuracy": 0.9499958890676499,
      "num_tokens": 249105699.0,
      "step": 2650
    },
    {
      "entropy": 0.57311468757689,
      "epoch": 0.7053061224489796,
      "grad_norm": 0.854145884513855,
      "learning_rate": 7.42042631005151e-05,
      "loss": 0.2548,
      "mean_token_accuracy": 0.948536623865366,
      "num_tokens": 253800051.0,
      "step": 2700
    },
    {
      "entropy": 0.5846149046719075,
      "epoch": 0.7183673469387755,
      "grad_norm": 0.7333242297172546,
      "learning_rate": 7.328319601432987e-05,
      "loss": 0.2424,
      "mean_token_accuracy": 0.951610994040966,
      "num_tokens": 258490956.0,
      "step": 2750
    },
    {
      "entropy": 0.5881183807551861,
      "epoch": 0.7314285714285714,
      "grad_norm": 0.34874463081359863,
      "learning_rate": 7.2351927033381e-05,
      "loss": 0.3103,
      "mean_token_accuracy": 0.9474381080269814,
      "num_tokens": 263183558.0,
      "step": 2800
    },
    {
      "entropy": 0.6000971308350563,
      "epoch": 0.7444897959183674,
      "grad_norm": 0.40719911456108093,
      "learning_rate": 7.141086420766875e-05,
      "loss": 0.2412,
      "mean_token_accuracy": 0.9557790002226829,
      "num_tokens": 267896256.0,
      "step": 2850
    },
    {
      "entropy": 0.6206524230539798,
      "epoch": 0.7575510204081632,
      "grad_norm": 0.3394398093223572,
      "learning_rate": 7.046041987851896e-05,
      "loss": 0.2431,
      "mean_token_accuracy": 0.9514926019310951,
      "num_tokens": 272576019.0,
      "step": 2900
    },
    {
      "entropy": 0.5987586285173893,
      "epoch": 0.7706122448979592,
      "grad_norm": 0.435951828956604,
      "learning_rate": 6.950101049790922e-05,
      "loss": 0.2353,
      "mean_token_accuracy": 0.9547148038446903,
      "num_tokens": 277288513.0,
      "step": 2950
    },
    {
      "entropy": 0.5921910824626684,
      "epoch": 0.7836734693877551,
      "grad_norm": 0.39050304889678955,
      "learning_rate": 6.853305644599397e-05,
      "loss": 0.2647,
      "mean_token_accuracy": 0.9546147619187831,
      "num_tokens": 281988009.0,
      "step": 3000
    },
    {
      "epoch": 0.7836734693877551,
      "eval_entropy": 0.5917929359436035,
      "eval_loss": 0.13952013850212097,
      "eval_mean_token_accuracy": 0.9494351725578308,
      "eval_num_tokens": 281988009.0,
      "eval_runtime": 105.8253,
      "eval_samples_per_second": 188.991,
      "eval_steps_per_second": 5.906,
      "step": 3000
    },
    {
      "entropy": 0.5995206294953823,
      "epoch": 0.796734693877551,
      "grad_norm": 0.5683656334877014,
      "learning_rate": 6.755698184690837e-05,
      "loss": 0.2358,
      "mean_token_accuracy": 0.954430370926857,
      "num_tokens": 286687009.0,
      "step": 3050
    },
    {
      "entropy": 0.6226249255239964,
      "epoch": 0.809795918367347,
      "grad_norm": 0.17246268689632416,
      "learning_rate": 6.657321438293174e-05,
      "loss": 0.2491,
      "mean_token_accuracy": 0.9509413394331933,
      "num_tokens": 291376119.0,
      "step": 3100
    },
    {
      "entropy": 0.5907960905134678,
      "epoch": 0.8228571428571428,
      "grad_norm": 0.14905746281147003,
      "learning_rate": 6.558218510709192e-05,
      "loss": 0.1754,
      "mean_token_accuracy": 0.9620982597768307,
      "num_tokens": 296068600.0,
      "step": 3150
    },
    {
      "entropy": 0.6032850205898285,
      "epoch": 0.8359183673469388,
      "grad_norm": 0.7370997071266174,
      "learning_rate": 6.458432825429263e-05,
      "loss": 0.2096,
      "mean_token_accuracy": 0.9580010569095612,
      "num_tokens": 300766641.0,
      "step": 3200
    },
    {
      "entropy": 0.6270367397367954,
      "epoch": 0.8489795918367347,
      "grad_norm": 0.25541967153549194,
      "learning_rate": 6.358008105104673e-05,
      "loss": 0.2168,
      "mean_token_accuracy": 0.9566573745012283,
      "num_tokens": 305487589.0,
      "step": 3250
    },
    {
      "entropy": 0.5874304908514023,
      "epoch": 0.8620408163265306,
      "grad_norm": 0.6642680764198303,
      "learning_rate": 6.256988352389859e-05,
      "loss": 0.2361,
      "mean_token_accuracy": 0.9588885164260864,
      "num_tokens": 310194698.0,
      "step": 3300
    },
    {
      "entropy": 0.6110400749742985,
      "epoch": 0.8751020408163265,
      "grad_norm": 0.26045140624046326,
      "learning_rate": 6.155417830661963e-05,
      "loss": 0.1988,
      "mean_token_accuracy": 0.9600324124097824,
      "num_tokens": 314871495.0,
      "step": 3350
    },
    {
      "entropy": 0.5896776075661182,
      "epoch": 0.8881632653061224,
      "grad_norm": 0.6913824677467346,
      "learning_rate": 6.0533410446261415e-05,
      "loss": 0.2404,
      "mean_token_accuracy": 0.9585775223374366,
      "num_tokens": 319562203.0,
      "step": 3400
    },
    {
      "entropy": 0.6021107806265354,
      "epoch": 0.9012244897959184,
      "grad_norm": 1.0586336851119995,
      "learning_rate": 5.950802720815143e-05,
      "loss": 0.242,
      "mean_token_accuracy": 0.9604564820230007,
      "num_tokens": 324272631.0,
      "step": 3450
    },
    {
      "entropy": 0.6005905717611313,
      "epoch": 0.9142857142857143,
      "grad_norm": 0.33950117230415344,
      "learning_rate": 5.8478477879916714e-05,
      "loss": 0.2179,
      "mean_token_accuracy": 0.9604296152293682,
      "num_tokens": 328972230.0,
      "step": 3500
    },
    {
      "entropy": 0.6018191055953502,
      "epoch": 0.9273469387755102,
      "grad_norm": 0.34331461787223816,
      "learning_rate": 5.744521357462156e-05,
      "loss": 0.2278,
      "mean_token_accuracy": 0.9604898960888386,
      "num_tokens": 333680964.0,
      "step": 3550
    },
    {
      "entropy": 0.6024402364343405,
      "epoch": 0.9404081632653061,
      "grad_norm": 0.14388936758041382,
      "learning_rate": 5.64086870331053e-05,
      "loss": 0.227,
      "mean_token_accuracy": 0.963430253714323,
      "num_tokens": 338368553.0,
      "step": 3600
    },
    {
      "entropy": 0.5977859281003475,
      "epoch": 0.953469387755102,
      "grad_norm": 0.7793987989425659,
      "learning_rate": 5.5369352425606816e-05,
      "loss": 0.1802,
      "mean_token_accuracy": 0.9638005784153938,
      "num_tokens": 343050682.0,
      "step": 3650
    },
    {
      "entropy": 0.5996729772537946,
      "epoch": 0.966530612244898,
      "grad_norm": 1.3826709985733032,
      "learning_rate": 5.4327665152762876e-05,
      "loss": 0.2326,
      "mean_token_accuracy": 0.9609452857077122,
      "num_tokens": 347754882.0,
      "step": 3700
    },
    {
      "entropy": 0.5990429550409317,
      "epoch": 0.9795918367346939,
      "grad_norm": 0.921048641204834,
      "learning_rate": 5.3284081646067105e-05,
      "loss": 0.2117,
      "mean_token_accuracy": 0.9624172969162464,
      "num_tokens": 352466908.0,
      "step": 3750
    },
    {
      "entropy": 0.5920401379466057,
      "epoch": 0.9926530612244898,
      "grad_norm": 0.9255926012992859,
      "learning_rate": 5.223905916787757e-05,
      "loss": 0.2149,
      "mean_token_accuracy": 0.9625219763815402,
      "num_tokens": 357183033.0,
      "step": 3800
    },
    {
      "entropy": 0.5910900549124215,
      "epoch": 1.0054857142857143,
      "grad_norm": 0.31576788425445557,
      "learning_rate": 5.119305561106008e-05,
      "loss": 0.1889,
      "mean_token_accuracy": 0.9667121869613803,
      "num_tokens": 361822524.0,
      "step": 3850
    },
    {
      "entropy": 0.5463350272923708,
      "epoch": 1.0185469387755102,
      "grad_norm": 0.15527017414569855,
      "learning_rate": 5.0146529298355305e-05,
      "loss": 0.1567,
      "mean_token_accuracy": 0.9730980744957924,
      "num_tokens": 366524404.0,
      "step": 3900
    },
    {
      "entropy": 0.5311270990222693,
      "epoch": 1.031608163265306,
      "grad_norm": 0.7653639316558838,
      "learning_rate": 4.909993878155755e-05,
      "loss": 0.1941,
      "mean_token_accuracy": 0.9723508515954018,
      "num_tokens": 371187218.0,
      "step": 3950
    },
    {
      "entropy": 0.5343203826993704,
      "epoch": 1.044669387755102,
      "grad_norm": 0.3685067594051361,
      "learning_rate": 4.805374264059317e-05,
      "loss": 0.1643,
      "mean_token_accuracy": 0.9716256402432919,
      "num_tokens": 375891451.0,
      "step": 4000
    },
    {
      "epoch": 1.044669387755102,
      "eval_entropy": 0.5199984492301941,
      "eval_loss": 0.10260085761547089,
      "eval_mean_token_accuracy": 0.9630540296554565,
      "eval_num_tokens": 375891451.0,
      "eval_runtime": 105.9615,
      "eval_samples_per_second": 188.748,
      "eval_steps_per_second": 5.898,
      "step": 4000
    },
    {
      "entropy": 0.5307158456742763,
      "epoch": 1.057730612244898,
      "grad_norm": 0.7387518882751465,
      "learning_rate": 4.700839928258657e-05,
      "loss": 0.1336,
      "mean_token_accuracy": 0.9701156315207481,
      "num_tokens": 380573790.0,
      "step": 4050
    },
    {
      "entropy": 0.5245926366001368,
      "epoch": 1.0707918367346938,
      "grad_norm": 0.5187281370162964,
      "learning_rate": 4.596436674100202e-05,
      "loss": 0.1243,
      "mean_token_accuracy": 0.9722427415847779,
      "num_tokens": 385276643.0,
      "step": 4100
    },
    {
      "entropy": 0.525925776809454,
      "epoch": 1.0838530612244899,
      "grad_norm": 0.11235808581113815,
      "learning_rate": 4.492210247494922e-05,
      "loss": 0.1594,
      "mean_token_accuracy": 0.974123645722866,
      "num_tokens": 390015359.0,
      "step": 4150
    },
    {
      "entropy": 0.5073333897441625,
      "epoch": 1.0969142857142857,
      "grad_norm": 0.3686864376068115,
      "learning_rate": 4.3882063168740405e-05,
      "loss": 0.1789,
      "mean_token_accuracy": 0.9707419118285179,
      "num_tokens": 394707917.0,
      "step": 4200
    },
    {
      "entropy": 0.5056077273190022,
      "epoch": 1.1099755102040816,
      "grad_norm": 0.7848389744758606,
      "learning_rate": 4.284470453178698e-05,
      "loss": 0.166,
      "mean_token_accuracy": 0.9751219192147255,
      "num_tokens": 399431065.0,
      "step": 4250
    },
    {
      "entropy": 0.5062515323609114,
      "epoch": 1.1230367346938777,
      "grad_norm": 0.2769508361816406,
      "learning_rate": 4.181048109892344e-05,
      "loss": 0.1537,
      "mean_token_accuracy": 0.9724229994416237,
      "num_tokens": 404132094.0,
      "step": 4300
    },
    {
      "entropy": 0.5368214490264654,
      "epoch": 1.1360979591836735,
      "grad_norm": 1.615638256072998,
      "learning_rate": 4.077984603124577e-05,
      "loss": 0.2671,
      "mean_token_accuracy": 0.9685475620627403,
      "num_tokens": 408819783.0,
      "step": 4350
    },
    {
      "entropy": 0.5090609015524388,
      "epoch": 1.1491591836734694,
      "grad_norm": 0.13154712319374084,
      "learning_rate": 3.975325091755189e-05,
      "loss": 0.2047,
      "mean_token_accuracy": 0.9727970556914807,
      "num_tokens": 413512051.0,
      "step": 4400
    },
    {
      "entropy": 0.5156180077046156,
      "epoch": 1.1622204081632652,
      "grad_norm": 0.33791911602020264,
      "learning_rate": 3.873114557647105e-05,
      "loss": 0.2171,
      "mean_token_accuracy": 0.9713401012122631,
      "num_tokens": 418236972.0,
      "step": 4450
    },
    {
      "entropy": 0.5125152183324099,
      "epoch": 1.1752816326530613,
      "grad_norm": 0.5354299545288086,
      "learning_rate": 3.771397785936868e-05,
      "loss": 0.1564,
      "mean_token_accuracy": 0.9742588277161122,
      "num_tokens": 422939153.0,
      "step": 4500
    },
    {
      "entropy": 0.5200927007943392,
      "epoch": 1.1883428571428571,
      "grad_norm": 0.7117024064064026,
      "learning_rate": 3.67021934541134e-05,
      "loss": 0.1567,
      "mean_token_accuracy": 0.9759668205678463,
      "num_tokens": 427638623.0,
      "step": 4550
    },
    {
      "entropy": 0.5100678735226393,
      "epoch": 1.201404081632653,
      "grad_norm": 0.7776734828948975,
      "learning_rate": 3.5696235689792e-05,
      "loss": 0.1744,
      "mean_token_accuracy": 0.9728824967145919,
      "num_tokens": 432324424.0,
      "step": 4600
    },
    {
      "entropy": 0.5009938656538725,
      "epoch": 1.214465306122449,
      "grad_norm": 0.45003074407577515,
      "learning_rate": 3.469654534245781e-05,
      "loss": 0.1523,
      "mean_token_accuracy": 0.975490598231554,
      "num_tokens": 437053744.0,
      "step": 4650
    },
    {
      "entropy": 0.5198498579859734,
      "epoch": 1.227526530612245,
      "grad_norm": 0.7395092248916626,
      "learning_rate": 3.370356044199785e-05,
      "loss": 0.1622,
      "mean_token_accuracy": 0.9760421338677406,
      "num_tokens": 441776589.0,
      "step": 4700
    },
    {
      "entropy": 0.5041036491096019,
      "epoch": 1.2405877551020408,
      "grad_norm": 0.6011821031570435,
      "learning_rate": 3.271771608020324e-05,
      "loss": 0.1282,
      "mean_token_accuracy": 0.9775787754356862,
      "num_tokens": 446474446.0,
      "step": 4750
    },
    {
      "entropy": 0.5190455333888531,
      "epoch": 1.2536489795918366,
      "grad_norm": 0.24861446022987366,
      "learning_rate": 3.173944422012688e-05,
      "loss": 0.1396,
      "mean_token_accuracy": 0.9773434112966061,
      "num_tokens": 451196726.0,
      "step": 4800
    },
    {
      "entropy": 0.5022503357380629,
      "epoch": 1.2667102040816327,
      "grad_norm": 0.24570924043655396,
      "learning_rate": 3.076917350681222e-05,
      "loss": 0.1592,
      "mean_token_accuracy": 0.9746822196245194,
      "num_tokens": 455888329.0,
      "step": 4850
    },
    {
      "entropy": 0.5063729017972947,
      "epoch": 1.2797714285714286,
      "grad_norm": 0.843909502029419,
      "learning_rate": 2.9807329079475686e-05,
      "loss": 0.1479,
      "mean_token_accuracy": 0.9775361755490303,
      "num_tokens": 460588507.0,
      "step": 4900
    },
    {
      "entropy": 0.5099849846214056,
      "epoch": 1.2928326530612244,
      "grad_norm": 0.5658782124519348,
      "learning_rate": 2.8854332385225336e-05,
      "loss": 0.1365,
      "mean_token_accuracy": 0.9764770893752575,
      "num_tokens": 465292875.0,
      "step": 4950
    },
    {
      "entropy": 0.5075208330154419,
      "epoch": 1.3058938775510205,
      "grad_norm": 0.5681100487709045,
      "learning_rate": 2.791060099439724e-05,
      "loss": 0.1547,
      "mean_token_accuracy": 0.9766012115776539,
      "num_tokens": 470018553.0,
      "step": 5000
    },
    {
      "epoch": 1.3058938775510205,
      "eval_entropy": 0.5093407655239105,
      "eval_loss": 0.07968976348638535,
      "eval_mean_token_accuracy": 0.9728172892570496,
      "eval_num_tokens": 470018553.0,
      "eval_runtime": 105.9871,
      "eval_samples_per_second": 188.702,
      "eval_steps_per_second": 5.897,
      "step": 5000
    },
    {
      "entropy": 0.5292845606803894,
      "epoch": 1.3189551020408163,
      "grad_norm": 0.3377012014389038,
      "learning_rate": 2.6976548417590607e-05,
      "loss": 0.1478,
      "mean_token_accuracy": 0.9774796655774116,
      "num_tokens": 474695442.0,
      "step": 5050
    },
    {
      "entropy": 0.5115040161460638,
      "epoch": 1.3320163265306122,
      "grad_norm": 0.31330540776252747,
      "learning_rate": 2.6052583924481532e-05,
      "loss": 0.1651,
      "mean_token_accuracy": 0.9753667731583119,
      "num_tokens": 479387979.0,
      "step": 5100
    },
    {
      "entropy": 0.518300534710288,
      "epoch": 1.3450775510204083,
      "grad_norm": 0.579721987247467,
      "learning_rate": 2.5139112364495286e-05,
      "loss": 0.2114,
      "mean_token_accuracy": 0.975457195341587,
      "num_tokens": 484096122.0,
      "step": 5150
    },
    {
      "entropy": 0.5092091238498688,
      "epoch": 1.3581387755102041,
      "grad_norm": 0.04276373237371445,
      "learning_rate": 2.4236533989415084e-05,
      "loss": 0.1473,
      "mean_token_accuracy": 0.978675073236227,
      "num_tokens": 488804070.0,
      "step": 5200
    },
    {
      "entropy": 0.5108030308783055,
      "epoch": 1.3712,
      "grad_norm": 0.34016624093055725,
      "learning_rate": 2.334524427800569e-05,
      "loss": 0.1871,
      "mean_token_accuracy": 0.9755242675542831,
      "num_tokens": 493512802.0,
      "step": 5250
    },
    {
      "entropy": 0.5037990070134402,
      "epoch": 1.384261224489796,
      "grad_norm": 0.058522939682006836,
      "learning_rate": 2.2465633762728094e-05,
      "loss": 0.1241,
      "mean_token_accuracy": 0.9779052963852882,
      "num_tokens": 498224036.0,
      "step": 5300
    },
    {
      "entropy": 0.5202232024073601,
      "epoch": 1.397322448979592,
      "grad_norm": 0.7593867182731628,
      "learning_rate": 2.1598087858621707e-05,
      "loss": 0.1528,
      "mean_token_accuracy": 0.9765554815530777,
      "num_tokens": 502927469.0,
      "step": 5350
    },
    {
      "entropy": 0.5276015343517065,
      "epoch": 1.4103836734693878,
      "grad_norm": 0.8394476175308228,
      "learning_rate": 2.074298669442881e-05,
      "loss": 0.1965,
      "mean_token_accuracy": 0.9732386584579945,
      "num_tokens": 507642470.0,
      "step": 5400
    },
    {
      "entropy": 0.5179370325058699,
      "epoch": 1.4234448979591836,
      "grad_norm": 0.280915766954422,
      "learning_rate": 1.9900704946035115e-05,
      "loss": 0.1817,
      "mean_token_accuracy": 0.9782391162216664,
      "num_tokens": 512330334.0,
      "step": 5450
    },
    {
      "entropy": 0.5077722806483507,
      "epoch": 1.4365061224489795,
      "grad_norm": 0.37315860390663147,
      "learning_rate": 1.9071611672299756e-05,
      "loss": 0.1516,
      "mean_token_accuracy": 0.9786237494647503,
      "num_tokens": 517049296.0,
      "step": 5500
    },
    {
      "entropy": 0.521270957365632,
      "epoch": 1.4495673469387755,
      "grad_norm": 0.8134558200836182,
      "learning_rate": 1.825607015334644e-05,
      "loss": 0.1509,
      "mean_token_accuracy": 0.9779170905053616,
      "num_tokens": 521739525.0,
      "step": 5550
    },
    {
      "entropy": 0.5181563084572554,
      "epoch": 1.4626285714285714,
      "grad_norm": 0.17686356604099274,
      "learning_rate": 1.7454437731386558e-05,
      "loss": 0.2467,
      "mean_token_accuracy": 0.9748677863925695,
      "num_tokens": 526451360.0,
      "step": 5600
    },
    {
      "entropy": 0.5103141793608665,
      "epoch": 1.4756897959183672,
      "grad_norm": 0.45316168665885925,
      "learning_rate": 1.6667065654144103e-05,
      "loss": 0.1965,
      "mean_token_accuracy": 0.9771437807381154,
      "num_tokens": 531176496.0,
      "step": 5650
    },
    {
      "entropy": 0.5023327280580997,
      "epoch": 1.4887510204081633,
      "grad_norm": 0.20157447457313538,
      "learning_rate": 1.5894298920951106e-05,
      "loss": 0.1909,
      "mean_token_accuracy": 0.9771021957695484,
      "num_tokens": 535864937.0,
      "step": 5700
    },
    {
      "entropy": 0.5016037753224373,
      "epoch": 1.5018122448979592,
      "grad_norm": 0.33676010370254517,
      "learning_rate": 1.5136476131580701e-05,
      "loss": 0.1675,
      "mean_token_accuracy": 0.9775694438815117,
      "num_tokens": 540571806.0,
      "step": 5750
    },
    {
      "entropy": 0.5027693736553193,
      "epoch": 1.514873469387755,
      "grad_norm": 0.5053212642669678,
      "learning_rate": 1.439392933788447e-05,
      "loss": 0.1807,
      "mean_token_accuracy": 0.9797284063696862,
      "num_tokens": 545276790.0,
      "step": 5800
    },
    {
      "entropy": 0.5009680130332709,
      "epoch": 1.527934693877551,
      "grad_norm": 0.09157374501228333,
      "learning_rate": 1.3666983898298657e-05,
      "loss": 0.2042,
      "mean_token_accuracy": 0.977972183674574,
      "num_tokens": 549964631.0,
      "step": 5850
    },
    {
      "entropy": 0.5023607714474201,
      "epoch": 1.540995918367347,
      "grad_norm": 0.23668606579303741,
      "learning_rate": 1.2955958335283353e-05,
      "loss": 0.2045,
      "mean_token_accuracy": 0.9760315857827664,
      "num_tokens": 554656554.0,
      "step": 5900
    },
    {
      "entropy": 0.505893104672432,
      "epoch": 1.5540571428571428,
      "grad_norm": 0.2585294246673584,
      "learning_rate": 1.2261164195756963e-05,
      "loss": 0.1519,
      "mean_token_accuracy": 0.9788727892935276,
      "num_tokens": 559319350.0,
      "step": 5950
    },
    {
      "entropy": 0.5045397817343473,
      "epoch": 1.5671183673469389,
      "grad_norm": 0.41178709268569946,
      "learning_rate": 1.1582905914586961e-05,
      "loss": 0.1618,
      "mean_token_accuracy": 0.9804812078177929,
      "num_tokens": 564036376.0,
      "step": 6000
    },
    {
      "epoch": 1.5671183673469389,
      "eval_entropy": 0.5037064853668213,
      "eval_loss": 0.06929752975702286,
      "eval_mean_token_accuracy": 0.9767320104598999,
      "eval_num_tokens": 564036376.0,
      "eval_runtime": 105.8974,
      "eval_samples_per_second": 188.862,
      "eval_steps_per_second": 5.902,
      "step": 6000
    }
  ],
  "logging_steps": 50,
  "max_steps": 7658,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 1000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 3.804321079366989e+19,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}