{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.6,
  "eval_steps": 100,
  "global_step": 1000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "entropy": 1.000200616195798,
      "epoch": 0.016,
      "grad_norm": 1776.0,
      "learning_rate": 4.5000000000000003e-07,
      "loss": 69.8616,
      "mean_token_accuracy": 0.460323916003108,
      "num_tokens": 195524.0,
      "step": 10
    },
    {
      "entropy": 0.9949515227228403,
      "epoch": 0.032,
      "grad_norm": 1624.0,
      "learning_rate": 9.500000000000001e-07,
      "loss": 68.3351,
      "mean_token_accuracy": 0.4633398361504078,
      "num_tokens": 390903.0,
      "step": 20
    },
    {
      "entropy": 1.0845801826566457,
      "epoch": 0.048,
      "grad_norm": 1728.0,
      "learning_rate": 1.45e-06,
      "loss": 63.0644,
      "mean_token_accuracy": 0.46960638258606197,
      "num_tokens": 589868.0,
      "step": 30
    },
    {
      "entropy": 1.247272988036275,
      "epoch": 0.064,
      "grad_norm": 488.0,
      "learning_rate": 1.9500000000000004e-06,
      "loss": 55.5844,
      "mean_token_accuracy": 0.480246701836586,
      "num_tokens": 791190.0,
      "step": 40
    },
    {
      "entropy": 1.4826370410621166,
      "epoch": 0.08,
      "grad_norm": 288.0,
      "learning_rate": 2.4500000000000003e-06,
      "loss": 50.1588,
      "mean_token_accuracy": 0.48712102621793746,
      "num_tokens": 989860.0,
      "step": 50
    },
    {
      "entropy": 1.814505684375763,
      "epoch": 0.096,
      "grad_norm": 131.0,
      "learning_rate": 2.95e-06,
      "loss": 45.3237,
      "mean_token_accuracy": 0.48977465480566024,
      "num_tokens": 1181777.0,
      "step": 60
    },
    {
      "entropy": 2.17904861420393,
      "epoch": 0.112,
      "grad_norm": 67.5,
      "learning_rate": 3.45e-06,
      "loss": 40.572,
      "mean_token_accuracy": 0.5055354550480843,
      "num_tokens": 1385513.0,
      "step": 70
    },
    {
      "entropy": 2.317786781489849,
      "epoch": 0.128,
      "grad_norm": 55.25,
      "learning_rate": 3.95e-06,
      "loss": 38.0783,
      "mean_token_accuracy": 0.5195932997390628,
      "num_tokens": 1582368.0,
      "step": 80
    },
    {
      "entropy": 2.2404811814427377,
      "epoch": 0.144,
      "grad_norm": 40.75,
      "learning_rate": 4.450000000000001e-06,
      "loss": 35.7283,
      "mean_token_accuracy": 0.5399348795413971,
      "num_tokens": 1773764.0,
      "step": 90
    },
    {
      "entropy": 2.1976587153971194,
      "epoch": 0.16,
      "grad_norm": 38.0,
      "learning_rate": 4.95e-06,
      "loss": 34.5297,
      "mean_token_accuracy": 0.5478190431371331,
      "num_tokens": 1970077.0,
      "step": 100
    },
    {
      "epoch": 0.16,
      "eval_biology_entropy": 2.1061868534088135,
      "eval_biology_loss": 2.0133121013641357,
      "eval_biology_mean_token_accuracy": 0.5668603234291076,
      "eval_biology_num_tokens": 1970077.0,
      "eval_biology_runtime": 47.9413,
      "eval_biology_samples_per_second": 10.429,
      "eval_biology_steps_per_second": 2.607,
      "step": 100
    },
    {
      "epoch": 0.16,
      "eval_chemistry_entropy": 1.8679178915023804,
      "eval_chemistry_loss": 1.844948410987854,
      "eval_chemistry_mean_token_accuracy": 0.5880863065719605,
      "eval_chemistry_num_tokens": 1970077.0,
      "eval_chemistry_runtime": 59.3684,
      "eval_chemistry_samples_per_second": 8.422,
      "eval_chemistry_steps_per_second": 2.105,
      "step": 100
    },
    {
      "epoch": 0.16,
      "eval_math_entropy": 1.6290378885269166,
      "eval_math_loss": 2.072597026824951,
      "eval_math_mean_token_accuracy": 0.5653720505237579,
      "eval_math_num_tokens": 1970077.0,
      "eval_math_runtime": 60.948,
      "eval_math_samples_per_second": 8.204,
      "eval_math_steps_per_second": 2.051,
      "step": 100
    },
    {
      "epoch": 0.16,
      "eval_physics_entropy": 1.907263669013977,
      "eval_physics_loss": 2.105146884918213,
      "eval_physics_mean_token_accuracy": 0.5574561676979065,
      "eval_physics_num_tokens": 1970077.0,
      "eval_physics_runtime": 69.3465,
      "eval_physics_samples_per_second": 7.21,
      "eval_physics_steps_per_second": 1.803,
      "step": 100
    },
    {
      "entropy": 2.0820947512984276,
      "epoch": 0.176,
      "grad_norm": 33.5,
      "learning_rate": 5.450000000000001e-06,
      "loss": 32.4406,
      "mean_token_accuracy": 0.5621457317844033,
      "num_tokens": 2168354.0,
      "step": 110
    },
    {
      "entropy": 2.0283117063343523,
      "epoch": 0.192,
      "grad_norm": 32.25,
      "learning_rate": 5.950000000000001e-06,
      "loss": 31.491,
      "mean_token_accuracy": 0.5688053227961063,
      "num_tokens": 2365822.0,
      "step": 120
    },
    {
      "entropy": 1.973121351003647,
      "epoch": 0.208,
      "grad_norm": 33.75,
      "learning_rate": 6.450000000000001e-06,
      "loss": 30.9217,
      "mean_token_accuracy": 0.5753367917612195,
      "num_tokens": 2558762.0,
      "step": 130
    },
    {
      "entropy": 1.9165778771042823,
      "epoch": 0.224,
      "grad_norm": 30.625,
      "learning_rate": 6.95e-06,
      "loss": 30.1267,
      "mean_token_accuracy": 0.583262656442821,
      "num_tokens": 2755347.0,
      "step": 140
    },
    {
      "entropy": 1.8870679274201394,
      "epoch": 0.24,
      "grad_norm": 31.5,
      "learning_rate": 7.450000000000001e-06,
      "loss": 29.6484,
      "mean_token_accuracy": 0.5872904889285564,
      "num_tokens": 2947346.0,
      "step": 150
    },
    {
      "entropy": 1.8710497491061688,
      "epoch": 0.256,
      "grad_norm": 29.875,
      "learning_rate": 7.950000000000002e-06,
      "loss": 29.4624,
      "mean_token_accuracy": 0.5887285262346268,
      "num_tokens": 3139957.0,
      "step": 160
    },
    {
      "entropy": 1.808907500654459,
      "epoch": 0.272,
      "grad_norm": 29.125,
      "learning_rate": 8.45e-06,
      "loss": 28.7075,
      "mean_token_accuracy": 0.5965064790099859,
      "num_tokens": 3335951.0,
      "step": 170
    },
    {
      "entropy": 1.774143099039793,
      "epoch": 0.288,
      "grad_norm": 28.125,
      "learning_rate": 8.95e-06,
      "loss": 28.1007,
      "mean_token_accuracy": 0.5996631737798452,
      "num_tokens": 3539731.0,
      "step": 180
    },
    {
      "entropy": 1.7740100547671318,
      "epoch": 0.304,
      "grad_norm": 27.75,
      "learning_rate": 9.450000000000001e-06,
      "loss": 28.1122,
      "mean_token_accuracy": 0.6012205760926008,
      "num_tokens": 3733488.0,
      "step": 190
    },
    {
      "entropy": 1.7725927896797657,
      "epoch": 0.32,
      "grad_norm": 35.25,
      "learning_rate": 9.950000000000001e-06,
      "loss": 28.1299,
      "mean_token_accuracy": 0.5997324202209711,
      "num_tokens": 3920545.0,
      "step": 200
    },
    {
      "epoch": 0.32,
      "eval_biology_entropy": 1.6843824968338013,
      "eval_biology_loss": 1.6854687929153442,
      "eval_biology_mean_token_accuracy": 0.6137065634727478,
      "eval_biology_num_tokens": 3920545.0,
      "eval_biology_runtime": 47.8268,
      "eval_biology_samples_per_second": 10.454,
      "eval_biology_steps_per_second": 2.614,
      "step": 200
    },
    {
      "epoch": 0.32,
      "eval_chemistry_entropy": 1.441379102230072,
      "eval_chemistry_loss": 1.4537148475646973,
      "eval_chemistry_mean_token_accuracy": 0.6551240344047546,
      "eval_chemistry_num_tokens": 3920545.0,
      "eval_chemistry_runtime": 59.3401,
      "eval_chemistry_samples_per_second": 8.426,
      "eval_chemistry_steps_per_second": 2.106,
      "step": 200
    },
    {
      "epoch": 0.32,
      "eval_math_entropy": 1.3251727666854858,
      "eval_math_loss": 1.6510798931121826,
      "eval_math_mean_token_accuracy": 0.6343780555725098,
      "eval_math_num_tokens": 3920545.0,
      "eval_math_runtime": 60.8747,
      "eval_math_samples_per_second": 8.214,
      "eval_math_steps_per_second": 2.053,
      "step": 200
    },
    {
      "epoch": 0.32,
      "eval_physics_entropy": 1.4914224824905395,
      "eval_physics_loss": 1.6194021701812744,
      "eval_physics_mean_token_accuracy": 0.6387534441947937,
      "eval_physics_num_tokens": 3920545.0,
      "eval_physics_runtime": 69.3484,
      "eval_physics_samples_per_second": 7.21,
      "eval_physics_steps_per_second": 1.802,
      "step": 200
    },
    {
      "entropy": 1.7085595309734345,
      "epoch": 0.336,
      "grad_norm": 30.375,
      "learning_rate": 1.045e-05,
      "loss": 27.2575,
      "mean_token_accuracy": 0.6091335583478212,
      "num_tokens": 4114077.0,
      "step": 210
    },
    {
      "entropy": 1.7301082991063594,
      "epoch": 0.352,
      "grad_norm": 31.0,
      "learning_rate": 1.095e-05,
      "loss": 27.6337,
      "mean_token_accuracy": 0.6071538373827934,
      "num_tokens": 4306949.0,
      "step": 220
    },
    {
      "entropy": 1.695884221047163,
      "epoch": 0.368,
      "grad_norm": 27.75,
      "learning_rate": 1.145e-05,
      "loss": 27.0338,
      "mean_token_accuracy": 0.6128786243498325,
      "num_tokens": 4504001.0,
      "step": 230
    },
    {
      "entropy": 1.6912743583321572,
      "epoch": 0.384,
      "grad_norm": 31.0,
      "learning_rate": 1.195e-05,
      "loss": 27.0143,
      "mean_token_accuracy": 0.6135674986988306,
      "num_tokens": 4693812.0,
      "step": 240
    },
    {
      "entropy": 1.6936707220971585,
      "epoch": 0.4,
      "grad_norm": 24.5,
      "learning_rate": 1.2450000000000003e-05,
      "loss": 26.8651,
      "mean_token_accuracy": 0.6143408048897981,
      "num_tokens": 4887094.0,
      "step": 250
    },
    {
      "entropy": 1.678075761348009,
      "epoch": 0.416,
      "grad_norm": 27.875,
      "learning_rate": 1.295e-05,
      "loss": 27.047,
      "mean_token_accuracy": 0.6129864633083344,
      "num_tokens": 5085369.0,
      "step": 260
    },
    {
      "entropy": 1.6657070152461528,
      "epoch": 0.432,
      "grad_norm": 29.75,
      "learning_rate": 1.3450000000000002e-05,
      "loss": 26.7844,
      "mean_token_accuracy": 0.6148818587884307,
      "num_tokens": 5271275.0,
      "step": 270
    },
    {
      "entropy": 1.6670012198388577,
      "epoch": 0.448,
      "grad_norm": 31.375,
      "learning_rate": 1.3950000000000002e-05,
      "loss": 26.617,
      "mean_token_accuracy": 0.6159537024796009,
      "num_tokens": 5460559.0,
      "step": 280
    },
    {
      "entropy": 1.6152651645243168,
      "epoch": 0.464,
      "grad_norm": 32.25,
      "learning_rate": 1.4450000000000002e-05,
      "loss": 25.6477,
      "mean_token_accuracy": 0.6256863068789243,
      "num_tokens": 5653809.0,
      "step": 290
    },
    {
      "entropy": 1.624342393130064,
      "epoch": 0.48,
      "grad_norm": 25.75,
      "learning_rate": 1.4950000000000003e-05,
      "loss": 26.2019,
      "mean_token_accuracy": 0.6199633304029704,
      "num_tokens": 5850176.0,
      "step": 300
    },
    {
      "epoch": 0.48,
      "eval_biology_entropy": 1.5827035818099975,
      "eval_biology_loss": 1.570924997329712,
      "eval_biology_mean_token_accuracy": 0.6328578324317932,
      "eval_biology_num_tokens": 5850176.0,
      "eval_biology_runtime": 47.7211,
      "eval_biology_samples_per_second": 10.478,
      "eval_biology_steps_per_second": 2.619,
      "step": 300
    },
    {
      "epoch": 0.48,
      "eval_chemistry_entropy": 1.3193063869476318,
      "eval_chemistry_loss": 1.3243473768234253,
      "eval_chemistry_mean_token_accuracy": 0.6796774597167968,
      "eval_chemistry_num_tokens": 5850176.0,
      "eval_chemistry_runtime": 59.1912,
      "eval_chemistry_samples_per_second": 8.447,
      "eval_chemistry_steps_per_second": 2.112,
      "step": 300
    },
    {
      "epoch": 0.48,
      "eval_math_entropy": 1.2024567556381225,
      "eval_math_loss": 1.5285311937332153,
      "eval_math_mean_token_accuracy": 0.6541934962272644,
      "eval_math_num_tokens": 5850176.0,
      "eval_math_runtime": 60.6728,
      "eval_math_samples_per_second": 8.241,
      "eval_math_steps_per_second": 2.06,
      "step": 300
    },
    {
      "epoch": 0.48,
      "eval_physics_entropy": 1.3696330366134644,
      "eval_physics_loss": 1.4624738693237305,
      "eval_physics_mean_token_accuracy": 0.6671717586517334,
      "eval_physics_num_tokens": 5850176.0,
      "eval_physics_runtime": 69.1435,
      "eval_physics_samples_per_second": 7.231,
      "eval_physics_steps_per_second": 1.808,
      "step": 300
    },
    {
      "entropy": 1.543294757604599,
      "epoch": 0.496,
      "grad_norm": 24.125,
      "learning_rate": 1.545e-05,
      "loss": 24.7301,
      "mean_token_accuracy": 0.6344853390008212,
      "num_tokens": 6046503.0,
      "step": 310
    },
    {
      "entropy": 1.6137450806796552,
      "epoch": 0.512,
      "grad_norm": 30.0,
      "learning_rate": 1.595e-05,
      "loss": 25.8005,
      "mean_token_accuracy": 0.624148428812623,
      "num_tokens": 6240456.0,
      "step": 320
    },
    {
      "entropy": 1.5901939049363136,
      "epoch": 0.528,
      "grad_norm": 26.75,
      "learning_rate": 1.645e-05,
      "loss": 25.5556,
      "mean_token_accuracy": 0.6283756781369447,
      "num_tokens": 6430555.0,
      "step": 330
    },
    {
      "entropy": 1.5919260658323764,
      "epoch": 0.544,
      "grad_norm": 25.375,
      "learning_rate": 1.6950000000000002e-05,
      "loss": 25.5888,
      "mean_token_accuracy": 0.6276101280003786,
      "num_tokens": 6626006.0,
      "step": 340
    },
    {
      "entropy": 1.5667011991143227,
      "epoch": 0.56,
      "grad_norm": 27.375,
      "learning_rate": 1.7450000000000004e-05,
      "loss": 25.1583,
      "mean_token_accuracy": 0.6304543860256672,
      "num_tokens": 6820754.0,
      "step": 350
    },
    {
      "entropy": 1.548937487602234,
      "epoch": 0.576,
      "grad_norm": 25.0,
      "learning_rate": 1.795e-05,
      "loss": 24.8095,
      "mean_token_accuracy": 0.6340219389647246,
      "num_tokens": 7021844.0,
      "step": 360
    },
    {
      "entropy": 1.5519795384258033,
      "epoch": 0.592,
      "grad_norm": 25.75,
      "learning_rate": 1.845e-05,
      "loss": 25.0794,
      "mean_token_accuracy": 0.6325468828901648,
      "num_tokens": 7213951.0,
      "step": 370
    },
    {
      "entropy": 1.5391070768237114,
      "epoch": 0.608,
      "grad_norm": 28.25,
      "learning_rate": 1.8950000000000003e-05,
      "loss": 24.6819,
      "mean_token_accuracy": 0.636061006039381,
      "num_tokens": 7416773.0,
      "step": 380
    },
    {
      "entropy": 1.5429081372916698,
      "epoch": 0.624,
      "grad_norm": 27.5,
      "learning_rate": 1.9450000000000002e-05,
      "loss": 24.8187,
      "mean_token_accuracy": 0.634391113743186,
      "num_tokens": 7612843.0,
      "step": 390
    },
    {
      "entropy": 1.5399384461343288,
      "epoch": 0.64,
      "grad_norm": 25.875,
      "learning_rate": 1.9950000000000004e-05,
      "loss": 24.8584,
      "mean_token_accuracy": 0.6340381380170583,
      "num_tokens": 7801633.0,
      "step": 400
    },
    {
      "epoch": 0.64,
      "eval_biology_entropy": 1.4887315626144408,
      "eval_biology_loss": 1.4859997034072876,
      "eval_biology_mean_token_accuracy": 0.6469658589363099,
      "eval_biology_num_tokens": 7801633.0,
      "eval_biology_runtime": 47.815,
      "eval_biology_samples_per_second": 10.457,
      "eval_biology_steps_per_second": 2.614,
      "step": 400
    },
    {
      "epoch": 0.64,
      "eval_chemistry_entropy": 1.216160174369812,
      "eval_chemistry_loss": 1.23386549949646,
      "eval_chemistry_mean_token_accuracy": 0.6968556141853333,
      "eval_chemistry_num_tokens": 7801633.0,
      "eval_chemistry_runtime": 59.3099,
      "eval_chemistry_samples_per_second": 8.43,
      "eval_chemistry_steps_per_second": 2.108,
      "step": 400
    },
    {
      "epoch": 0.64,
      "eval_math_entropy": 1.0936446270942688,
      "eval_math_loss": 1.4483542442321777,
      "eval_math_mean_token_accuracy": 0.6711465301513672,
      "eval_math_num_tokens": 7801633.0,
      "eval_math_runtime": 60.6,
      "eval_math_samples_per_second": 8.251,
      "eval_math_steps_per_second": 2.063,
      "step": 400
    },
    {
      "epoch": 0.64,
      "eval_physics_entropy": 1.2649442830085755,
      "eval_physics_loss": 1.3500477075576782,
      "eval_physics_mean_token_accuracy": 0.6875980172157288,
      "eval_physics_num_tokens": 7801633.0,
      "eval_physics_runtime": 68.83,
      "eval_physics_samples_per_second": 7.264,
      "eval_physics_steps_per_second": 1.816,
      "step": 400
    },
    {
      "entropy": 1.5343165121972562,
      "epoch": 0.656,
      "grad_norm": 26.375,
      "learning_rate": 1.9950000000000004e-05,
      "loss": 24.7076,
      "mean_token_accuracy": 0.6344317115843296,
      "num_tokens": 7995843.0,
      "step": 410
    },
    {
      "entropy": 1.5481104589998722,
      "epoch": 0.672,
      "grad_norm": 24.75,
      "learning_rate": 1.9894444444444446e-05,
      "loss": 24.7524,
      "mean_token_accuracy": 0.6355006828904152,
      "num_tokens": 8183103.0,
      "step": 420
    },
    {
      "entropy": 1.5163978930562734,
      "epoch": 0.688,
      "grad_norm": 22.75,
      "learning_rate": 1.983888888888889e-05,
      "loss": 24.5063,
      "mean_token_accuracy": 0.6392664317041635,
      "num_tokens": 8385976.0,
      "step": 430
    },
    {
      "entropy": 1.5027333579957485,
      "epoch": 0.704,
      "grad_norm": 21.875,
      "learning_rate": 1.9783333333333334e-05,
      "loss": 24.2448,
      "mean_token_accuracy": 0.641432274878025,
      "num_tokens": 8578431.0,
      "step": 440
    },
    {
      "entropy": 1.4510110400617122,
      "epoch": 0.72,
      "grad_norm": 22.625,
      "learning_rate": 1.972777777777778e-05,
      "loss": 23.3137,
      "mean_token_accuracy": 0.6509528819471597,
      "num_tokens": 8781342.0,
      "step": 450
    },
    {
      "entropy": 1.471246886998415,
      "epoch": 0.736,
      "grad_norm": 22.0,
      "learning_rate": 1.9672222222222222e-05,
      "loss": 23.5955,
      "mean_token_accuracy": 0.6463174737989903,
      "num_tokens": 8977918.0,
      "step": 460
    },
    {
      "entropy": 1.4814870670437812,
      "epoch": 0.752,
      "grad_norm": 22.25,
      "learning_rate": 1.9616666666666668e-05,
      "loss": 23.9505,
      "mean_token_accuracy": 0.6442026626318693,
      "num_tokens": 9169322.0,
      "step": 470
    },
    {
      "entropy": 1.4392801743000745,
      "epoch": 0.768,
      "grad_norm": 24.125,
      "learning_rate": 1.9561111111111114e-05,
      "loss": 23.1521,
      "mean_token_accuracy": 0.6515602272003889,
      "num_tokens": 9368141.0,
      "step": 480
    },
    {
      "entropy": 1.4612651772797107,
      "epoch": 0.784,
      "grad_norm": 25.875,
      "learning_rate": 1.9505555555555556e-05,
      "loss": 23.5729,
      "mean_token_accuracy": 0.6485392026603222,
      "num_tokens": 9565236.0,
      "step": 490
    },
    {
      "entropy": 1.4798763774335384,
      "epoch": 0.8,
      "grad_norm": 24.0,
      "learning_rate": 1.9450000000000002e-05,
      "loss": 23.9628,
      "mean_token_accuracy": 0.6433830507099628,
      "num_tokens": 9761227.0,
      "step": 500
    },
    {
      "epoch": 0.8,
      "eval_biology_entropy": 1.4178470463752746,
      "eval_biology_loss": 1.4294027090072632,
      "eval_biology_mean_token_accuracy": 0.6571813192367554,
      "eval_biology_num_tokens": 9761227.0,
      "eval_biology_runtime": 47.9552,
      "eval_biology_samples_per_second": 10.426,
      "eval_biology_steps_per_second": 2.607,
      "step": 500
    },
    {
      "epoch": 0.8,
      "eval_chemistry_entropy": 1.1406473817825318,
      "eval_chemistry_loss": 1.1737452745437622,
      "eval_chemistry_mean_token_accuracy": 0.7092447648048401,
      "eval_chemistry_num_tokens": 9761227.0,
      "eval_chemistry_runtime": 58.9872,
      "eval_chemistry_samples_per_second": 8.476,
      "eval_chemistry_steps_per_second": 2.119,
      "step": 500
    },
    {
      "epoch": 0.8,
      "eval_math_entropy": 1.0307734375,
      "eval_math_loss": 1.3721623420715332,
      "eval_math_mean_token_accuracy": 0.6881859846115113,
      "eval_math_num_tokens": 9761227.0,
      "eval_math_runtime": 60.7077,
      "eval_math_samples_per_second": 8.236,
      "eval_math_steps_per_second": 2.059,
      "step": 500
    },
    {
      "epoch": 0.8,
      "eval_physics_entropy": 1.1864667072296142,
      "eval_physics_loss": 1.28458571434021,
      "eval_physics_mean_token_accuracy": 0.701751223564148,
      "eval_physics_num_tokens": 9761227.0,
      "eval_physics_runtime": 69.206,
      "eval_physics_samples_per_second": 7.225,
      "eval_physics_steps_per_second": 1.806,
      "step": 500
    },
    {
      "entropy": 1.4389251872897149,
      "epoch": 0.816,
      "grad_norm": 24.75,
      "learning_rate": 1.9394444444444448e-05,
      "loss": 23.1005,
      "mean_token_accuracy": 0.6514821499586105,
      "num_tokens": 9958727.0,
      "step": 510
    },
    {
      "entropy": 1.4328688051551581,
      "epoch": 0.832,
      "grad_norm": 21.625,
      "learning_rate": 1.933888888888889e-05,
      "loss": 23.1468,
      "mean_token_accuracy": 0.6508167646825314,
      "num_tokens": 10155771.0,
      "step": 520
    },
    {
      "entropy": 1.4297623857855797,
      "epoch": 0.848,
      "grad_norm": 24.375,
      "learning_rate": 1.9283333333333332e-05,
      "loss": 23.0084,
      "mean_token_accuracy": 0.6537799876183271,
      "num_tokens": 10357721.0,
      "step": 530
    },
    {
      "entropy": 1.4593383342027664,
      "epoch": 0.864,
      "grad_norm": 23.375,
      "learning_rate": 1.9227777777777778e-05,
      "loss": 23.4158,
      "mean_token_accuracy": 0.6497054308652878,
      "num_tokens": 10552495.0,
      "step": 540
    },
    {
      "entropy": 1.4262017775326967,
      "epoch": 0.88,
      "grad_norm": 26.125,
      "learning_rate": 1.9172222222222224e-05,
      "loss": 23.0791,
      "mean_token_accuracy": 0.6533483766019345,
      "num_tokens": 10748749.0,
      "step": 550
    },
    {
      "entropy": 1.449614517390728,
      "epoch": 0.896,
      "grad_norm": 23.5,
      "learning_rate": 1.911666666666667e-05,
      "loss": 23.4292,
      "mean_token_accuracy": 0.6497185111045838,
      "num_tokens": 10943319.0,
      "step": 560
    },
    {
      "entropy": 1.4316417548805476,
      "epoch": 0.912,
      "grad_norm": 24.25,
      "learning_rate": 1.9061111111111112e-05,
      "loss": 23.1123,
      "mean_token_accuracy": 0.6529628891497851,
      "num_tokens": 11136935.0,
      "step": 570
    },
    {
      "entropy": 1.4503454469144343,
      "epoch": 0.928,
      "grad_norm": 22.875,
      "learning_rate": 1.9005555555555557e-05,
      "loss": 23.2073,
      "mean_token_accuracy": 0.6513680387288332,
      "num_tokens": 11331098.0,
      "step": 580
    },
    {
      "entropy": 1.3842108126729726,
      "epoch": 0.944,
      "grad_norm": 20.375,
      "learning_rate": 1.8950000000000003e-05,
      "loss": 22.459,
      "mean_token_accuracy": 0.657518332451582,
      "num_tokens": 11530550.0,
      "step": 590
    },
    {
      "entropy": 1.4227556131780148,
      "epoch": 0.96,
      "grad_norm": 24.625,
      "learning_rate": 1.8894444444444446e-05,
      "loss": 22.9116,
      "mean_token_accuracy": 0.6550851400941611,
      "num_tokens": 11729645.0,
      "step": 600
    },
    {
      "epoch": 0.96,
      "eval_biology_entropy": 1.408647123813629,
      "eval_biology_loss": 1.3943630456924438,
      "eval_biology_mean_token_accuracy": 0.6631479191780091,
      "eval_biology_num_tokens": 11729645.0,
      "eval_biology_runtime": 47.7888,
      "eval_biology_samples_per_second": 10.463,
      "eval_biology_steps_per_second": 2.616,
      "step": 600
    },
    {
      "epoch": 0.96,
      "eval_chemistry_entropy": 1.146325659751892,
      "eval_chemistry_loss": 1.1449472904205322,
      "eval_chemistry_mean_token_accuracy": 0.714297685623169,
      "eval_chemistry_num_tokens": 11729645.0,
      "eval_chemistry_runtime": 59.1809,
      "eval_chemistry_samples_per_second": 8.449,
      "eval_chemistry_steps_per_second": 2.112,
      "step": 600
    },
    {
      "epoch": 0.96,
      "eval_math_entropy": 1.0387896914482118,
      "eval_math_loss": 1.3466871976852417,
      "eval_math_mean_token_accuracy": 0.6919928450584412,
      "eval_math_num_tokens": 11729645.0,
      "eval_math_runtime": 60.7385,
      "eval_math_samples_per_second": 8.232,
      "eval_math_steps_per_second": 2.058,
      "step": 600
    },
    {
      "epoch": 0.96,
      "eval_physics_entropy": 1.2005870642662049,
      "eval_physics_loss": 1.254885196685791,
      "eval_physics_mean_token_accuracy": 0.707194664478302,
      "eval_physics_num_tokens": 11729645.0,
      "eval_physics_runtime": 69.1629,
      "eval_physics_samples_per_second": 7.229,
      "eval_physics_steps_per_second": 1.807,
      "step": 600
    },
    {
      "entropy": 1.3916049905121326,
      "epoch": 0.976,
      "grad_norm": 26.0,
      "learning_rate": 1.883888888888889e-05,
      "loss": 22.4356,
      "mean_token_accuracy": 0.6580898258835077,
      "num_tokens": 11924644.0,
      "step": 610
    },
    {
      "entropy": 1.415398271009326,
      "epoch": 0.992,
      "grad_norm": 22.0,
      "learning_rate": 1.8783333333333337e-05,
      "loss": 22.9259,
      "mean_token_accuracy": 0.6539454534649849,
      "num_tokens": 12123059.0,
      "step": 620
    },
    {
      "entropy": 1.4149091124534607,
      "epoch": 1.008,
      "grad_norm": 22.0,
      "learning_rate": 1.872777777777778e-05,
      "loss": 22.5914,
      "mean_token_accuracy": 0.6590565942227841,
      "num_tokens": 12319366.0,
      "step": 630
    },
    {
      "entropy": 1.3635457135736941,
      "epoch": 1.024,
      "grad_norm": 22.25,
      "learning_rate": 1.8672222222222222e-05,
      "loss": 21.891,
      "mean_token_accuracy": 0.6659590568393469,
      "num_tokens": 12524183.0,
      "step": 640
    },
    {
      "entropy": 1.4019929181784392,
      "epoch": 1.04,
      "grad_norm": 23.25,
      "learning_rate": 1.8616666666666667e-05,
      "loss": 22.5727,
      "mean_token_accuracy": 0.6585873100906611,
      "num_tokens": 12718593.0,
      "step": 650
    },
    {
      "entropy": 1.3632093496620654,
      "epoch": 1.056,
      "grad_norm": 23.375,
      "learning_rate": 1.8561111111111113e-05,
      "loss": 21.9258,
      "mean_token_accuracy": 0.6668245669454336,
      "num_tokens": 12917803.0,
      "step": 660
    },
    {
      "entropy": 1.384382402151823,
      "epoch": 1.072,
      "grad_norm": 25.125,
      "learning_rate": 1.8505555555555556e-05,
      "loss": 22.2729,
      "mean_token_accuracy": 0.661271245405078,
      "num_tokens": 13105826.0,
      "step": 670
    },
    {
      "entropy": 1.3646291557699441,
      "epoch": 1.088,
      "grad_norm": 22.75,
      "learning_rate": 1.845e-05,
      "loss": 22.0115,
      "mean_token_accuracy": 0.6647654321044684,
      "num_tokens": 13298619.0,
      "step": 680
    },
    {
      "entropy": 1.3707259617745877,
      "epoch": 1.104,
      "grad_norm": 21.875,
      "learning_rate": 1.8394444444444447e-05,
      "loss": 22.1735,
      "mean_token_accuracy": 0.6635601755231619,
      "num_tokens": 13491486.0,
      "step": 690
    },
    {
      "entropy": 1.3868375834077598,
      "epoch": 1.12,
      "grad_norm": 25.375,
      "learning_rate": 1.833888888888889e-05,
      "loss": 22.2951,
      "mean_token_accuracy": 0.6626851093024015,
      "num_tokens": 13674663.0,
      "step": 700
    },
    {
      "epoch": 1.12,
      "eval_biology_entropy": 1.3246132278442382,
      "eval_biology_loss": 1.372668981552124,
      "eval_biology_mean_token_accuracy": 0.6674273037910461,
      "eval_biology_num_tokens": 13674663.0,
      "eval_biology_runtime": 47.9969,
      "eval_biology_samples_per_second": 10.417,
      "eval_biology_steps_per_second": 2.604,
      "step": 700
    },
    {
      "epoch": 1.12,
      "eval_chemistry_entropy": 1.0691780796051025,
      "eval_chemistry_loss": 1.127761960029602,
      "eval_chemistry_mean_token_accuracy": 0.7181085128784179,
      "eval_chemistry_num_tokens": 13674663.0,
      "eval_chemistry_runtime": 59.4472,
      "eval_chemistry_samples_per_second": 8.411,
      "eval_chemistry_steps_per_second": 2.103,
      "step": 700
    },
    {
      "epoch": 1.12,
      "eval_math_entropy": 0.99389643907547,
      "eval_math_loss": 1.3389809131622314,
      "eval_math_mean_token_accuracy": 0.6951848978996277,
      "eval_math_num_tokens": 13674663.0,
      "eval_math_runtime": 60.9994,
      "eval_math_samples_per_second": 8.197,
      "eval_math_steps_per_second": 2.049,
      "step": 700
    },
    {
      "epoch": 1.12,
      "eval_physics_entropy": 1.1285274119377136,
      "eval_physics_loss": 1.2335925102233887,
      "eval_physics_mean_token_accuracy": 0.7106850080490112,
      "eval_physics_num_tokens": 13674663.0,
      "eval_physics_runtime": 69.3966,
      "eval_physics_samples_per_second": 7.205,
      "eval_physics_steps_per_second": 1.801,
      "step": 700
    },
    {
      "entropy": 1.3608252748847007,
      "epoch": 1.1360000000000001,
      "grad_norm": 25.375,
      "learning_rate": 1.8283333333333335e-05,
      "loss": 22.0408,
      "mean_token_accuracy": 0.6657578565180302,
      "num_tokens": 13869134.0,
      "step": 710
    },
    {
      "entropy": 1.3435391496866942,
      "epoch": 1.152,
      "grad_norm": 20.75,
      "learning_rate": 1.822777777777778e-05,
      "loss": 21.6423,
      "mean_token_accuracy": 0.6683840617537499,
      "num_tokens": 14078365.0,
      "step": 720
    },
    {
      "entropy": 1.3709587834775447,
      "epoch": 1.168,
      "grad_norm": 21.75,
      "learning_rate": 1.8172222222222223e-05,
      "loss": 21.9547,
      "mean_token_accuracy": 0.6655889604240656,
      "num_tokens": 14266831.0,
      "step": 730
    },
    {
      "entropy": 1.35636080019176,
      "epoch": 1.184,
      "grad_norm": 23.375,
      "learning_rate": 1.811666666666667e-05,
      "loss": 21.9976,
      "mean_token_accuracy": 0.6647399220615625,
      "num_tokens": 14465660.0,
      "step": 740
    },
    {
      "entropy": 1.3615092240273952,
      "epoch": 1.2,
      "grad_norm": 23.5,
      "learning_rate": 1.806111111111111e-05,
      "loss": 21.7382,
      "mean_token_accuracy": 0.6674605444073677,
      "num_tokens": 14653228.0,
      "step": 750
    },
    {
      "entropy": 1.3184621524065734,
      "epoch": 1.216,
      "grad_norm": 22.875,
      "learning_rate": 1.8005555555555557e-05,
      "loss": 21.2498,
      "mean_token_accuracy": 0.6712751515209675,
      "num_tokens": 14857782.0,
      "step": 760
    },
    {
      "entropy": 1.3402907609939576,
      "epoch": 1.232,
      "grad_norm": 21.25,
      "learning_rate": 1.795e-05,
      "loss": 21.5866,
      "mean_token_accuracy": 0.6716990198940038,
      "num_tokens": 15047356.0,
      "step": 770
    },
    {
      "entropy": 1.3448184553533793,
      "epoch": 1.248,
      "grad_norm": 22.125,
      "learning_rate": 1.7894444444444445e-05,
      "loss": 21.7789,
      "mean_token_accuracy": 0.666607766225934,
      "num_tokens": 15241098.0,
      "step": 780
    },
    {
      "entropy": 1.3665424205362797,
      "epoch": 1.264,
      "grad_norm": 21.625,
      "learning_rate": 1.783888888888889e-05,
      "loss": 21.9676,
      "mean_token_accuracy": 0.6660240691155195,
      "num_tokens": 15437657.0,
      "step": 790
    },
    {
      "entropy": 1.3752684637904167,
      "epoch": 1.28,
      "grad_norm": 24.25,
      "learning_rate": 1.7783333333333333e-05,
      "loss": 22.1845,
      "mean_token_accuracy": 0.663004144653678,
      "num_tokens": 15630795.0,
      "step": 800
    },
    {
      "epoch": 1.28,
      "eval_biology_entropy": 1.3029982619285583,
      "eval_biology_loss": 1.3570163249969482,
      "eval_biology_mean_token_accuracy": 0.6696099963188171,
      "eval_biology_num_tokens": 15630795.0,
      "eval_biology_runtime": 47.881,
      "eval_biology_samples_per_second": 10.443,
      "eval_biology_steps_per_second": 2.611,
      "step": 800
    },
    {
      "epoch": 1.28,
      "eval_chemistry_entropy": 1.0498044610023498,
      "eval_chemistry_loss": 1.1165025234222412,
      "eval_chemistry_mean_token_accuracy": 0.7205243277549743,
      "eval_chemistry_num_tokens": 15630795.0,
      "eval_chemistry_runtime": 59.37,
      "eval_chemistry_samples_per_second": 8.422,
      "eval_chemistry_steps_per_second": 2.105,
      "step": 800
    },
    {
      "epoch": 1.28,
      "eval_math_entropy": 0.9808836612701416,
      "eval_math_loss": 1.3316479921340942,
      "eval_math_mean_token_accuracy": 0.6963300676345825,
      "eval_math_num_tokens": 15630795.0,
      "eval_math_runtime": 60.9288,
      "eval_math_samples_per_second": 8.206,
      "eval_math_steps_per_second": 2.052,
      "step": 800
    },
    {
      "epoch": 1.28,
      "eval_physics_entropy": 1.105515335559845,
      "eval_physics_loss": 1.2220377922058105,
      "eval_physics_mean_token_accuracy": 0.7136437473297119,
      "eval_physics_num_tokens": 15630795.0,
      "eval_physics_runtime": 69.455,
      "eval_physics_samples_per_second": 7.199,
      "eval_physics_steps_per_second": 1.8,
      "step": 800
    },
    {
      "entropy": 1.3728813920170069,
      "epoch": 1.296,
      "grad_norm": 22.0,
      "learning_rate": 1.772777777777778e-05,
      "loss": 22.0803,
      "mean_token_accuracy": 0.6633372567594051,
      "num_tokens": 15827105.0,
      "step": 810
    },
    {
      "entropy": 1.3588534504175187,
      "epoch": 1.312,
      "grad_norm": 22.75,
      "learning_rate": 1.7672222222222224e-05,
      "loss": 21.9941,
      "mean_token_accuracy": 0.6649123594164849,
      "num_tokens": 16019645.0,
      "step": 820
    },
    {
      "entropy": 1.3583826199173927,
      "epoch": 1.328,
      "grad_norm": 22.5,
      "learning_rate": 1.761666666666667e-05,
      "loss": 22.0033,
      "mean_token_accuracy": 0.6650775659829378,
      "num_tokens": 16221726.0,
      "step": 830
    },
    {
      "entropy": 1.3320345506072044,
      "epoch": 1.3439999999999999,
      "grad_norm": 21.625,
      "learning_rate": 1.7561111111111113e-05,
      "loss": 21.3812,
      "mean_token_accuracy": 0.6717754557728768,
      "num_tokens": 16427594.0,
      "step": 840
    },
    {
      "entropy": 1.3404277559369802,
      "epoch": 1.3599999999999999,
      "grad_norm": 23.125,
      "learning_rate": 1.7505555555555558e-05,
      "loss": 21.7611,
      "mean_token_accuracy": 0.6681062959134578,
      "num_tokens": 16621605.0,
      "step": 850
    },
    {
      "entropy": 1.3873565673828125,
      "epoch": 1.376,
      "grad_norm": 24.375,
      "learning_rate": 1.7450000000000004e-05,
      "loss": 22.2736,
      "mean_token_accuracy": 0.6620242651551962,
      "num_tokens": 16813444.0,
      "step": 860
    },
    {
      "entropy": 1.350209017843008,
      "epoch": 1.392,
      "grad_norm": 21.75,
      "learning_rate": 1.7394444444444446e-05,
      "loss": 21.8414,
      "mean_token_accuracy": 0.6684158366173506,
      "num_tokens": 17006509.0,
      "step": 870
    },
    {
      "entropy": 1.3420262347906828,
      "epoch": 1.408,
      "grad_norm": 19.75,
      "learning_rate": 1.733888888888889e-05,
      "loss": 21.5337,
      "mean_token_accuracy": 0.6690315004438162,
      "num_tokens": 17197870.0,
      "step": 880
    },
    {
      "entropy": 1.350721249729395,
      "epoch": 1.424,
      "grad_norm": 22.875,
      "learning_rate": 1.7283333333333334e-05,
      "loss": 21.7077,
      "mean_token_accuracy": 0.6683476880192757,
      "num_tokens": 17394390.0,
      "step": 890
    },
    {
      "entropy": 1.3492400590330362,
      "epoch": 1.44,
      "grad_norm": 26.625,
      "learning_rate": 1.722777777777778e-05,
      "loss": 21.7337,
      "mean_token_accuracy": 0.6685651421546936,
      "num_tokens": 17587777.0,
      "step": 900
    },
    {
      "epoch": 1.44,
      "eval_biology_entropy": 1.3251069784164429,
      "eval_biology_loss": 1.3422290086746216,
      "eval_biology_mean_token_accuracy": 0.6725588870048523,
      "eval_biology_num_tokens": 17587777.0,
      "eval_biology_runtime": 47.7727,
      "eval_biology_samples_per_second": 10.466,
      "eval_biology_steps_per_second": 2.617,
      "step": 900
    },
    {
      "epoch": 1.44,
      "eval_chemistry_entropy": 1.0663507118225097,
      "eval_chemistry_loss": 1.1076302528381348,
      "eval_chemistry_mean_token_accuracy": 0.7229596424102783,
      "eval_chemistry_num_tokens": 17587777.0,
      "eval_chemistry_runtime": 59.2788,
      "eval_chemistry_samples_per_second": 8.435,
      "eval_chemistry_steps_per_second": 2.109,
      "step": 900
    },
    {
      "epoch": 1.44,
      "eval_math_entropy": 0.9817979979515076,
      "eval_math_loss": 1.3171546459197998,
      "eval_math_mean_token_accuracy": 0.69857652759552,
      "eval_math_num_tokens": 17587777.0,
      "eval_math_runtime": 60.7675,
      "eval_math_samples_per_second": 8.228,
      "eval_math_steps_per_second": 2.057,
      "step": 900
    },
    {
      "epoch": 1.44,
      "eval_physics_entropy": 1.1204471597671508,
      "eval_physics_loss": 1.2109025716781616,
      "eval_physics_mean_token_accuracy": 0.7157044930458069,
      "eval_physics_num_tokens": 17587777.0,
      "eval_physics_runtime": 69.3094,
      "eval_physics_samples_per_second": 7.214,
      "eval_physics_steps_per_second": 1.804,
      "step": 900
    },
    {
      "entropy": 1.3289990171790123,
      "epoch": 1.456,
      "grad_norm": 22.625,
      "learning_rate": 1.7172222222222223e-05,
      "loss": 21.5738,
      "mean_token_accuracy": 0.6697568029165268,
      "num_tokens": 17788456.0,
      "step": 910
    },
    {
      "entropy": 1.3400883872061968,
      "epoch": 1.472,
      "grad_norm": 21.0,
      "learning_rate": 1.7116666666666668e-05,
      "loss": 21.477,
      "mean_token_accuracy": 0.671789263933897,
      "num_tokens": 17984063.0,
      "step": 920
    },
    {
      "entropy": 1.329145884141326,
      "epoch": 1.488,
      "grad_norm": 22.125,
      "learning_rate": 1.7061111111111114e-05,
      "loss": 21.5347,
      "mean_token_accuracy": 0.672481220588088,
      "num_tokens": 18175640.0,
      "step": 930
    },
    {
      "entropy": 1.3450588792562486,
      "epoch": 1.504,
      "grad_norm": 21.625,
      "learning_rate": 1.7005555555555556e-05,
      "loss": 21.7376,
      "mean_token_accuracy": 0.6689606472849846,
      "num_tokens": 18367857.0,
      "step": 940
    },
    {
      "entropy": 1.334338391199708,
      "epoch": 1.52,
      "grad_norm": 21.125,
      "learning_rate": 1.6950000000000002e-05,
      "loss": 21.5544,
      "mean_token_accuracy": 0.6698691464960576,
      "num_tokens": 18569146.0,
      "step": 950
    },
    {
      "entropy": 1.3566543370485307,
      "epoch": 1.536,
      "grad_norm": 24.375,
      "learning_rate": 1.6894444444444448e-05,
      "loss": 21.8168,
      "mean_token_accuracy": 0.6673684451729059,
      "num_tokens": 18755079.0,
      "step": 960
    },
    {
      "entropy": 1.3215898990631103,
      "epoch": 1.552,
      "grad_norm": 23.5,
      "learning_rate": 1.683888888888889e-05,
      "loss": 21.3841,
      "mean_token_accuracy": 0.6729550462216138,
      "num_tokens": 18956248.0,
      "step": 970
    },
    {
      "entropy": 1.3184372022747994,
      "epoch": 1.568,
      "grad_norm": 24.25,
      "learning_rate": 1.6783333333333332e-05,
      "loss": 21.2336,
      "mean_token_accuracy": 0.6743004772812128,
      "num_tokens": 19150315.0,
      "step": 980
    },
    {
      "entropy": 1.3297088388353586,
      "epoch": 1.584,
      "grad_norm": 23.125,
      "learning_rate": 1.6727777777777778e-05,
      "loss": 21.35,
      "mean_token_accuracy": 0.673252772539854,
      "num_tokens": 19344260.0,
      "step": 990
    },
    {
      "entropy": 1.356522634997964,
      "epoch": 1.6,
      "grad_norm": 22.0,
      "learning_rate": 1.6672222222222224e-05,
      "loss": 21.8441,
      "mean_token_accuracy": 0.6669803854078055,
      "num_tokens": 19532552.0,
      "step": 1000
    },
    {
      "epoch": 1.6,
      "eval_biology_entropy": 1.2830810399055481,
      "eval_biology_loss": 1.3320355415344238,
      "eval_biology_mean_token_accuracy": 0.6747092785835266,
      "eval_biology_num_tokens": 19532552.0,
      "eval_biology_runtime": 47.7764,
      "eval_biology_samples_per_second": 10.465,
      "eval_biology_steps_per_second": 2.616,
      "step": 1000
    },
    {
      "epoch": 1.6,
      "eval_chemistry_entropy": 1.0297417635917663,
      "eval_chemistry_loss": 1.0974704027175903,
      "eval_chemistry_mean_token_accuracy": 0.725460883140564,
      "eval_chemistry_num_tokens": 19532552.0,
      "eval_chemistry_runtime": 59.2295,
      "eval_chemistry_samples_per_second": 8.442,
      "eval_chemistry_steps_per_second": 2.11,
      "step": 1000
    },
    {
      "epoch": 1.6,
      "eval_math_entropy": 0.954737386226654,
      "eval_math_loss": 1.3101767301559448,
      "eval_math_mean_token_accuracy": 0.7013816771507263,
      "eval_math_num_tokens": 19532552.0,
      "eval_math_runtime": 60.8529,
      "eval_math_samples_per_second": 8.217,
      "eval_math_steps_per_second": 2.054,
      "step": 1000
    },
    {
      "epoch": 1.6,
      "eval_physics_entropy": 1.0814638199806212,
      "eval_physics_loss": 1.2015507221221924,
      "eval_physics_mean_token_accuracy": 0.7181381497383118,
      "eval_physics_num_tokens": 19532552.0,
      "eval_physics_runtime": 69.3718,
      "eval_physics_samples_per_second": 7.208,
      "eval_physics_steps_per_second": 1.802,
      "step": 1000
    }
  ],
  "logging_steps": 10,
  "max_steps": 4000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 7,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 2.2020838985302374e+18,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}