{
  "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": 1.0750614512711762,
      "epoch": 0.016,
      "grad_norm": 162.0,
      "learning_rate": 6.000000000000001e-07,
      "loss": 41.6591,
      "mean_token_accuracy": 0.6122298210859298,
      "num_tokens": 338352.0,
      "step": 10
    },
    {
      "entropy": 1.0839035604149103,
      "epoch": 0.032,
      "grad_norm": 155.0,
      "learning_rate": 1.2666666666666669e-06,
      "loss": 41.6397,
      "mean_token_accuracy": 0.6166786661371588,
      "num_tokens": 671193.0,
      "step": 20
    },
    {
      "entropy": 1.0894881710410118,
      "epoch": 0.048,
      "grad_norm": 132.0,
      "learning_rate": 1.9333333333333336e-06,
      "loss": 41.0461,
      "mean_token_accuracy": 0.6147234473377466,
      "num_tokens": 1001008.0,
      "step": 30
    },
    {
      "entropy": 1.138040168955922,
      "epoch": 0.064,
      "grad_norm": 126.0,
      "learning_rate": 2.6e-06,
      "loss": 41.8193,
      "mean_token_accuracy": 0.6091248352080584,
      "num_tokens": 1331208.0,
      "step": 40
    },
    {
      "entropy": 1.1673120155930519,
      "epoch": 0.08,
      "grad_norm": 114.5,
      "learning_rate": 3.266666666666667e-06,
      "loss": 39.7083,
      "mean_token_accuracy": 0.6180085053667426,
      "num_tokens": 1667532.0,
      "step": 50
    },
    {
      "entropy": 1.19383430108428,
      "epoch": 0.096,
      "grad_norm": 91.0,
      "learning_rate": 3.9333333333333335e-06,
      "loss": 37.505,
      "mean_token_accuracy": 0.6295596646144986,
      "num_tokens": 2007176.0,
      "step": 60
    },
    {
      "entropy": 1.354680097848177,
      "epoch": 0.112,
      "grad_norm": 78.5,
      "learning_rate": 4.600000000000001e-06,
      "loss": 37.0959,
      "mean_token_accuracy": 0.6256176400929689,
      "num_tokens": 2342259.0,
      "step": 70
    },
    {
      "entropy": 1.6070489153265952,
      "epoch": 0.128,
      "grad_norm": 55.0,
      "learning_rate": 5.2666666666666665e-06,
      "loss": 36.6385,
      "mean_token_accuracy": 0.6223620869219303,
      "num_tokens": 2663987.0,
      "step": 80
    },
    {
      "entropy": 1.682814510539174,
      "epoch": 0.144,
      "grad_norm": 37.0,
      "learning_rate": 5.933333333333335e-06,
      "loss": 32.9488,
      "mean_token_accuracy": 0.6449975268915296,
      "num_tokens": 2997824.0,
      "step": 90
    },
    {
      "entropy": 1.7247826386243106,
      "epoch": 0.16,
      "grad_norm": 32.0,
      "learning_rate": 6.600000000000001e-06,
      "loss": 30.8315,
      "mean_token_accuracy": 0.6568072270601988,
      "num_tokens": 3330597.0,
      "step": 100
    },
    {
      "epoch": 0.16,
      "eval_biology_entropy": 3.8070496501922606,
      "eval_biology_loss": 4.702436447143555,
      "eval_biology_mean_token_accuracy": 0.3792178168296814,
      "eval_biology_num_tokens": 3330597.0,
      "eval_biology_runtime": 57.2423,
      "eval_biology_samples_per_second": 8.735,
      "eval_biology_steps_per_second": 2.184,
      "step": 100
    },
    {
      "epoch": 0.16,
      "eval_chemistry_entropy": 2.312845594406128,
      "eval_chemistry_loss": 2.5579307079315186,
      "eval_chemistry_mean_token_accuracy": 0.5897436618804932,
      "eval_chemistry_num_tokens": 3330597.0,
      "eval_chemistry_runtime": 71.9394,
      "eval_chemistry_samples_per_second": 6.95,
      "eval_chemistry_steps_per_second": 1.738,
      "step": 100
    },
    {
      "epoch": 0.16,
      "eval_math_entropy": 1.3987004208564757,
      "eval_math_loss": 1.9521548748016357,
      "eval_math_mean_token_accuracy": 0.6628584680557251,
      "eval_math_num_tokens": 3330597.0,
      "eval_math_runtime": 74.5287,
      "eval_math_samples_per_second": 6.709,
      "eval_math_steps_per_second": 1.677,
      "step": 100
    },
    {
      "epoch": 0.16,
      "eval_physics_entropy": 1.725432638168335,
      "eval_physics_loss": 1.8779996633529663,
      "eval_physics_mean_token_accuracy": 0.6590415091514588,
      "eval_physics_num_tokens": 3330597.0,
      "eval_physics_runtime": 85.2266,
      "eval_physics_samples_per_second": 5.867,
      "eval_physics_steps_per_second": 1.467,
      "step": 100
    },
    {
      "entropy": 1.7152024820446967,
      "epoch": 0.176,
      "grad_norm": 26.875,
      "learning_rate": 7.266666666666668e-06,
      "loss": 29.6512,
      "mean_token_accuracy": 0.6657887665554881,
      "num_tokens": 3658264.0,
      "step": 110
    },
    {
      "entropy": 1.622603354603052,
      "epoch": 0.192,
      "grad_norm": 19.375,
      "learning_rate": 7.933333333333334e-06,
      "loss": 28.3031,
      "mean_token_accuracy": 0.6764084205031395,
      "num_tokens": 3995568.0,
      "step": 120
    },
    {
      "entropy": 1.555612001568079,
      "epoch": 0.208,
      "grad_norm": 19.875,
      "learning_rate": 8.6e-06,
      "loss": 26.6387,
      "mean_token_accuracy": 0.6902163974940777,
      "num_tokens": 4321436.0,
      "step": 130
    },
    {
      "entropy": 1.512053081765771,
      "epoch": 0.224,
      "grad_norm": 17.875,
      "learning_rate": 9.266666666666667e-06,
      "loss": 26.0858,
      "mean_token_accuracy": 0.694270471855998,
      "num_tokens": 4648491.0,
      "step": 140
    },
    {
      "entropy": 1.4869094472378492,
      "epoch": 0.24,
      "grad_norm": 19.75,
      "learning_rate": 9.933333333333334e-06,
      "loss": 25.4417,
      "mean_token_accuracy": 0.6989999961107969,
      "num_tokens": 4986175.0,
      "step": 150
    },
    {
      "entropy": 1.4632187210023404,
      "epoch": 0.256,
      "grad_norm": 19.375,
      "learning_rate": 1.0600000000000002e-05,
      "loss": 25.0091,
      "mean_token_accuracy": 0.7029873106628657,
      "num_tokens": 5329320.0,
      "step": 160
    },
    {
      "entropy": 1.4810095187276602,
      "epoch": 0.272,
      "grad_norm": 20.25,
      "learning_rate": 1.1266666666666668e-05,
      "loss": 25.2607,
      "mean_token_accuracy": 0.7004018183797598,
      "num_tokens": 5658796.0,
      "step": 170
    },
    {
      "entropy": 1.4340823762118817,
      "epoch": 0.288,
      "grad_norm": 18.25,
      "learning_rate": 1.1933333333333335e-05,
      "loss": 24.6883,
      "mean_token_accuracy": 0.7079170107841491,
      "num_tokens": 5980201.0,
      "step": 180
    },
    {
      "entropy": 1.4303428437560797,
      "epoch": 0.304,
      "grad_norm": 22.125,
      "learning_rate": 1.2600000000000001e-05,
      "loss": 24.4539,
      "mean_token_accuracy": 0.7080751962959766,
      "num_tokens": 6298858.0,
      "step": 190
    },
    {
      "entropy": 1.399438175559044,
      "epoch": 0.32,
      "grad_norm": 18.875,
      "learning_rate": 1.3266666666666668e-05,
      "loss": 23.9145,
      "mean_token_accuracy": 0.7128454186022282,
      "num_tokens": 6622798.0,
      "step": 200
    },
    {
      "epoch": 0.32,
      "eval_biology_entropy": 3.444120632171631,
      "eval_biology_loss": 4.076927661895752,
      "eval_biology_mean_token_accuracy": 0.42938668751716613,
      "eval_biology_num_tokens": 6622798.0,
      "eval_biology_runtime": 57.3705,
      "eval_biology_samples_per_second": 8.715,
      "eval_biology_steps_per_second": 2.179,
      "step": 200
    },
    {
      "epoch": 0.32,
      "eval_chemistry_entropy": 1.8959890480041504,
      "eval_chemistry_loss": 2.08449649810791,
      "eval_chemistry_mean_token_accuracy": 0.6399502665996551,
      "eval_chemistry_num_tokens": 6622798.0,
      "eval_chemistry_runtime": 71.8716,
      "eval_chemistry_samples_per_second": 6.957,
      "eval_chemistry_steps_per_second": 1.739,
      "step": 200
    },
    {
      "epoch": 0.32,
      "eval_math_entropy": 1.2845887546539307,
      "eval_math_loss": 1.6543570756912231,
      "eval_math_mean_token_accuracy": 0.6954182968139648,
      "eval_math_num_tokens": 6622798.0,
      "eval_math_runtime": 74.7507,
      "eval_math_samples_per_second": 6.689,
      "eval_math_steps_per_second": 1.672,
      "step": 200
    },
    {
      "epoch": 0.32,
      "eval_physics_entropy": 1.3866976137161255,
      "eval_physics_loss": 1.4872251749038696,
      "eval_physics_mean_token_accuracy": 0.7125752453804016,
      "eval_physics_num_tokens": 6622798.0,
      "eval_physics_runtime": 85.9206,
      "eval_physics_samples_per_second": 5.819,
      "eval_physics_steps_per_second": 1.455,
      "step": 200
    },
    {
      "entropy": 1.3804268836975098,
      "epoch": 0.336,
      "grad_norm": 18.5,
      "learning_rate": 1.3933333333333334e-05,
      "loss": 23.6643,
      "mean_token_accuracy": 0.714659295603633,
      "num_tokens": 6953611.0,
      "step": 210
    },
    {
      "entropy": 1.3354483414441347,
      "epoch": 0.352,
      "grad_norm": 18.125,
      "learning_rate": 1.46e-05,
      "loss": 22.9062,
      "mean_token_accuracy": 0.722674485296011,
      "num_tokens": 7290378.0,
      "step": 220
    },
    {
      "entropy": 1.3261277750134468,
      "epoch": 0.368,
      "grad_norm": 18.5,
      "learning_rate": 1.5266666666666667e-05,
      "loss": 22.7576,
      "mean_token_accuracy": 0.7249834384769202,
      "num_tokens": 7621082.0,
      "step": 230
    },
    {
      "entropy": 1.3795554626733064,
      "epoch": 0.384,
      "grad_norm": 17.875,
      "learning_rate": 1.5933333333333336e-05,
      "loss": 23.5828,
      "mean_token_accuracy": 0.7127871386706829,
      "num_tokens": 7955570.0,
      "step": 240
    },
    {
      "entropy": 1.3256618689745665,
      "epoch": 0.4,
      "grad_norm": 20.0,
      "learning_rate": 1.66e-05,
      "loss": 22.7361,
      "mean_token_accuracy": 0.7222114279866219,
      "num_tokens": 8291049.0,
      "step": 250
    },
    {
      "entropy": 1.298596952483058,
      "epoch": 0.416,
      "grad_norm": 22.0,
      "learning_rate": 1.726666666666667e-05,
      "loss": 22.2416,
      "mean_token_accuracy": 0.7284637104719878,
      "num_tokens": 8627310.0,
      "step": 260
    },
    {
      "entropy": 1.311914747953415,
      "epoch": 0.432,
      "grad_norm": 19.375,
      "learning_rate": 1.7933333333333333e-05,
      "loss": 22.4984,
      "mean_token_accuracy": 0.7248033240437508,
      "num_tokens": 8958371.0,
      "step": 270
    },
    {
      "entropy": 1.2482076751068234,
      "epoch": 0.448,
      "grad_norm": 18.0,
      "learning_rate": 1.86e-05,
      "loss": 21.4775,
      "mean_token_accuracy": 0.7356370355933904,
      "num_tokens": 9294388.0,
      "step": 280
    },
    {
      "entropy": 1.2598353076726199,
      "epoch": 0.464,
      "grad_norm": 20.5,
      "learning_rate": 1.926666666666667e-05,
      "loss": 21.5863,
      "mean_token_accuracy": 0.7334068417549133,
      "num_tokens": 9620366.0,
      "step": 290
    },
    {
      "entropy": 1.2661810796707869,
      "epoch": 0.48,
      "grad_norm": 19.5,
      "learning_rate": 1.9933333333333334e-05,
      "loss": 21.8336,
      "mean_token_accuracy": 0.7298611674457789,
      "num_tokens": 9955431.0,
      "step": 300
    },
    {
      "epoch": 0.48,
      "eval_biology_entropy": 3.3976770553588866,
      "eval_biology_loss": 3.904201030731201,
      "eval_biology_mean_token_accuracy": 0.43922377967834475,
      "eval_biology_num_tokens": 9955431.0,
      "eval_biology_runtime": 57.4472,
      "eval_biology_samples_per_second": 8.704,
      "eval_biology_steps_per_second": 2.176,
      "step": 300
    },
    {
      "epoch": 0.48,
      "eval_chemistry_entropy": 1.8206102285385133,
      "eval_chemistry_loss": 1.956045389175415,
      "eval_chemistry_mean_token_accuracy": 0.6531638257503509,
      "eval_chemistry_num_tokens": 9955431.0,
      "eval_chemistry_runtime": 71.9347,
      "eval_chemistry_samples_per_second": 6.951,
      "eval_chemistry_steps_per_second": 1.738,
      "step": 300
    },
    {
      "epoch": 0.48,
      "eval_math_entropy": 1.2509758601188659,
      "eval_math_loss": 1.5806307792663574,
      "eval_math_mean_token_accuracy": 0.7045728435516357,
      "eval_math_num_tokens": 9955431.0,
      "eval_math_runtime": 75.0056,
      "eval_math_samples_per_second": 6.666,
      "eval_math_steps_per_second": 1.667,
      "step": 300
    },
    {
      "epoch": 0.48,
      "eval_physics_entropy": 1.304166310787201,
      "eval_physics_loss": 1.3740965127944946,
      "eval_physics_mean_token_accuracy": 0.7282451548576355,
      "eval_physics_num_tokens": 9955431.0,
      "eval_physics_runtime": 85.2346,
      "eval_physics_samples_per_second": 5.866,
      "eval_physics_steps_per_second": 1.467,
      "step": 300
    },
    {
      "entropy": 1.2894000235944987,
      "epoch": 0.496,
      "grad_norm": 18.875,
      "learning_rate": 1.9933333333333334e-05,
      "loss": 22.125,
      "mean_token_accuracy": 0.7272629871964454,
      "num_tokens": 10289883.0,
      "step": 310
    },
    {
      "entropy": 1.258072443306446,
      "epoch": 0.512,
      "grad_norm": 20.5,
      "learning_rate": 1.985925925925926e-05,
      "loss": 21.7519,
      "mean_token_accuracy": 0.7323984511196613,
      "num_tokens": 10619468.0,
      "step": 320
    },
    {
      "entropy": 1.2734836369752884,
      "epoch": 0.528,
      "grad_norm": 16.375,
      "learning_rate": 1.9785185185185187e-05,
      "loss": 21.9114,
      "mean_token_accuracy": 0.7292184431105853,
      "num_tokens": 10944904.0,
      "step": 330
    },
    {
      "entropy": 1.2483941707760096,
      "epoch": 0.544,
      "grad_norm": 19.25,
      "learning_rate": 1.971111111111111e-05,
      "loss": 21.5727,
      "mean_token_accuracy": 0.7335125673562288,
      "num_tokens": 11266409.0,
      "step": 340
    },
    {
      "entropy": 1.2357387099415065,
      "epoch": 0.56,
      "grad_norm": 16.625,
      "learning_rate": 1.963703703703704e-05,
      "loss": 21.1173,
      "mean_token_accuracy": 0.7381447162479162,
      "num_tokens": 11605544.0,
      "step": 350
    },
    {
      "entropy": 1.2104782849550246,
      "epoch": 0.576,
      "grad_norm": 19.25,
      "learning_rate": 1.9562962962962964e-05,
      "loss": 20.9845,
      "mean_token_accuracy": 0.740512777864933,
      "num_tokens": 11931715.0,
      "step": 360
    },
    {
      "entropy": 1.2183124080300332,
      "epoch": 0.592,
      "grad_norm": 19.375,
      "learning_rate": 1.948888888888889e-05,
      "loss": 21.0928,
      "mean_token_accuracy": 0.7366720300167799,
      "num_tokens": 12257243.0,
      "step": 370
    },
    {
      "entropy": 1.2145392388105392,
      "epoch": 0.608,
      "grad_norm": 16.625,
      "learning_rate": 1.9414814814814817e-05,
      "loss": 20.8598,
      "mean_token_accuracy": 0.740410690754652,
      "num_tokens": 12583154.0,
      "step": 380
    },
    {
      "entropy": 1.2197224445641042,
      "epoch": 0.624,
      "grad_norm": 17.5,
      "learning_rate": 1.9340740740740743e-05,
      "loss": 21.1959,
      "mean_token_accuracy": 0.7364733774214983,
      "num_tokens": 12905392.0,
      "step": 390
    },
    {
      "entropy": 1.2305814437568188,
      "epoch": 0.64,
      "grad_norm": 18.625,
      "learning_rate": 1.926666666666667e-05,
      "loss": 21.2803,
      "mean_token_accuracy": 0.7369419563561678,
      "num_tokens": 13232198.0,
      "step": 400
    },
    {
      "epoch": 0.64,
      "eval_biology_entropy": 3.2428079328536987,
      "eval_biology_loss": 3.81221866607666,
      "eval_biology_mean_token_accuracy": 0.44962965536117555,
      "eval_biology_num_tokens": 13232198.0,
      "eval_biology_runtime": 57.9921,
      "eval_biology_samples_per_second": 8.622,
      "eval_biology_steps_per_second": 2.155,
      "step": 400
    },
    {
      "epoch": 0.64,
      "eval_chemistry_entropy": 1.71326642370224,
      "eval_chemistry_loss": 1.8857945203781128,
      "eval_chemistry_mean_token_accuracy": 0.6629956634044647,
      "eval_chemistry_num_tokens": 13232198.0,
      "eval_chemistry_runtime": 72.1018,
      "eval_chemistry_samples_per_second": 6.935,
      "eval_chemistry_steps_per_second": 1.734,
      "step": 400
    },
    {
      "epoch": 0.64,
      "eval_math_entropy": 1.1944113321304322,
      "eval_math_loss": 1.5440242290496826,
      "eval_math_mean_token_accuracy": 0.7086545734405517,
      "eval_math_num_tokens": 13232198.0,
      "eval_math_runtime": 75.1501,
      "eval_math_samples_per_second": 6.653,
      "eval_math_steps_per_second": 1.663,
      "step": 400
    },
    {
      "epoch": 0.64,
      "eval_physics_entropy": 1.211931779384613,
      "eval_physics_loss": 1.3088711500167847,
      "eval_physics_mean_token_accuracy": 0.737203718662262,
      "eval_physics_num_tokens": 13232198.0,
      "eval_physics_runtime": 85.7441,
      "eval_physics_samples_per_second": 5.831,
      "eval_physics_steps_per_second": 1.458,
      "step": 400
    },
    {
      "entropy": 1.2212170030921699,
      "epoch": 0.656,
      "grad_norm": 17.875,
      "learning_rate": 1.9192592592592593e-05,
      "loss": 20.9332,
      "mean_token_accuracy": 0.7387504555284977,
      "num_tokens": 13575902.0,
      "step": 410
    },
    {
      "entropy": 1.1882116444408894,
      "epoch": 0.672,
      "grad_norm": 20.25,
      "learning_rate": 1.911851851851852e-05,
      "loss": 20.6082,
      "mean_token_accuracy": 0.743332651630044,
      "num_tokens": 13895997.0,
      "step": 420
    },
    {
      "entropy": 1.207483420521021,
      "epoch": 0.688,
      "grad_norm": 18.0,
      "learning_rate": 1.9044444444444446e-05,
      "loss": 20.654,
      "mean_token_accuracy": 0.7427127715200186,
      "num_tokens": 14234888.0,
      "step": 430
    },
    {
      "entropy": 1.2052442319691181,
      "epoch": 0.704,
      "grad_norm": 19.0,
      "learning_rate": 1.8970370370370372e-05,
      "loss": 20.8893,
      "mean_token_accuracy": 0.7387381002306939,
      "num_tokens": 14567908.0,
      "step": 440
    },
    {
      "entropy": 1.181990123167634,
      "epoch": 0.72,
      "grad_norm": 21.625,
      "learning_rate": 1.8896296296296295e-05,
      "loss": 20.4926,
      "mean_token_accuracy": 0.7439979754388333,
      "num_tokens": 14882927.0,
      "step": 450
    },
    {
      "entropy": 1.1967087093740703,
      "epoch": 0.736,
      "grad_norm": 20.375,
      "learning_rate": 1.8822222222222225e-05,
      "loss": 20.5403,
      "mean_token_accuracy": 0.7425415780395269,
      "num_tokens": 15217342.0,
      "step": 460
    },
    {
      "entropy": 1.1514527130872012,
      "epoch": 0.752,
      "grad_norm": 16.75,
      "learning_rate": 1.874814814814815e-05,
      "loss": 19.9152,
      "mean_token_accuracy": 0.7488316483795643,
      "num_tokens": 15547710.0,
      "step": 470
    },
    {
      "entropy": 1.2158014349639417,
      "epoch": 0.768,
      "grad_norm": 21.375,
      "learning_rate": 1.8674074074074075e-05,
      "loss": 21.052,
      "mean_token_accuracy": 0.7373728647828102,
      "num_tokens": 15877177.0,
      "step": 480
    },
    {
      "entropy": 1.1517657596617936,
      "epoch": 0.784,
      "grad_norm": 16.125,
      "learning_rate": 1.86e-05,
      "loss": 19.9164,
      "mean_token_accuracy": 0.7503607623279095,
      "num_tokens": 16219640.0,
      "step": 490
    },
    {
      "entropy": 1.180898468568921,
      "epoch": 0.8,
      "grad_norm": 21.75,
      "learning_rate": 1.8525925925925928e-05,
      "loss": 20.3875,
      "mean_token_accuracy": 0.7441588096320629,
      "num_tokens": 16548261.0,
      "step": 500
    },
    {
      "epoch": 0.8,
      "eval_biology_entropy": 3.2802926797866823,
      "eval_biology_loss": 3.785775899887085,
      "eval_biology_mean_token_accuracy": 0.4522908911705017,
      "eval_biology_num_tokens": 16548261.0,
      "eval_biology_runtime": 58.1484,
      "eval_biology_samples_per_second": 8.599,
      "eval_biology_steps_per_second": 2.15,
      "step": 500
    },
    {
      "epoch": 0.8,
      "eval_chemistry_entropy": 1.712562572002411,
      "eval_chemistry_loss": 1.8553112745285034,
      "eval_chemistry_mean_token_accuracy": 0.6659346613883972,
      "eval_chemistry_num_tokens": 16548261.0,
      "eval_chemistry_runtime": 71.5095,
      "eval_chemistry_samples_per_second": 6.992,
      "eval_chemistry_steps_per_second": 1.748,
      "step": 500
    },
    {
      "epoch": 0.8,
      "eval_math_entropy": 1.1794157824516296,
      "eval_math_loss": 1.5209077596664429,
      "eval_math_mean_token_accuracy": 0.7121200709342956,
      "eval_math_num_tokens": 16548261.0,
      "eval_math_runtime": 74.5115,
      "eval_math_samples_per_second": 6.71,
      "eval_math_steps_per_second": 1.678,
      "step": 500
    },
    {
      "epoch": 0.8,
      "eval_physics_entropy": 1.1884380350112915,
      "eval_physics_loss": 1.2718405723571777,
      "eval_physics_mean_token_accuracy": 0.742140392780304,
      "eval_physics_num_tokens": 16548261.0,
      "eval_physics_runtime": 85.7525,
      "eval_physics_samples_per_second": 5.831,
      "eval_physics_steps_per_second": 1.458,
      "step": 500
    },
    {
      "entropy": 1.167030330747366,
      "epoch": 0.816,
      "grad_norm": 18.625,
      "learning_rate": 1.8451851851851855e-05,
      "loss": 20.2717,
      "mean_token_accuracy": 0.7453148473054171,
      "num_tokens": 16870533.0,
      "step": 510
    },
    {
      "entropy": 1.1830362804234027,
      "epoch": 0.832,
      "grad_norm": 21.25,
      "learning_rate": 1.8377777777777778e-05,
      "loss": 20.3805,
      "mean_token_accuracy": 0.743823428824544,
      "num_tokens": 17196371.0,
      "step": 520
    },
    {
      "entropy": 1.1646822672337294,
      "epoch": 0.848,
      "grad_norm": 16.375,
      "learning_rate": 1.8303703703703704e-05,
      "loss": 20.0962,
      "mean_token_accuracy": 0.7472099754959345,
      "num_tokens": 17527066.0,
      "step": 530
    },
    {
      "entropy": 1.1752454269677401,
      "epoch": 0.864,
      "grad_norm": 17.875,
      "learning_rate": 1.822962962962963e-05,
      "loss": 20.2686,
      "mean_token_accuracy": 0.7448990754783154,
      "num_tokens": 17859605.0,
      "step": 540
    },
    {
      "entropy": 1.1999774873256683,
      "epoch": 0.88,
      "grad_norm": 20.25,
      "learning_rate": 1.8155555555555557e-05,
      "loss": 20.6029,
      "mean_token_accuracy": 0.7425056725740433,
      "num_tokens": 18187315.0,
      "step": 550
    },
    {
      "entropy": 1.1115015875548124,
      "epoch": 0.896,
      "grad_norm": 16.75,
      "learning_rate": 1.8081481481481484e-05,
      "loss": 19.1904,
      "mean_token_accuracy": 0.755654676631093,
      "num_tokens": 18536790.0,
      "step": 560
    },
    {
      "entropy": 1.1791554648429154,
      "epoch": 0.912,
      "grad_norm": 19.5,
      "learning_rate": 1.800740740740741e-05,
      "loss": 20.3039,
      "mean_token_accuracy": 0.7430271036922932,
      "num_tokens": 18869344.0,
      "step": 570
    },
    {
      "entropy": 1.160981795936823,
      "epoch": 0.928,
      "grad_norm": 24.0,
      "learning_rate": 1.7933333333333333e-05,
      "loss": 20.2025,
      "mean_token_accuracy": 0.7445277851074934,
      "num_tokens": 19199834.0,
      "step": 580
    },
    {
      "entropy": 1.1458067331463098,
      "epoch": 0.944,
      "grad_norm": 20.875,
      "learning_rate": 1.785925925925926e-05,
      "loss": 19.7237,
      "mean_token_accuracy": 0.7498494196683169,
      "num_tokens": 19525385.0,
      "step": 590
    },
    {
      "entropy": 1.1434341490268707,
      "epoch": 0.96,
      "grad_norm": 22.0,
      "learning_rate": 1.7785185185185186e-05,
      "loss": 19.8511,
      "mean_token_accuracy": 0.7467116083949804,
      "num_tokens": 19856973.0,
      "step": 600
    },
    {
      "epoch": 0.96,
      "eval_biology_entropy": 3.09391410446167,
      "eval_biology_loss": 3.76809024810791,
      "eval_biology_mean_token_accuracy": 0.4543556985855103,
      "eval_biology_num_tokens": 19856973.0,
      "eval_biology_runtime": 57.6512,
      "eval_biology_samples_per_second": 8.673,
      "eval_biology_steps_per_second": 2.168,
      "step": 600
    },
    {
      "epoch": 0.96,
      "eval_chemistry_entropy": 1.6207774982452392,
      "eval_chemistry_loss": 1.829699993133545,
      "eval_chemistry_mean_token_accuracy": 0.6696915452480316,
      "eval_chemistry_num_tokens": 19856973.0,
      "eval_chemistry_runtime": 71.7673,
      "eval_chemistry_samples_per_second": 6.967,
      "eval_chemistry_steps_per_second": 1.742,
      "step": 600
    },
    {
      "epoch": 0.96,
      "eval_math_entropy": 1.1423851866722108,
      "eval_math_loss": 1.512243390083313,
      "eval_math_mean_token_accuracy": 0.7139626431465149,
      "eval_math_num_tokens": 19856973.0,
      "eval_math_runtime": 74.6709,
      "eval_math_samples_per_second": 6.696,
      "eval_math_steps_per_second": 1.674,
      "step": 600
    },
    {
      "epoch": 0.96,
      "eval_physics_entropy": 1.139296302318573,
      "eval_physics_loss": 1.2478581666946411,
      "eval_physics_mean_token_accuracy": 0.7462703113555909,
      "eval_physics_num_tokens": 19856973.0,
      "eval_physics_runtime": 85.2751,
      "eval_physics_samples_per_second": 5.863,
      "eval_physics_steps_per_second": 1.466,
      "step": 600
    },
    {
      "entropy": 1.147014268860221,
      "epoch": 0.976,
      "grad_norm": 20.875,
      "learning_rate": 1.7711111111111113e-05,
      "loss": 19.8475,
      "mean_token_accuracy": 0.7499787386506795,
      "num_tokens": 20190600.0,
      "step": 610
    },
    {
      "entropy": 1.1680770445615054,
      "epoch": 0.992,
      "grad_norm": 19.75,
      "learning_rate": 1.763703703703704e-05,
      "loss": 20.112,
      "mean_token_accuracy": 0.7469341777265072,
      "num_tokens": 20518014.0,
      "step": 620
    },
    {
      "entropy": 1.1299722768366336,
      "epoch": 1.008,
      "grad_norm": 20.375,
      "learning_rate": 1.7562962962962962e-05,
      "loss": 19.3733,
      "mean_token_accuracy": 0.7531220152974129,
      "num_tokens": 20859215.0,
      "step": 630
    },
    {
      "entropy": 1.1239742942154407,
      "epoch": 1.024,
      "grad_norm": 20.25,
      "learning_rate": 1.7488888888888892e-05,
      "loss": 19.6045,
      "mean_token_accuracy": 0.7504096325486899,
      "num_tokens": 21186658.0,
      "step": 640
    },
    {
      "entropy": 1.1457833401858806,
      "epoch": 1.04,
      "grad_norm": 18.125,
      "learning_rate": 1.7414814814814815e-05,
      "loss": 19.5576,
      "mean_token_accuracy": 0.7513953182846308,
      "num_tokens": 21517237.0,
      "step": 650
    },
    {
      "entropy": 1.0998740132898093,
      "epoch": 1.056,
      "grad_norm": 18.625,
      "learning_rate": 1.7340740740740742e-05,
      "loss": 19.212,
      "mean_token_accuracy": 0.7547836307436228,
      "num_tokens": 21852528.0,
      "step": 660
    },
    {
      "entropy": 1.1379598241299391,
      "epoch": 1.072,
      "grad_norm": 16.625,
      "learning_rate": 1.726666666666667e-05,
      "loss": 19.5156,
      "mean_token_accuracy": 0.7515444949269294,
      "num_tokens": 22192380.0,
      "step": 670
    },
    {
      "entropy": 1.1368626669049262,
      "epoch": 1.088,
      "grad_norm": 22.5,
      "learning_rate": 1.7192592592592595e-05,
      "loss": 19.7833,
      "mean_token_accuracy": 0.7488166857510805,
      "num_tokens": 22517971.0,
      "step": 680
    },
    {
      "entropy": 1.1213505759835243,
      "epoch": 1.104,
      "grad_norm": 17.625,
      "learning_rate": 1.711851851851852e-05,
      "loss": 19.3223,
      "mean_token_accuracy": 0.7535126812756061,
      "num_tokens": 22848694.0,
      "step": 690
    },
    {
      "entropy": 1.1373495645821095,
      "epoch": 1.12,
      "grad_norm": 18.125,
      "learning_rate": 1.7044444444444445e-05,
      "loss": 19.7462,
      "mean_token_accuracy": 0.7481394957751035,
      "num_tokens": 23183889.0,
      "step": 700
    },
    {
      "epoch": 1.12,
      "eval_biology_entropy": 3.119599843978882,
      "eval_biology_loss": 3.7430577278137207,
      "eval_biology_mean_token_accuracy": 0.45569487643241885,
      "eval_biology_num_tokens": 23183889.0,
      "eval_biology_runtime": 57.756,
      "eval_biology_samples_per_second": 8.657,
      "eval_biology_steps_per_second": 2.164,
      "step": 700
    },
    {
      "epoch": 1.12,
      "eval_chemistry_entropy": 1.62452432012558,
      "eval_chemistry_loss": 1.812204360961914,
      "eval_chemistry_mean_token_accuracy": 0.67113920378685,
      "eval_chemistry_num_tokens": 23183889.0,
      "eval_chemistry_runtime": 72.0304,
      "eval_chemistry_samples_per_second": 6.942,
      "eval_chemistry_steps_per_second": 1.735,
      "step": 700
    },
    {
      "epoch": 1.12,
      "eval_math_entropy": 1.1421763606071471,
      "eval_math_loss": 1.4993647336959839,
      "eval_math_mean_token_accuracy": 0.715766981601715,
      "eval_math_num_tokens": 23183889.0,
      "eval_math_runtime": 75.1797,
      "eval_math_samples_per_second": 6.651,
      "eval_math_steps_per_second": 1.663,
      "step": 700
    },
    {
      "epoch": 1.12,
      "eval_physics_entropy": 1.137205484867096,
      "eval_physics_loss": 1.2303540706634521,
      "eval_physics_mean_token_accuracy": 0.7483545050621033,
      "eval_physics_num_tokens": 23183889.0,
      "eval_physics_runtime": 85.8421,
      "eval_physics_samples_per_second": 5.825,
      "eval_physics_steps_per_second": 1.456,
      "step": 700
    },
    {
      "entropy": 1.100171994790435,
      "epoch": 1.1360000000000001,
      "grad_norm": 18.375,
      "learning_rate": 1.697037037037037e-05,
      "loss": 18.9149,
      "mean_token_accuracy": 0.7582110617309809,
      "num_tokens": 23523903.0,
      "step": 710
    },
    {
      "entropy": 1.106547114998102,
      "epoch": 1.152,
      "grad_norm": 20.25,
      "learning_rate": 1.6896296296296298e-05,
      "loss": 19.3555,
      "mean_token_accuracy": 0.755762941762805,
      "num_tokens": 23848848.0,
      "step": 720
    },
    {
      "entropy": 1.1538855239748955,
      "epoch": 1.168,
      "grad_norm": 19.375,
      "learning_rate": 1.6822222222222224e-05,
      "loss": 19.7298,
      "mean_token_accuracy": 0.748001191392541,
      "num_tokens": 24174589.0,
      "step": 730
    },
    {
      "entropy": 1.1084443144500256,
      "epoch": 1.184,
      "grad_norm": 21.125,
      "learning_rate": 1.6748148148148147e-05,
      "loss": 19.4071,
      "mean_token_accuracy": 0.7534149281680584,
      "num_tokens": 24495981.0,
      "step": 740
    },
    {
      "entropy": 1.1029753223061562,
      "epoch": 1.2,
      "grad_norm": 18.25,
      "learning_rate": 1.6674074074074077e-05,
      "loss": 18.9968,
      "mean_token_accuracy": 0.7581038322299719,
      "num_tokens": 24826245.0,
      "step": 750
    },
    {
      "entropy": 1.1198973149061202,
      "epoch": 1.216,
      "grad_norm": 22.5,
      "learning_rate": 1.66e-05,
      "loss": 19.4468,
      "mean_token_accuracy": 0.75144985280931,
      "num_tokens": 25147775.0,
      "step": 760
    },
    {
      "entropy": 1.1077666085213422,
      "epoch": 1.232,
      "grad_norm": 21.5,
      "learning_rate": 1.6525925925925927e-05,
      "loss": 19.2235,
      "mean_token_accuracy": 0.7555838622152805,
      "num_tokens": 25480033.0,
      "step": 770
    },
    {
      "entropy": 1.1241751793771981,
      "epoch": 1.248,
      "grad_norm": 21.5,
      "learning_rate": 1.6451851851851853e-05,
      "loss": 19.354,
      "mean_token_accuracy": 0.754213061928749,
      "num_tokens": 25811978.0,
      "step": 780
    },
    {
      "entropy": 1.1038431864231826,
      "epoch": 1.264,
      "grad_norm": 20.75,
      "learning_rate": 1.637777777777778e-05,
      "loss": 19.1684,
      "mean_token_accuracy": 0.7550360012799502,
      "num_tokens": 26146299.0,
      "step": 790
    },
    {
      "entropy": 1.154108800366521,
      "epoch": 1.28,
      "grad_norm": 20.25,
      "learning_rate": 1.6303703703703706e-05,
      "loss": 19.8371,
      "mean_token_accuracy": 0.7468210935592652,
      "num_tokens": 26478207.0,
      "step": 800
    },
    {
      "epoch": 1.28,
      "eval_biology_entropy": 3.1355883588790894,
      "eval_biology_loss": 3.745408773422241,
      "eval_biology_mean_token_accuracy": 0.45794152569770813,
      "eval_biology_num_tokens": 26478207.0,
      "eval_biology_runtime": 57.5193,
      "eval_biology_samples_per_second": 8.693,
      "eval_biology_steps_per_second": 2.173,
      "step": 800
    },
    {
      "epoch": 1.28,
      "eval_chemistry_entropy": 1.6176510167121887,
      "eval_chemistry_loss": 1.7935805320739746,
      "eval_chemistry_mean_token_accuracy": 0.6746818594932557,
      "eval_chemistry_num_tokens": 26478207.0,
      "eval_chemistry_runtime": 72.1166,
      "eval_chemistry_samples_per_second": 6.933,
      "eval_chemistry_steps_per_second": 1.733,
      "step": 800
    },
    {
      "epoch": 1.28,
      "eval_math_entropy": 1.1283973746299745,
      "eval_math_loss": 1.4904797077178955,
      "eval_math_mean_token_accuracy": 0.7178232269287109,
      "eval_math_num_tokens": 26478207.0,
      "eval_math_runtime": 75.0358,
      "eval_math_samples_per_second": 6.663,
      "eval_math_steps_per_second": 1.666,
      "step": 800
    },
    {
      "epoch": 1.28,
      "eval_physics_entropy": 1.1170437216758728,
      "eval_physics_loss": 1.2174808979034424,
      "eval_physics_mean_token_accuracy": 0.7503818984031677,
      "eval_physics_num_tokens": 26478207.0,
      "eval_physics_runtime": 85.6504,
      "eval_physics_samples_per_second": 5.838,
      "eval_physics_steps_per_second": 1.459,
      "step": 800
    },
    {
      "entropy": 1.0819556184113026,
      "epoch": 1.296,
      "grad_norm": 21.875,
      "learning_rate": 1.622962962962963e-05,
      "loss": 18.8799,
      "mean_token_accuracy": 0.759453434124589,
      "num_tokens": 26806477.0,
      "step": 810
    },
    {
      "entropy": 1.1482679098844528,
      "epoch": 1.312,
      "grad_norm": 21.875,
      "learning_rate": 1.6155555555555556e-05,
      "loss": 19.6809,
      "mean_token_accuracy": 0.7498604208230972,
      "num_tokens": 27137528.0,
      "step": 820
    },
    {
      "entropy": 1.1160907682031393,
      "epoch": 1.328,
      "grad_norm": 24.25,
      "learning_rate": 1.6081481481481482e-05,
      "loss": 19.4066,
      "mean_token_accuracy": 0.7520301517099142,
      "num_tokens": 27465859.0,
      "step": 830
    },
    {
      "entropy": 1.1234158642590046,
      "epoch": 1.3439999999999999,
      "grad_norm": 21.875,
      "learning_rate": 1.600740740740741e-05,
      "loss": 19.5899,
      "mean_token_accuracy": 0.7511072501540184,
      "num_tokens": 27795568.0,
      "step": 840
    },
    {
      "entropy": 1.138951735943556,
      "epoch": 1.3599999999999999,
      "grad_norm": 20.875,
      "learning_rate": 1.5933333333333336e-05,
      "loss": 19.5108,
      "mean_token_accuracy": 0.7510471243411303,
      "num_tokens": 28129988.0,
      "step": 850
    },
    {
      "entropy": 1.0953974552452563,
      "epoch": 1.376,
      "grad_norm": 23.625,
      "learning_rate": 1.5859259259259262e-05,
      "loss": 19.0035,
      "mean_token_accuracy": 0.7561734985560179,
      "num_tokens": 28465903.0,
      "step": 860
    },
    {
      "entropy": 1.0946024935692549,
      "epoch": 1.392,
      "grad_norm": 21.375,
      "learning_rate": 1.5785185185185185e-05,
      "loss": 18.8209,
      "mean_token_accuracy": 0.7587534476071596,
      "num_tokens": 28801554.0,
      "step": 870
    },
    {
      "entropy": 1.085593356192112,
      "epoch": 1.408,
      "grad_norm": 19.5,
      "learning_rate": 1.571111111111111e-05,
      "loss": 18.9846,
      "mean_token_accuracy": 0.7570537071675062,
      "num_tokens": 29125114.0,
      "step": 880
    },
    {
      "entropy": 1.1382916759699584,
      "epoch": 1.424,
      "grad_norm": 18.25,
      "learning_rate": 1.5637037037037038e-05,
      "loss": 19.6386,
      "mean_token_accuracy": 0.7486919555813074,
      "num_tokens": 29454334.0,
      "step": 890
    },
    {
      "entropy": 1.0628450240939855,
      "epoch": 1.44,
      "grad_norm": 19.875,
      "learning_rate": 1.5562962962962965e-05,
      "loss": 18.4404,
      "mean_token_accuracy": 0.7623223628848791,
      "num_tokens": 29789253.0,
      "step": 900
    },
    {
      "epoch": 1.44,
      "eval_biology_entropy": 3.1081992044448854,
      "eval_biology_loss": 3.7266616821289062,
      "eval_biology_mean_token_accuracy": 0.45933370518684385,
      "eval_biology_num_tokens": 29789253.0,
      "eval_biology_runtime": 58.0106,
      "eval_biology_samples_per_second": 8.619,
      "eval_biology_steps_per_second": 2.155,
      "step": 900
    },
    {
      "epoch": 1.44,
      "eval_chemistry_entropy": 1.5958426017761231,
      "eval_chemistry_loss": 1.7839316129684448,
      "eval_chemistry_mean_token_accuracy": 0.675512980222702,
      "eval_chemistry_num_tokens": 29789253.0,
      "eval_chemistry_runtime": 71.9545,
      "eval_chemistry_samples_per_second": 6.949,
      "eval_chemistry_steps_per_second": 1.737,
      "step": 900
    },
    {
      "epoch": 1.44,
      "eval_math_entropy": 1.1177569427490235,
      "eval_math_loss": 1.485580563545227,
      "eval_math_mean_token_accuracy": 0.7180696530342102,
      "eval_math_num_tokens": 29789253.0,
      "eval_math_runtime": 75.2492,
      "eval_math_samples_per_second": 6.645,
      "eval_math_steps_per_second": 1.661,
      "step": 900
    },
    {
      "epoch": 1.44,
      "eval_physics_entropy": 1.1015179448127748,
      "eval_physics_loss": 1.20570707321167,
      "eval_physics_mean_token_accuracy": 0.7521858010292053,
      "eval_physics_num_tokens": 29789253.0,
      "eval_physics_runtime": 86.0235,
      "eval_physics_samples_per_second": 5.812,
      "eval_physics_steps_per_second": 1.453,
      "step": 900
    },
    {
      "entropy": 1.0630920186638833,
      "epoch": 1.456,
      "grad_norm": 16.875,
      "learning_rate": 1.548888888888889e-05,
      "loss": 18.4751,
      "mean_token_accuracy": 0.7624705236405134,
      "num_tokens": 30128576.0,
      "step": 910
    },
    {
      "entropy": 1.1466537602245808,
      "epoch": 1.472,
      "grad_norm": 21.625,
      "learning_rate": 1.5414814814814814e-05,
      "loss": 19.847,
      "mean_token_accuracy": 0.7475753095000982,
      "num_tokens": 30452719.0,
      "step": 920
    },
    {
      "entropy": 1.116706547141075,
      "epoch": 1.488,
      "grad_norm": 18.75,
      "learning_rate": 1.5340740740740744e-05,
      "loss": 19.3164,
      "mean_token_accuracy": 0.7528045650571584,
      "num_tokens": 30782343.0,
      "step": 930
    },
    {
      "entropy": 1.0954343430697917,
      "epoch": 1.504,
      "grad_norm": 18.75,
      "learning_rate": 1.5266666666666667e-05,
      "loss": 18.8761,
      "mean_token_accuracy": 0.7587096191942692,
      "num_tokens": 31113238.0,
      "step": 940
    },
    {
      "entropy": 1.1062404930591583,
      "epoch": 1.52,
      "grad_norm": 20.625,
      "learning_rate": 1.5192592592592594e-05,
      "loss": 19.2361,
      "mean_token_accuracy": 0.7530912421643734,
      "num_tokens": 31444589.0,
      "step": 950
    },
    {
      "entropy": 1.1350872155278922,
      "epoch": 1.536,
      "grad_norm": 22.125,
      "learning_rate": 1.5118518518518519e-05,
      "loss": 19.5966,
      "mean_token_accuracy": 0.7506488811224699,
      "num_tokens": 31774137.0,
      "step": 960
    },
    {
      "entropy": 1.0756664339452982,
      "epoch": 1.552,
      "grad_norm": 22.375,
      "learning_rate": 1.5044444444444445e-05,
      "loss": 18.6153,
      "mean_token_accuracy": 0.760128814727068,
      "num_tokens": 32101776.0,
      "step": 970
    },
    {
      "entropy": 1.075929443538189,
      "epoch": 1.568,
      "grad_norm": 20.5,
      "learning_rate": 1.497037037037037e-05,
      "loss": 18.6699,
      "mean_token_accuracy": 0.7611421547830105,
      "num_tokens": 32428021.0,
      "step": 980
    },
    {
      "entropy": 1.1277167603373528,
      "epoch": 1.584,
      "grad_norm": 21.875,
      "learning_rate": 1.4896296296296298e-05,
      "loss": 19.5224,
      "mean_token_accuracy": 0.749526159465313,
      "num_tokens": 32752735.0,
      "step": 990
    },
    {
      "entropy": 1.1047262009233236,
      "epoch": 1.6,
      "grad_norm": 19.375,
      "learning_rate": 1.4822222222222225e-05,
      "loss": 19.3018,
      "mean_token_accuracy": 0.7532294802367687,
      "num_tokens": 33075822.0,
      "step": 1000
    },
    {
      "epoch": 1.6,
      "eval_biology_entropy": 3.1174922523498534,
      "eval_biology_loss": 3.690537691116333,
      "eval_biology_mean_token_accuracy": 0.46170222663879396,
      "eval_biology_num_tokens": 33075822.0,
      "eval_biology_runtime": 57.7123,
      "eval_biology_samples_per_second": 8.664,
      "eval_biology_steps_per_second": 2.166,
      "step": 1000
    },
    {
      "epoch": 1.6,
      "eval_chemistry_entropy": 1.6017069320678712,
      "eval_chemistry_loss": 1.7750109434127808,
      "eval_chemistry_mean_token_accuracy": 0.6766609756946563,
      "eval_chemistry_num_tokens": 33075822.0,
      "eval_chemistry_runtime": 71.5342,
      "eval_chemistry_samples_per_second": 6.99,
      "eval_chemistry_steps_per_second": 1.747,
      "step": 1000
    },
    {
      "epoch": 1.6,
      "eval_math_entropy": 1.1217774868011474,
      "eval_math_loss": 1.4800317287445068,
      "eval_math_mean_token_accuracy": 0.7193166851997376,
      "eval_math_num_tokens": 33075822.0,
      "eval_math_runtime": 75.3052,
      "eval_math_samples_per_second": 6.64,
      "eval_math_steps_per_second": 1.66,
      "step": 1000
    },
    {
      "epoch": 1.6,
      "eval_physics_entropy": 1.0995629110336305,
      "eval_physics_loss": 1.1977922916412354,
      "eval_physics_mean_token_accuracy": 0.7535692925453186,
      "eval_physics_num_tokens": 33075822.0,
      "eval_physics_runtime": 85.6367,
      "eval_physics_samples_per_second": 5.839,
      "eval_physics_steps_per_second": 1.46,
      "step": 1000
    }
  ],
  "logging_steps": 10,
  "max_steps": 3000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 3.3005113980506685e+18,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}