{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.2,
  "eval_steps": 100,
  "global_step": 2000,
  "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
    },
    {
      "entropy": 1.0596268713474273,
      "epoch": 1.616,
      "grad_norm": 20.75,
      "learning_rate": 1.474814814814815e-05,
      "loss": 18.3912,
      "mean_token_accuracy": 0.7637467764317989,
      "num_tokens": 33411850.0,
      "step": 1010
    },
    {
      "entropy": 1.0854996472597123,
      "epoch": 1.6320000000000001,
      "grad_norm": 19.75,
      "learning_rate": 1.4674074074074076e-05,
      "loss": 18.8433,
      "mean_token_accuracy": 0.759448166936636,
      "num_tokens": 33739456.0,
      "step": 1020
    },
    {
      "entropy": 1.0560027658939362,
      "epoch": 1.6480000000000001,
      "grad_norm": 21.375,
      "learning_rate": 1.46e-05,
      "loss": 18.2989,
      "mean_token_accuracy": 0.7635872703045606,
      "num_tokens": 34072022.0,
      "step": 1030
    },
    {
      "entropy": 1.0859289176762104,
      "epoch": 1.6640000000000001,
      "grad_norm": 24.5,
      "learning_rate": 1.4525925925925927e-05,
      "loss": 18.8689,
      "mean_token_accuracy": 0.7598234452307224,
      "num_tokens": 34398959.0,
      "step": 1040
    },
    {
      "entropy": 1.065019664540887,
      "epoch": 1.6800000000000002,
      "grad_norm": 18.875,
      "learning_rate": 1.4451851851851852e-05,
      "loss": 18.3802,
      "mean_token_accuracy": 0.7630509950220585,
      "num_tokens": 34732284.0,
      "step": 1050
    },
    {
      "entropy": 1.0720367562025785,
      "epoch": 1.696,
      "grad_norm": 22.375,
      "learning_rate": 1.4377777777777779e-05,
      "loss": 18.7745,
      "mean_token_accuracy": 0.7593208305537701,
      "num_tokens": 35053712.0,
      "step": 1060
    },
    {
      "entropy": 1.112621883302927,
      "epoch": 1.712,
      "grad_norm": 22.875,
      "learning_rate": 1.4303703703703703e-05,
      "loss": 19.0989,
      "mean_token_accuracy": 0.7542798519134521,
      "num_tokens": 35393088.0,
      "step": 1070
    },
    {
      "entropy": 1.0726498935371638,
      "epoch": 1.728,
      "grad_norm": 21.75,
      "learning_rate": 1.4229629629629632e-05,
      "loss": 18.7565,
      "mean_token_accuracy": 0.7590801615267992,
      "num_tokens": 35729007.0,
      "step": 1080
    },
    {
      "entropy": 1.082063402235508,
      "epoch": 1.744,
      "grad_norm": 20.5,
      "learning_rate": 1.4155555555555556e-05,
      "loss": 18.5636,
      "mean_token_accuracy": 0.7609590470790863,
      "num_tokens": 36072489.0,
      "step": 1090
    },
    {
      "entropy": 1.0809564728289842,
      "epoch": 1.76,
      "grad_norm": 19.25,
      "learning_rate": 1.4081481481481483e-05,
      "loss": 18.6747,
      "mean_token_accuracy": 0.7598331365734339,
      "num_tokens": 36398285.0,
      "step": 1100
    },
    {
      "epoch": 1.76,
      "eval_biology_entropy": 3.0216426620483396,
      "eval_biology_loss": 3.7136104106903076,
      "eval_biology_mean_token_accuracy": 0.4623387033939362,
      "eval_biology_num_tokens": 36398285.0,
      "eval_biology_runtime": 57.9472,
      "eval_biology_samples_per_second": 8.629,
      "eval_biology_steps_per_second": 2.157,
      "step": 1100
    },
    {
      "epoch": 1.76,
      "eval_chemistry_entropy": 1.5544043154716491,
      "eval_chemistry_loss": 1.7709003686904907,
      "eval_chemistry_mean_token_accuracy": 0.6778159828186036,
      "eval_chemistry_num_tokens": 36398285.0,
      "eval_chemistry_runtime": 71.9353,
      "eval_chemistry_samples_per_second": 6.951,
      "eval_chemistry_steps_per_second": 1.738,
      "step": 1100
    },
    {
      "epoch": 1.76,
      "eval_math_entropy": 1.1053983087539674,
      "eval_math_loss": 1.478028655052185,
      "eval_math_mean_token_accuracy": 0.7200283484458924,
      "eval_math_num_tokens": 36398285.0,
      "eval_math_runtime": 75.8155,
      "eval_math_samples_per_second": 6.595,
      "eval_math_steps_per_second": 1.649,
      "step": 1100
    },
    {
      "epoch": 1.76,
      "eval_physics_entropy": 1.0777991194725036,
      "eval_physics_loss": 1.1916557550430298,
      "eval_physics_mean_token_accuracy": 0.7549497890472412,
      "eval_physics_num_tokens": 36398285.0,
      "eval_physics_runtime": 85.2243,
      "eval_physics_samples_per_second": 5.867,
      "eval_physics_steps_per_second": 1.467,
      "step": 1100
    },
    {
      "entropy": 1.114939071238041,
      "epoch": 1.776,
      "grad_norm": 20.5,
      "learning_rate": 1.400740740740741e-05,
      "loss": 19.4472,
      "mean_token_accuracy": 0.7514877334237099,
      "num_tokens": 36730479.0,
      "step": 1110
    },
    {
      "entropy": 1.0648007225245237,
      "epoch": 1.792,
      "grad_norm": 23.5,
      "learning_rate": 1.3933333333333334e-05,
      "loss": 18.5582,
      "mean_token_accuracy": 0.7612718716263771,
      "num_tokens": 37057425.0,
      "step": 1120
    },
    {
      "entropy": 1.0584326587617396,
      "epoch": 1.808,
      "grad_norm": 19.375,
      "learning_rate": 1.385925925925926e-05,
      "loss": 18.2426,
      "mean_token_accuracy": 0.764327023923397,
      "num_tokens": 37395449.0,
      "step": 1130
    },
    {
      "entropy": 1.0700030837208032,
      "epoch": 1.8239999999999998,
      "grad_norm": 19.625,
      "learning_rate": 1.3785185185185186e-05,
      "loss": 18.6209,
      "mean_token_accuracy": 0.7612834881991148,
      "num_tokens": 37729535.0,
      "step": 1140
    },
    {
      "entropy": 1.0742189947515726,
      "epoch": 1.8399999999999999,
      "grad_norm": 21.5,
      "learning_rate": 1.3711111111111112e-05,
      "loss": 18.6079,
      "mean_token_accuracy": 0.7612603317946196,
      "num_tokens": 38057839.0,
      "step": 1150
    },
    {
      "entropy": 1.0742695450782775,
      "epoch": 1.8559999999999999,
      "grad_norm": 21.375,
      "learning_rate": 1.3637037037037037e-05,
      "loss": 18.6078,
      "mean_token_accuracy": 0.7604191195219755,
      "num_tokens": 38398033.0,
      "step": 1160
    },
    {
      "entropy": 1.077582898736,
      "epoch": 1.8719999999999999,
      "grad_norm": 19.0,
      "learning_rate": 1.3562962962962965e-05,
      "loss": 18.6337,
      "mean_token_accuracy": 0.7588818453252315,
      "num_tokens": 38729203.0,
      "step": 1170
    },
    {
      "entropy": 1.0918789152055979,
      "epoch": 1.888,
      "grad_norm": 21.5,
      "learning_rate": 1.3488888888888888e-05,
      "loss": 18.9929,
      "mean_token_accuracy": 0.7574284754693508,
      "num_tokens": 39058791.0,
      "step": 1180
    },
    {
      "entropy": 1.1133687142282724,
      "epoch": 1.904,
      "grad_norm": 19.875,
      "learning_rate": 1.3414814814814817e-05,
      "loss": 19.2692,
      "mean_token_accuracy": 0.7535256497561932,
      "num_tokens": 39386170.0,
      "step": 1190
    },
    {
      "entropy": 1.1121761441230773,
      "epoch": 1.92,
      "grad_norm": 20.625,
      "learning_rate": 1.3340740740740741e-05,
      "loss": 19.3217,
      "mean_token_accuracy": 0.7535242728888989,
      "num_tokens": 39710381.0,
      "step": 1200
    },
    {
      "epoch": 1.92,
      "eval_biology_entropy": 3.068399606704712,
      "eval_biology_loss": 3.697348117828369,
      "eval_biology_mean_token_accuracy": 0.46416072607040404,
      "eval_biology_num_tokens": 39710381.0,
      "eval_biology_runtime": 57.9205,
      "eval_biology_samples_per_second": 8.633,
      "eval_biology_steps_per_second": 2.158,
      "step": 1200
    },
    {
      "epoch": 1.92,
      "eval_chemistry_entropy": 1.5695023431777955,
      "eval_chemistry_loss": 1.7615015506744385,
      "eval_chemistry_mean_token_accuracy": 0.679024073600769,
      "eval_chemistry_num_tokens": 39710381.0,
      "eval_chemistry_runtime": 72.0985,
      "eval_chemistry_samples_per_second": 6.935,
      "eval_chemistry_steps_per_second": 1.734,
      "step": 1200
    },
    {
      "epoch": 1.92,
      "eval_math_entropy": 1.1080369915962218,
      "eval_math_loss": 1.4733725786209106,
      "eval_math_mean_token_accuracy": 0.720601568698883,
      "eval_math_num_tokens": 39710381.0,
      "eval_math_runtime": 76.2304,
      "eval_math_samples_per_second": 6.559,
      "eval_math_steps_per_second": 1.64,
      "step": 1200
    },
    {
      "epoch": 1.92,
      "eval_physics_entropy": 1.0821580514907836,
      "eval_physics_loss": 1.1858406066894531,
      "eval_physics_mean_token_accuracy": 0.7558628296852112,
      "eval_physics_num_tokens": 39710381.0,
      "eval_physics_runtime": 86.5869,
      "eval_physics_samples_per_second": 5.775,
      "eval_physics_steps_per_second": 1.444,
      "step": 1200
    },
    {
      "entropy": 1.080176205933094,
      "epoch": 1.936,
      "grad_norm": 20.125,
      "learning_rate": 1.3266666666666668e-05,
      "loss": 18.7227,
      "mean_token_accuracy": 0.7583677973598242,
      "num_tokens": 40043354.0,
      "step": 1210
    },
    {
      "entropy": 1.0193040519952774,
      "epoch": 1.952,
      "grad_norm": 19.375,
      "learning_rate": 1.3192592592592594e-05,
      "loss": 17.7008,
      "mean_token_accuracy": 0.7700962536036968,
      "num_tokens": 40375790.0,
      "step": 1220
    },
    {
      "entropy": 1.0746948149055242,
      "epoch": 1.968,
      "grad_norm": 20.25,
      "learning_rate": 1.311851851851852e-05,
      "loss": 18.7088,
      "mean_token_accuracy": 0.7597304716706276,
      "num_tokens": 40705451.0,
      "step": 1230
    },
    {
      "entropy": 1.1093663211911917,
      "epoch": 1.984,
      "grad_norm": 22.75,
      "learning_rate": 1.3044444444444446e-05,
      "loss": 19.1019,
      "mean_token_accuracy": 0.7561715345829725,
      "num_tokens": 41040313.0,
      "step": 1240
    },
    {
      "entropy": 1.0923482172191143,
      "epoch": 2.0,
      "grad_norm": 21.125,
      "learning_rate": 1.297037037037037e-05,
      "loss": 18.8945,
      "mean_token_accuracy": 0.7591748807579279,
      "num_tokens": 41373608.0,
      "step": 1250
    },
    {
      "entropy": 1.0565200865268707,
      "epoch": 2.016,
      "grad_norm": 17.75,
      "learning_rate": 1.2896296296296299e-05,
      "loss": 18.208,
      "mean_token_accuracy": 0.7639335673302412,
      "num_tokens": 41708226.0,
      "step": 1260
    },
    {
      "entropy": 1.0461373936384917,
      "epoch": 2.032,
      "grad_norm": 18.25,
      "learning_rate": 1.2822222222222222e-05,
      "loss": 18.2669,
      "mean_token_accuracy": 0.7649948503822088,
      "num_tokens": 42045433.0,
      "step": 1270
    },
    {
      "entropy": 1.102442101202905,
      "epoch": 2.048,
      "grad_norm": 21.75,
      "learning_rate": 1.274814814814815e-05,
      "loss": 19.144,
      "mean_token_accuracy": 0.7540556605905294,
      "num_tokens": 42372694.0,
      "step": 1280
    },
    {
      "entropy": 1.0597637433558702,
      "epoch": 2.064,
      "grad_norm": 19.625,
      "learning_rate": 1.2674074074074075e-05,
      "loss": 18.2938,
      "mean_token_accuracy": 0.7644871883094311,
      "num_tokens": 42710240.0,
      "step": 1290
    },
    {
      "entropy": 1.079553159326315,
      "epoch": 2.08,
      "grad_norm": 20.75,
      "learning_rate": 1.2600000000000001e-05,
      "loss": 18.7297,
      "mean_token_accuracy": 0.7595991253852844,
      "num_tokens": 43037844.0,
      "step": 1300
    },
    {
      "epoch": 2.08,
      "eval_biology_entropy": 3.1228656520843505,
      "eval_biology_loss": 3.680032253265381,
      "eval_biology_mean_token_accuracy": 0.46285973715782164,
      "eval_biology_num_tokens": 43037844.0,
      "eval_biology_runtime": 58.3769,
      "eval_biology_samples_per_second": 8.565,
      "eval_biology_steps_per_second": 2.141,
      "step": 1300
    },
    {
      "epoch": 2.08,
      "eval_chemistry_entropy": 1.5790047402381897,
      "eval_chemistry_loss": 1.7592977285385132,
      "eval_chemistry_mean_token_accuracy": 0.6785329551696777,
      "eval_chemistry_num_tokens": 43037844.0,
      "eval_chemistry_runtime": 72.3659,
      "eval_chemistry_samples_per_second": 6.909,
      "eval_chemistry_steps_per_second": 1.727,
      "step": 1300
    },
    {
      "epoch": 2.08,
      "eval_math_entropy": 1.1097278380393982,
      "eval_math_loss": 1.4702972173690796,
      "eval_math_mean_token_accuracy": 0.7207047066688538,
      "eval_math_num_tokens": 43037844.0,
      "eval_math_runtime": 74.6984,
      "eval_math_samples_per_second": 6.694,
      "eval_math_steps_per_second": 1.673,
      "step": 1300
    },
    {
      "epoch": 2.08,
      "eval_physics_entropy": 1.0787066140174866,
      "eval_physics_loss": 1.181474208831787,
      "eval_physics_mean_token_accuracy": 0.7564280052185058,
      "eval_physics_num_tokens": 43037844.0,
      "eval_physics_runtime": 86.4265,
      "eval_physics_samples_per_second": 5.785,
      "eval_physics_steps_per_second": 1.446,
      "step": 1300
    },
    {
      "entropy": 1.0772028099745512,
      "epoch": 2.096,
      "grad_norm": 21.625,
      "learning_rate": 1.2525925925925928e-05,
      "loss": 18.7627,
      "mean_token_accuracy": 0.7594522505998611,
      "num_tokens": 43368675.0,
      "step": 1310
    },
    {
      "entropy": 1.0694199606776238,
      "epoch": 2.112,
      "grad_norm": 20.125,
      "learning_rate": 1.2451851851851853e-05,
      "loss": 18.5009,
      "mean_token_accuracy": 0.7623933359980584,
      "num_tokens": 43699544.0,
      "step": 1320
    },
    {
      "entropy": 1.0775703921914102,
      "epoch": 2.128,
      "grad_norm": 20.375,
      "learning_rate": 1.237777777777778e-05,
      "loss": 18.7445,
      "mean_token_accuracy": 0.7581640217453242,
      "num_tokens": 44012632.0,
      "step": 1330
    },
    {
      "entropy": 1.0589148577302694,
      "epoch": 2.144,
      "grad_norm": 18.375,
      "learning_rate": 1.2303703703703704e-05,
      "loss": 18.3475,
      "mean_token_accuracy": 0.7610369019210339,
      "num_tokens": 44349732.0,
      "step": 1340
    },
    {
      "entropy": 1.106671568751335,
      "epoch": 2.16,
      "grad_norm": 21.25,
      "learning_rate": 1.222962962962963e-05,
      "loss": 19.2264,
      "mean_token_accuracy": 0.7524396948516369,
      "num_tokens": 44680410.0,
      "step": 1350
    },
    {
      "entropy": 1.0581391721963882,
      "epoch": 2.176,
      "grad_norm": 18.5,
      "learning_rate": 1.2155555555555555e-05,
      "loss": 18.277,
      "mean_token_accuracy": 0.7640965595841408,
      "num_tokens": 45017922.0,
      "step": 1360
    },
    {
      "entropy": 1.0631261829286813,
      "epoch": 2.192,
      "grad_norm": 20.75,
      "learning_rate": 1.2081481481481484e-05,
      "loss": 18.3296,
      "mean_token_accuracy": 0.7635702408850193,
      "num_tokens": 45348179.0,
      "step": 1370
    },
    {
      "entropy": 1.048299814015627,
      "epoch": 2.208,
      "grad_norm": 22.0,
      "learning_rate": 1.2007407407407408e-05,
      "loss": 18.2386,
      "mean_token_accuracy": 0.7643856957554818,
      "num_tokens": 45678397.0,
      "step": 1380
    },
    {
      "entropy": 1.0520006015896797,
      "epoch": 2.224,
      "grad_norm": 19.375,
      "learning_rate": 1.1933333333333335e-05,
      "loss": 18.2818,
      "mean_token_accuracy": 0.7641717221587896,
      "num_tokens": 46010052.0,
      "step": 1390
    },
    {
      "entropy": 1.0484760511666535,
      "epoch": 2.24,
      "grad_norm": 18.25,
      "learning_rate": 1.185925925925926e-05,
      "loss": 18.184,
      "mean_token_accuracy": 0.7641188941895962,
      "num_tokens": 46347116.0,
      "step": 1400
    },
    {
      "epoch": 2.24,
      "eval_biology_entropy": 3.069058748245239,
      "eval_biology_loss": 3.6843225955963135,
      "eval_biology_mean_token_accuracy": 0.4650956363677978,
      "eval_biology_num_tokens": 46347116.0,
      "eval_biology_runtime": 57.3787,
      "eval_biology_samples_per_second": 8.714,
      "eval_biology_steps_per_second": 2.179,
      "step": 1400
    },
    {
      "epoch": 2.24,
      "eval_chemistry_entropy": 1.5667169632911682,
      "eval_chemistry_loss": 1.755946397781372,
      "eval_chemistry_mean_token_accuracy": 0.6802433683872223,
      "eval_chemistry_num_tokens": 46347116.0,
      "eval_chemistry_runtime": 72.6351,
      "eval_chemistry_samples_per_second": 6.884,
      "eval_chemistry_steps_per_second": 1.721,
      "step": 1400
    },
    {
      "epoch": 2.24,
      "eval_math_entropy": 1.105882652759552,
      "eval_math_loss": 1.4684650897979736,
      "eval_math_mean_token_accuracy": 0.720993902683258,
      "eval_math_num_tokens": 46347116.0,
      "eval_math_runtime": 75.1932,
      "eval_math_samples_per_second": 6.65,
      "eval_math_steps_per_second": 1.662,
      "step": 1400
    },
    {
      "epoch": 2.24,
      "eval_physics_entropy": 1.0758399648666381,
      "eval_physics_loss": 1.1777132749557495,
      "eval_physics_mean_token_accuracy": 0.7566362113952637,
      "eval_physics_num_tokens": 46347116.0,
      "eval_physics_runtime": 85.4605,
      "eval_physics_samples_per_second": 5.851,
      "eval_physics_steps_per_second": 1.463,
      "step": 1400
    },
    {
      "entropy": 1.1059522338211536,
      "epoch": 2.2560000000000002,
      "grad_norm": 21.375,
      "learning_rate": 1.1785185185185186e-05,
      "loss": 19.1498,
      "mean_token_accuracy": 0.7556869950145483,
      "num_tokens": 46673074.0,
      "step": 1410
    },
    {
      "entropy": 1.042980781197548,
      "epoch": 2.2720000000000002,
      "grad_norm": 20.0,
      "learning_rate": 1.1711111111111113e-05,
      "loss": 18.0571,
      "mean_token_accuracy": 0.7675418592989445,
      "num_tokens": 47009386.0,
      "step": 1420
    },
    {
      "entropy": 1.0891771662980319,
      "epoch": 2.288,
      "grad_norm": 19.375,
      "learning_rate": 1.1637037037037037e-05,
      "loss": 18.8901,
      "mean_token_accuracy": 0.7568176455795765,
      "num_tokens": 47344231.0,
      "step": 1430
    },
    {
      "entropy": 1.0886630911380053,
      "epoch": 2.304,
      "grad_norm": 22.875,
      "learning_rate": 1.1562962962962964e-05,
      "loss": 18.7894,
      "mean_token_accuracy": 0.7597312886267901,
      "num_tokens": 47658392.0,
      "step": 1440
    },
    {
      "entropy": 1.0535034578293563,
      "epoch": 2.32,
      "grad_norm": 20.125,
      "learning_rate": 1.1488888888888889e-05,
      "loss": 18.3235,
      "mean_token_accuracy": 0.7618763439357281,
      "num_tokens": 47995015.0,
      "step": 1450
    },
    {
      "entropy": 1.034178711846471,
      "epoch": 2.336,
      "grad_norm": 19.0,
      "learning_rate": 1.1414814814814817e-05,
      "loss": 17.9277,
      "mean_token_accuracy": 0.7673741526901722,
      "num_tokens": 48336708.0,
      "step": 1460
    },
    {
      "entropy": 1.0358063239604234,
      "epoch": 2.352,
      "grad_norm": 18.75,
      "learning_rate": 1.1340740740740742e-05,
      "loss": 17.9535,
      "mean_token_accuracy": 0.7666664138436318,
      "num_tokens": 48665311.0,
      "step": 1470
    },
    {
      "entropy": 1.062301841750741,
      "epoch": 2.368,
      "grad_norm": 19.0,
      "learning_rate": 1.1266666666666668e-05,
      "loss": 18.4889,
      "mean_token_accuracy": 0.76162573620677,
      "num_tokens": 49005334.0,
      "step": 1480
    },
    {
      "entropy": 1.100458450242877,
      "epoch": 2.384,
      "grad_norm": 22.625,
      "learning_rate": 1.1192592592592593e-05,
      "loss": 18.7753,
      "mean_token_accuracy": 0.7570323292165995,
      "num_tokens": 49331936.0,
      "step": 1490
    },
    {
      "entropy": 1.0668535258620977,
      "epoch": 2.4,
      "grad_norm": 19.75,
      "learning_rate": 1.111851851851852e-05,
      "loss": 18.6209,
      "mean_token_accuracy": 0.7592197827994823,
      "num_tokens": 49671048.0,
      "step": 1500
    },
    {
      "epoch": 2.4,
      "eval_biology_entropy": 3.032345582008362,
      "eval_biology_loss": 3.680708885192871,
      "eval_biology_mean_token_accuracy": 0.4642630183696747,
      "eval_biology_num_tokens": 49671048.0,
      "eval_biology_runtime": 58.1209,
      "eval_biology_samples_per_second": 8.603,
      "eval_biology_steps_per_second": 2.151,
      "step": 1500
    },
    {
      "epoch": 2.4,
      "eval_chemistry_entropy": 1.5479856276512145,
      "eval_chemistry_loss": 1.7515379190444946,
      "eval_chemistry_mean_token_accuracy": 0.6798697869777679,
      "eval_chemistry_num_tokens": 49671048.0,
      "eval_chemistry_runtime": 71.7682,
      "eval_chemistry_samples_per_second": 6.967,
      "eval_chemistry_steps_per_second": 1.742,
      "step": 1500
    },
    {
      "epoch": 2.4,
      "eval_math_entropy": 1.100146502494812,
      "eval_math_loss": 1.467644453048706,
      "eval_math_mean_token_accuracy": 0.7213256497383118,
      "eval_math_num_tokens": 49671048.0,
      "eval_math_runtime": 74.9283,
      "eval_math_samples_per_second": 6.673,
      "eval_math_steps_per_second": 1.668,
      "step": 1500
    },
    {
      "epoch": 2.4,
      "eval_physics_entropy": 1.0670854773521423,
      "eval_physics_loss": 1.1747907400131226,
      "eval_physics_mean_token_accuracy": 0.7574667348861694,
      "eval_physics_num_tokens": 49671048.0,
      "eval_physics_runtime": 85.2734,
      "eval_physics_samples_per_second": 5.863,
      "eval_physics_steps_per_second": 1.466,
      "step": 1500
    },
    {
      "entropy": 1.05430528819561,
      "epoch": 2.416,
      "grad_norm": 19.0,
      "learning_rate": 1.1044444444444444e-05,
      "loss": 18.1552,
      "mean_token_accuracy": 0.7650518540292979,
      "num_tokens": 50018182.0,
      "step": 1510
    },
    {
      "entropy": 1.0577623914927243,
      "epoch": 2.432,
      "grad_norm": 21.5,
      "learning_rate": 1.0970370370370371e-05,
      "loss": 18.3544,
      "mean_token_accuracy": 0.7647016257047653,
      "num_tokens": 50341187.0,
      "step": 1520
    },
    {
      "entropy": 1.0907797906547785,
      "epoch": 2.448,
      "grad_norm": 17.75,
      "learning_rate": 1.0896296296296298e-05,
      "loss": 18.9549,
      "mean_token_accuracy": 0.7548936065286398,
      "num_tokens": 50667019.0,
      "step": 1530
    },
    {
      "entropy": 1.0868097092956304,
      "epoch": 2.464,
      "grad_norm": 20.875,
      "learning_rate": 1.0822222222222222e-05,
      "loss": 18.7252,
      "mean_token_accuracy": 0.7591790940612555,
      "num_tokens": 50991113.0,
      "step": 1540
    },
    {
      "entropy": 1.0638412170112133,
      "epoch": 2.48,
      "grad_norm": 20.625,
      "learning_rate": 1.074814814814815e-05,
      "loss": 18.5812,
      "mean_token_accuracy": 0.7602099806070328,
      "num_tokens": 51319624.0,
      "step": 1550
    },
    {
      "entropy": 1.0854915376752614,
      "epoch": 2.496,
      "grad_norm": 19.75,
      "learning_rate": 1.0674074074074074e-05,
      "loss": 18.6781,
      "mean_token_accuracy": 0.7581930819898843,
      "num_tokens": 51654633.0,
      "step": 1560
    },
    {
      "entropy": 1.059612356312573,
      "epoch": 2.512,
      "grad_norm": 21.25,
      "learning_rate": 1.0600000000000002e-05,
      "loss": 18.4191,
      "mean_token_accuracy": 0.7628906007856131,
      "num_tokens": 51980051.0,
      "step": 1570
    },
    {
      "entropy": 1.0589963756501675,
      "epoch": 2.528,
      "grad_norm": 19.375,
      "learning_rate": 1.0525925925925927e-05,
      "loss": 18.318,
      "mean_token_accuracy": 0.7630537264049053,
      "num_tokens": 52316109.0,
      "step": 1580
    },
    {
      "entropy": 1.0746777534484864,
      "epoch": 2.544,
      "grad_norm": 23.375,
      "learning_rate": 1.0451851851851853e-05,
      "loss": 18.7474,
      "mean_token_accuracy": 0.7583229299634695,
      "num_tokens": 52634528.0,
      "step": 1590
    },
    {
      "entropy": 1.058685540780425,
      "epoch": 2.56,
      "grad_norm": 20.875,
      "learning_rate": 1.0377777777777778e-05,
      "loss": 18.3369,
      "mean_token_accuracy": 0.7622545655816794,
      "num_tokens": 52964035.0,
      "step": 1600
    },
    {
      "epoch": 2.56,
      "eval_biology_entropy": 3.0418846673965456,
      "eval_biology_loss": 3.6649463176727295,
      "eval_biology_mean_token_accuracy": 0.46480821561813357,
      "eval_biology_num_tokens": 52964035.0,
      "eval_biology_runtime": 58.2135,
      "eval_biology_samples_per_second": 8.589,
      "eval_biology_steps_per_second": 2.147,
      "step": 1600
    },
    {
      "epoch": 2.56,
      "eval_chemistry_entropy": 1.556763877391815,
      "eval_chemistry_loss": 1.7491346597671509,
      "eval_chemistry_mean_token_accuracy": 0.6798153126239777,
      "eval_chemistry_num_tokens": 52964035.0,
      "eval_chemistry_runtime": 72.5086,
      "eval_chemistry_samples_per_second": 6.896,
      "eval_chemistry_steps_per_second": 1.724,
      "step": 1600
    },
    {
      "epoch": 2.56,
      "eval_math_entropy": 1.104836820602417,
      "eval_math_loss": 1.4643486738204956,
      "eval_math_mean_token_accuracy": 0.721966543674469,
      "eval_math_num_tokens": 52964035.0,
      "eval_math_runtime": 74.6076,
      "eval_math_samples_per_second": 6.702,
      "eval_math_steps_per_second": 1.675,
      "step": 1600
    },
    {
      "epoch": 2.56,
      "eval_physics_entropy": 1.073439978122711,
      "eval_physics_loss": 1.171987533569336,
      "eval_physics_mean_token_accuracy": 0.75812575674057,
      "eval_physics_num_tokens": 52964035.0,
      "eval_physics_runtime": 85.0876,
      "eval_physics_samples_per_second": 5.876,
      "eval_physics_steps_per_second": 1.469,
      "step": 1600
    },
    {
      "entropy": 1.0514010008424521,
      "epoch": 2.576,
      "grad_norm": 20.875,
      "learning_rate": 1.0303703703703705e-05,
      "loss": 18.144,
      "mean_token_accuracy": 0.7653789080679416,
      "num_tokens": 53293345.0,
      "step": 1610
    },
    {
      "entropy": 1.0642182644456626,
      "epoch": 2.592,
      "grad_norm": 18.0,
      "learning_rate": 1.0229629629629631e-05,
      "loss": 18.5106,
      "mean_token_accuracy": 0.7617163117974997,
      "num_tokens": 53620896.0,
      "step": 1620
    },
    {
      "entropy": 1.030896332114935,
      "epoch": 2.608,
      "grad_norm": 18.375,
      "learning_rate": 1.0155555555555556e-05,
      "loss": 17.9262,
      "mean_token_accuracy": 0.7679114989936352,
      "num_tokens": 53950989.0,
      "step": 1630
    },
    {
      "entropy": 1.052972572296858,
      "epoch": 2.624,
      "grad_norm": 18.375,
      "learning_rate": 1.0081481481481484e-05,
      "loss": 18.2908,
      "mean_token_accuracy": 0.7629249345511198,
      "num_tokens": 54282732.0,
      "step": 1640
    },
    {
      "entropy": 1.0540472943335772,
      "epoch": 2.64,
      "grad_norm": 20.75,
      "learning_rate": 1.0007407407407407e-05,
      "loss": 18.288,
      "mean_token_accuracy": 0.764120651781559,
      "num_tokens": 54609748.0,
      "step": 1650
    },
    {
      "entropy": 1.093755446560681,
      "epoch": 2.656,
      "grad_norm": 21.75,
      "learning_rate": 9.933333333333334e-06,
      "loss": 18.9771,
      "mean_token_accuracy": 0.7559192903339863,
      "num_tokens": 54953583.0,
      "step": 1660
    },
    {
      "entropy": 1.068393959850073,
      "epoch": 2.672,
      "grad_norm": 18.875,
      "learning_rate": 9.85925925925926e-06,
      "loss": 18.4853,
      "mean_token_accuracy": 0.7603901419788599,
      "num_tokens": 55281945.0,
      "step": 1670
    },
    {
      "entropy": 1.0409850925207138,
      "epoch": 2.6879999999999997,
      "grad_norm": 23.625,
      "learning_rate": 9.785185185185187e-06,
      "loss": 17.9454,
      "mean_token_accuracy": 0.7679447285830975,
      "num_tokens": 55612391.0,
      "step": 1680
    },
    {
      "entropy": 1.033861081302166,
      "epoch": 2.7039999999999997,
      "grad_norm": 18.125,
      "learning_rate": 9.711111111111111e-06,
      "loss": 18.1138,
      "mean_token_accuracy": 0.7650211125612258,
      "num_tokens": 55949791.0,
      "step": 1690
    },
    {
      "entropy": 1.0701947890222072,
      "epoch": 2.7199999999999998,
      "grad_norm": 24.0,
      "learning_rate": 9.637037037037038e-06,
      "loss": 18.3555,
      "mean_token_accuracy": 0.7623727094382048,
      "num_tokens": 56282204.0,
      "step": 1700
    },
    {
      "epoch": 2.7199999999999998,
      "eval_biology_entropy": 3.0319280109405518,
      "eval_biology_loss": 3.670518159866333,
      "eval_biology_mean_token_accuracy": 0.46465519547462464,
      "eval_biology_num_tokens": 56282204.0,
      "eval_biology_runtime": 57.7598,
      "eval_biology_samples_per_second": 8.657,
      "eval_biology_steps_per_second": 2.164,
      "step": 1700
    },
    {
      "epoch": 2.7199999999999998,
      "eval_chemistry_entropy": 1.5443168606758118,
      "eval_chemistry_loss": 1.747648000717163,
      "eval_chemistry_mean_token_accuracy": 0.6801766774654389,
      "eval_chemistry_num_tokens": 56282204.0,
      "eval_chemistry_runtime": 72.2239,
      "eval_chemistry_samples_per_second": 6.923,
      "eval_chemistry_steps_per_second": 1.731,
      "step": 1700
    },
    {
      "epoch": 2.7199999999999998,
      "eval_math_entropy": 1.0994245495796204,
      "eval_math_loss": 1.464704155921936,
      "eval_math_mean_token_accuracy": 0.7219137578010559,
      "eval_math_num_tokens": 56282204.0,
      "eval_math_runtime": 75.0727,
      "eval_math_samples_per_second": 6.66,
      "eval_math_steps_per_second": 1.665,
      "step": 1700
    },
    {
      "epoch": 2.7199999999999998,
      "eval_physics_entropy": 1.06145493888855,
      "eval_physics_loss": 1.1699579954147339,
      "eval_physics_mean_token_accuracy": 0.7579556694030761,
      "eval_physics_num_tokens": 56282204.0,
      "eval_physics_runtime": 85.528,
      "eval_physics_samples_per_second": 5.846,
      "eval_physics_steps_per_second": 1.462,
      "step": 1700
    },
    {
      "entropy": 1.093729916960001,
      "epoch": 2.7359999999999998,
      "grad_norm": 20.875,
      "learning_rate": 9.562962962962965e-06,
      "loss": 19.1324,
      "mean_token_accuracy": 0.7557784903794527,
      "num_tokens": 56605478.0,
      "step": 1710
    },
    {
      "entropy": 1.0620717525482177,
      "epoch": 2.752,
      "grad_norm": 18.125,
      "learning_rate": 9.48888888888889e-06,
      "loss": 18.4694,
      "mean_token_accuracy": 0.7616514947265387,
      "num_tokens": 56935874.0,
      "step": 1720
    },
    {
      "entropy": 1.0455114886164665,
      "epoch": 2.768,
      "grad_norm": 19.75,
      "learning_rate": 9.414814814814816e-06,
      "loss": 18.0225,
      "mean_token_accuracy": 0.766771863028407,
      "num_tokens": 57270552.0,
      "step": 1730
    },
    {
      "entropy": 1.0662406332790852,
      "epoch": 2.784,
      "grad_norm": 21.25,
      "learning_rate": 9.34074074074074e-06,
      "loss": 18.4719,
      "mean_token_accuracy": 0.7604681506752968,
      "num_tokens": 57602968.0,
      "step": 1740
    },
    {
      "entropy": 1.1209681322798133,
      "epoch": 2.8,
      "grad_norm": 21.875,
      "learning_rate": 9.266666666666667e-06,
      "loss": 19.574,
      "mean_token_accuracy": 0.7499242343008519,
      "num_tokens": 57925886.0,
      "step": 1750
    },
    {
      "entropy": 1.045488651841879,
      "epoch": 2.816,
      "grad_norm": 19.5,
      "learning_rate": 9.192592592592594e-06,
      "loss": 17.9029,
      "mean_token_accuracy": 0.7672869808971882,
      "num_tokens": 58262960.0,
      "step": 1760
    },
    {
      "entropy": 1.0733379565179348,
      "epoch": 2.832,
      "grad_norm": 21.25,
      "learning_rate": 9.118518518518518e-06,
      "loss": 18.6542,
      "mean_token_accuracy": 0.75910082757473,
      "num_tokens": 58586324.0,
      "step": 1770
    },
    {
      "entropy": 1.048666688427329,
      "epoch": 2.848,
      "grad_norm": 20.5,
      "learning_rate": 9.044444444444445e-06,
      "loss": 18.272,
      "mean_token_accuracy": 0.7628897685557604,
      "num_tokens": 58912949.0,
      "step": 1780
    },
    {
      "entropy": 1.037505378201604,
      "epoch": 2.864,
      "grad_norm": 17.375,
      "learning_rate": 8.970370370370372e-06,
      "loss": 17.9683,
      "mean_token_accuracy": 0.7664824549108744,
      "num_tokens": 59252494.0,
      "step": 1790
    },
    {
      "entropy": 1.0653341230005027,
      "epoch": 2.88,
      "grad_norm": 23.625,
      "learning_rate": 8.896296296296298e-06,
      "loss": 18.4816,
      "mean_token_accuracy": 0.7624582652002573,
      "num_tokens": 59574951.0,
      "step": 1800
    },
    {
      "epoch": 2.88,
      "eval_biology_entropy": 3.0604784240722656,
      "eval_biology_loss": 3.653829336166382,
      "eval_biology_mean_token_accuracy": 0.4658286116123199,
      "eval_biology_num_tokens": 59574951.0,
      "eval_biology_runtime": 58.2002,
      "eval_biology_samples_per_second": 8.591,
      "eval_biology_steps_per_second": 2.148,
      "step": 1800
    },
    {
      "epoch": 2.88,
      "eval_chemistry_entropy": 1.5504239439964294,
      "eval_chemistry_loss": 1.7466660737991333,
      "eval_chemistry_mean_token_accuracy": 0.6807781052589417,
      "eval_chemistry_num_tokens": 59574951.0,
      "eval_chemistry_runtime": 71.811,
      "eval_chemistry_samples_per_second": 6.963,
      "eval_chemistry_steps_per_second": 1.741,
      "step": 1800
    },
    {
      "epoch": 2.88,
      "eval_math_entropy": 1.0987313556671143,
      "eval_math_loss": 1.463243842124939,
      "eval_math_mean_token_accuracy": 0.722051815032959,
      "eval_math_num_tokens": 59574951.0,
      "eval_math_runtime": 74.7495,
      "eval_math_samples_per_second": 6.689,
      "eval_math_steps_per_second": 1.672,
      "step": 1800
    },
    {
      "epoch": 2.88,
      "eval_physics_entropy": 1.0614235019683838,
      "eval_physics_loss": 1.1685162782669067,
      "eval_physics_mean_token_accuracy": 0.7582965307235717,
      "eval_physics_num_tokens": 59574951.0,
      "eval_physics_runtime": 85.1707,
      "eval_physics_samples_per_second": 5.871,
      "eval_physics_steps_per_second": 1.468,
      "step": 1800
    },
    {
      "entropy": 1.041395665332675,
      "epoch": 2.896,
      "grad_norm": 20.5,
      "learning_rate": 8.822222222222223e-06,
      "loss": 18.2273,
      "mean_token_accuracy": 0.7648600865155458,
      "num_tokens": 59903274.0,
      "step": 1810
    },
    {
      "entropy": 1.0448328781872989,
      "epoch": 2.912,
      "grad_norm": 19.875,
      "learning_rate": 8.74814814814815e-06,
      "loss": 18.1075,
      "mean_token_accuracy": 0.765788996964693,
      "num_tokens": 60234203.0,
      "step": 1820
    },
    {
      "entropy": 1.052156737446785,
      "epoch": 2.928,
      "grad_norm": 21.0,
      "learning_rate": 8.674074074074074e-06,
      "loss": 18.1653,
      "mean_token_accuracy": 0.7660945057868958,
      "num_tokens": 60559123.0,
      "step": 1830
    },
    {
      "entropy": 1.0404166903346777,
      "epoch": 2.944,
      "grad_norm": 22.75,
      "learning_rate": 8.6e-06,
      "loss": 18.0472,
      "mean_token_accuracy": 0.7664348762482405,
      "num_tokens": 60893423.0,
      "step": 1840
    },
    {
      "entropy": 1.0563691571354865,
      "epoch": 2.96,
      "grad_norm": 18.375,
      "learning_rate": 8.525925925925927e-06,
      "loss": 18.4101,
      "mean_token_accuracy": 0.7621017336845398,
      "num_tokens": 61228692.0,
      "step": 1850
    },
    {
      "entropy": 1.0723481249064206,
      "epoch": 2.976,
      "grad_norm": 19.5,
      "learning_rate": 8.451851851851852e-06,
      "loss": 18.5154,
      "mean_token_accuracy": 0.7603589288890362,
      "num_tokens": 61560852.0,
      "step": 1860
    },
    {
      "entropy": 1.0210791245102881,
      "epoch": 2.992,
      "grad_norm": 20.375,
      "learning_rate": 8.377777777777779e-06,
      "loss": 17.7535,
      "mean_token_accuracy": 0.769398371502757,
      "num_tokens": 61898951.0,
      "step": 1870
    },
    {
      "entropy": 1.0361403860151768,
      "epoch": 3.008,
      "grad_norm": 19.125,
      "learning_rate": 8.303703703703705e-06,
      "loss": 18.034,
      "mean_token_accuracy": 0.7676448699086904,
      "num_tokens": 62223875.0,
      "step": 1880
    },
    {
      "entropy": 1.054644862189889,
      "epoch": 3.024,
      "grad_norm": 20.5,
      "learning_rate": 8.229629629629632e-06,
      "loss": 18.2441,
      "mean_token_accuracy": 0.7643492009490729,
      "num_tokens": 62553166.0,
      "step": 1890
    },
    {
      "entropy": 1.0314039301127196,
      "epoch": 3.04,
      "grad_norm": 20.375,
      "learning_rate": 8.155555555555556e-06,
      "loss": 17.829,
      "mean_token_accuracy": 0.7673805825412273,
      "num_tokens": 62893113.0,
      "step": 1900
    },
    {
      "epoch": 3.04,
      "eval_biology_entropy": 3.0260161380767823,
      "eval_biology_loss": 3.6554806232452393,
      "eval_biology_mean_token_accuracy": 0.4659731407165527,
      "eval_biology_num_tokens": 62893113.0,
      "eval_biology_runtime": 58.103,
      "eval_biology_samples_per_second": 8.605,
      "eval_biology_steps_per_second": 2.151,
      "step": 1900
    },
    {
      "epoch": 3.04,
      "eval_chemistry_entropy": 1.5367405834197998,
      "eval_chemistry_loss": 1.7443947792053223,
      "eval_chemistry_mean_token_accuracy": 0.6808655200004577,
      "eval_chemistry_num_tokens": 62893113.0,
      "eval_chemistry_runtime": 72.9446,
      "eval_chemistry_samples_per_second": 6.855,
      "eval_chemistry_steps_per_second": 1.714,
      "step": 1900
    },
    {
      "epoch": 3.04,
      "eval_math_entropy": 1.0913958082199096,
      "eval_math_loss": 1.4638071060180664,
      "eval_math_mean_token_accuracy": 0.7221347332000733,
      "eval_math_num_tokens": 62893113.0,
      "eval_math_runtime": 74.5849,
      "eval_math_samples_per_second": 6.704,
      "eval_math_steps_per_second": 1.676,
      "step": 1900
    },
    {
      "epoch": 3.04,
      "eval_physics_entropy": 1.0539872069358827,
      "eval_physics_loss": 1.1672000885009766,
      "eval_physics_mean_token_accuracy": 0.7585978369712829,
      "eval_physics_num_tokens": 62893113.0,
      "eval_physics_runtime": 85.1536,
      "eval_physics_samples_per_second": 5.872,
      "eval_physics_steps_per_second": 1.468,
      "step": 1900
    },
    {
      "entropy": 1.0623472094535829,
      "epoch": 3.056,
      "grad_norm": 20.875,
      "learning_rate": 8.081481481481483e-06,
      "loss": 18.5719,
      "mean_token_accuracy": 0.7610293030738831,
      "num_tokens": 63218088.0,
      "step": 1910
    },
    {
      "entropy": 1.0648277301341296,
      "epoch": 3.072,
      "grad_norm": 21.25,
      "learning_rate": 8.007407407407408e-06,
      "loss": 18.3262,
      "mean_token_accuracy": 0.7616520181298256,
      "num_tokens": 63551201.0,
      "step": 1920
    },
    {
      "entropy": 1.0729082938283683,
      "epoch": 3.088,
      "grad_norm": 21.125,
      "learning_rate": 7.933333333333334e-06,
      "loss": 18.5813,
      "mean_token_accuracy": 0.7592061251401901,
      "num_tokens": 63876362.0,
      "step": 1930
    },
    {
      "entropy": 1.0287240067496897,
      "epoch": 3.104,
      "grad_norm": 22.5,
      "learning_rate": 7.859259259259259e-06,
      "loss": 17.8296,
      "mean_token_accuracy": 0.7683002337813377,
      "num_tokens": 64208171.0,
      "step": 1940
    },
    {
      "entropy": 1.0403265412896872,
      "epoch": 3.12,
      "grad_norm": 20.125,
      "learning_rate": 7.785185185185185e-06,
      "loss": 17.9084,
      "mean_token_accuracy": 0.7681071911007166,
      "num_tokens": 64544493.0,
      "step": 1950
    },
    {
      "entropy": 1.0773630373179912,
      "epoch": 3.136,
      "grad_norm": 21.375,
      "learning_rate": 7.711111111111112e-06,
      "loss": 18.7301,
      "mean_token_accuracy": 0.7578829418867826,
      "num_tokens": 64872488.0,
      "step": 1960
    },
    {
      "entropy": 1.0316934179514647,
      "epoch": 3.152,
      "grad_norm": 19.875,
      "learning_rate": 7.637037037037037e-06,
      "loss": 17.9357,
      "mean_token_accuracy": 0.767891078069806,
      "num_tokens": 65201124.0,
      "step": 1970
    },
    {
      "entropy": 1.0546127382665873,
      "epoch": 3.168,
      "grad_norm": 17.75,
      "learning_rate": 7.562962962962963e-06,
      "loss": 18.3228,
      "mean_token_accuracy": 0.7644848499447108,
      "num_tokens": 65525868.0,
      "step": 1980
    },
    {
      "entropy": 1.0422284375876187,
      "epoch": 3.184,
      "grad_norm": 20.875,
      "learning_rate": 7.48888888888889e-06,
      "loss": 18.0602,
      "mean_token_accuracy": 0.7683299861848354,
      "num_tokens": 65861603.0,
      "step": 1990
    },
    {
      "entropy": 1.0573079880326985,
      "epoch": 3.2,
      "grad_norm": 19.0,
      "learning_rate": 7.4148148148148155e-06,
      "loss": 18.277,
      "mean_token_accuracy": 0.7637478534132243,
      "num_tokens": 66193675.0,
      "step": 2000
    },
    {
      "epoch": 3.2,
      "eval_biology_entropy": 3.033109013557434,
      "eval_biology_loss": 3.662965774536133,
      "eval_biology_mean_token_accuracy": 0.46735861563682557,
      "eval_biology_num_tokens": 66193675.0,
      "eval_biology_runtime": 57.5764,
      "eval_biology_samples_per_second": 8.684,
      "eval_biology_steps_per_second": 2.171,
      "step": 2000
    },
    {
      "epoch": 3.2,
      "eval_chemistry_entropy": 1.5397213683128357,
      "eval_chemistry_loss": 1.7426356077194214,
      "eval_chemistry_mean_token_accuracy": 0.6820196812152862,
      "eval_chemistry_num_tokens": 66193675.0,
      "eval_chemistry_runtime": 72.7193,
      "eval_chemistry_samples_per_second": 6.876,
      "eval_chemistry_steps_per_second": 1.719,
      "step": 2000
    },
    {
      "epoch": 3.2,
      "eval_math_entropy": 1.0944099469184876,
      "eval_math_loss": 1.4623637199401855,
      "eval_math_mean_token_accuracy": 0.7220344128608703,
      "eval_math_num_tokens": 66193675.0,
      "eval_math_runtime": 75.2724,
      "eval_math_samples_per_second": 6.643,
      "eval_math_steps_per_second": 1.661,
      "step": 2000
    },
    {
      "epoch": 3.2,
      "eval_physics_entropy": 1.056751304626465,
      "eval_physics_loss": 1.1659578084945679,
      "eval_physics_mean_token_accuracy": 0.7591085810661315,
      "eval_physics_num_tokens": 66193675.0,
      "eval_physics_runtime": 86.2245,
      "eval_physics_samples_per_second": 5.799,
      "eval_physics_steps_per_second": 1.45,
      "step": 2000
    }
  ],
  "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": 6.614072648666411e+18,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}