{ "best_global_step": null, "best_metric": null, "best_model_checkpoint": null, "epoch": 0.11160714285714286, "eval_steps": 500, "global_step": 50, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 697.140625, "completions/mean_terminated_length": 653.5645141601562, "completions/min_length": 15.0, "completions/min_terminated_length": 15.0, "epoch": 0.002232142857142857, "frac_reward_zero_std": 0.0, "grad_norm": 1.256870985031128, "learning_rate": 0.0, "loss": -0.0, "num_tokens": 173481.0, "reward": 0.46328121423721313, "reward_std": 0.4737887382507324, "rewards/format_reward/mean": 0.4140625, "rewards/format_reward/std": 0.19012710452079773, "rewards/mcq_exact_match_reward/mean": 0.421875, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "step": 1 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1016.0, "completions/max_terminated_length": 1016.0, "completions/mean_length": 473.921875, "completions/mean_terminated_length": 473.921875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.004464285714285714, "frac_reward_zero_std": 0.0, "grad_norm": 5.316381931304932, "learning_rate": 2e-07, "loss": -0.0, "num_tokens": 281476.0, "reward": 0.1875, "reward_std": 0.28762155771255493, "rewards/format_reward/mean": 0.3125, "rewards/format_reward/std": 0.28867512941360474, "rewards/mcq_exact_match_reward/mean": 0.15625, "rewards/mcq_exact_match_reward/std": 0.36596253514289856, "step": 2 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1823.0, "completions/mean_length": 550.640625, "completions/mean_terminated_length": 502.33868408203125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.006696428571428571, "frac_reward_zero_std": 0.0, "grad_norm": 3.0207016468048096, "learning_rate": 4e-07, "loss": -0.0, "num_tokens": 434925.0, "reward": 0.20781248807907104, "reward_std": 0.31630992889404297, "rewards/format_reward/mean": 0.359375, "rewards/format_reward/std": 0.24346621334552765, "rewards/mcq_exact_match_reward/mean": 0.171875, "rewards/mcq_exact_match_reward/std": 0.38025420904159546, "step": 3 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1503.0, "completions/mean_length": 542.578125, "completions/mean_terminated_length": 518.6825561523438, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.008928571428571428, "frac_reward_zero_std": 0.125, "grad_norm": 7.995953559875488, "learning_rate": 6e-07, "loss": -0.0, "num_tokens": 578562.0, "reward": 0.2749999761581421, "reward_std": 0.27574247121810913, "rewards/format_reward/mean": 0.40625, "rewards/format_reward/std": 0.19669894874095917, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "step": 4 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1591.0, "completions/max_terminated_length": 1591.0, "completions/mean_length": 461.65625, "completions/mean_terminated_length": 461.65625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.011160714285714286, "frac_reward_zero_std": 0.0, "grad_norm": 9.621064186096191, "learning_rate": 8e-07, "loss": -0.0, "num_tokens": 689084.0, "reward": 0.3179687261581421, "reward_std": 0.3347626328468323, "rewards/format_reward/mean": 0.3671875, "rewards/format_reward/std": 0.2708333432674408, "rewards/mcq_exact_match_reward/mean": 0.28125, "rewards/mcq_exact_match_reward/std": 0.4531635046005249, "step": 5 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1962.0, "completions/mean_length": 673.515625, "completions/mean_terminated_length": 651.6984252929688, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.013392857142857142, "frac_reward_zero_std": 0.125, "grad_norm": 7.209977149963379, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 826877.0, "reward": 0.32265621423721313, "reward_std": 0.2884722650051117, "rewards/format_reward/mean": 0.4140625, "rewards/format_reward/std": 0.209963858127594, "rewards/mcq_exact_match_reward/mean": 0.28125, "rewards/mcq_exact_match_reward/std": 0.4531635046005249, "step": 6 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1664.0, "completions/mean_length": 579.359375, "completions/mean_terminated_length": 531.9838256835938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.015625, "frac_reward_zero_std": 0.125, "grad_norm": 13.31259536743164, "learning_rate": 9.99726628670463e-07, "loss": 0.0, "num_tokens": 981732.0, "reward": 0.18828123807907104, "reward_std": 0.3201846480369568, "rewards/format_reward/mean": 0.3203125, "rewards/format_reward/std": 0.24180518090724945, "rewards/mcq_exact_match_reward/mean": 0.15625, "rewards/mcq_exact_match_reward/std": 0.36596253514289856, "step": 7 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1863.0, "completions/mean_length": 628.1875, "completions/mean_terminated_length": 558.360595703125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.017857142857142856, "frac_reward_zero_std": 0.0, "grad_norm": 5.400537967681885, "learning_rate": 9.989068136093872e-07, "loss": 0.0, "num_tokens": 1135112.0, "reward": 0.26406246423721313, "reward_std": 0.3958834111690521, "rewards/format_reward/mean": 0.296875, "rewards/format_reward/std": 0.24750742316246033, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "step": 8 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1559.0, "completions/mean_length": 558.4375, "completions/mean_terminated_length": 534.793701171875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.020089285714285716, "frac_reward_zero_std": 0.125, "grad_norm": 10.620858192443848, "learning_rate": 9.975414512725056e-07, "loss": 0.0, "num_tokens": 1268916.0, "reward": 0.3070312440395355, "reward_std": 0.3649529218673706, "rewards/format_reward/mean": 0.4140625, "rewards/format_reward/std": 0.2605654299259186, "rewards/mcq_exact_match_reward/mean": 0.265625, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "step": 9 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1356.0, "completions/mean_length": 521.921875, "completions/mean_terminated_length": 497.6984558105469, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.022321428571428572, "frac_reward_zero_std": 0.0, "grad_norm": 2.200150728225708, "learning_rate": 9.956320346634875e-07, "loss": -0.0, "num_tokens": 1407095.0, "reward": 0.2749999761581421, "reward_std": 0.38830214738845825, "rewards/format_reward/mean": 0.40625, "rewards/format_reward/std": 0.25, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "step": 10 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1313.0, "completions/mean_length": 555.203125, "completions/mean_terminated_length": 481.786865234375, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "epoch": 0.024553571428571428, "frac_reward_zero_std": 0.125, "grad_norm": 2.5358588695526123, "learning_rate": 9.931806517013612e-07, "loss": 0.0, "num_tokens": 1532588.0, "reward": 0.33281248807907104, "reward_std": 0.33862143754959106, "rewards/format_reward/mean": 0.359375, "rewards/format_reward/std": 0.3145764470100403, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "step": 11 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 2008.0, "completions/max_terminated_length": 2008.0, "completions/mean_length": 579.921875, "completions/mean_terminated_length": 579.921875, "completions/min_length": 10.0, "completions/min_terminated_length": 10.0, "epoch": 0.026785714285714284, "frac_reward_zero_std": 0.125, "grad_norm": 3.9451773166656494, "learning_rate": 9.901899829374047e-07, "loss": 0.0, "num_tokens": 1688031.0, "reward": 0.2593749761581421, "reward_std": 0.29294344782829285, "rewards/format_reward/mean": 0.40625, "rewards/format_reward/std": 0.21593283116817474, "rewards/mcq_exact_match_reward/mean": 0.21875, "rewards/mcq_exact_match_reward/std": 0.4166666865348816, "step": 12 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1496.0, "completions/mean_length": 490.40625, "completions/mean_terminated_length": 413.8032531738281, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.029017857142857144, "frac_reward_zero_std": 0.125, "grad_norm": 4.638545036315918, "learning_rate": 9.866632986240029e-07, "loss": -0.0, "num_tokens": 1805081.0, "reward": 0.36953121423721313, "reward_std": 0.2620023190975189, "rewards/format_reward/mean": 0.4140625, "rewards/format_reward/std": 0.244862899184227, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "step": 13 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1262.0, "completions/mean_length": 498.953125, "completions/mean_terminated_length": 448.9838562011719, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.03125, "frac_reward_zero_std": 0.0, "grad_norm": 7.494521141052246, "learning_rate": 9.826044551386742e-07, "loss": 0.0, "num_tokens": 1941862.0, "reward": 0.24140626192092896, "reward_std": 0.38149911165237427, "rewards/format_reward/mean": 0.3828125, "rewards/format_reward/std": 0.2780628502368927, "rewards/mcq_exact_match_reward/mean": 0.203125, "rewards/mcq_exact_match_reward/std": 0.40550529956817627, "step": 14 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1587.0, "completions/max_terminated_length": 1587.0, "completions/mean_length": 470.21875, "completions/mean_terminated_length": 470.21875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.033482142857142856, "frac_reward_zero_std": 0.125, "grad_norm": 3.4959218502044678, "learning_rate": 9.780178907671788e-07, "loss": -0.0, "num_tokens": 2079620.0, "reward": 0.41093745827674866, "reward_std": 0.3220180869102478, "rewards/format_reward/mean": 0.359375, "rewards/format_reward/std": 0.2592533528804779, "rewards/mcq_exact_match_reward/mean": 0.375, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "step": 15 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1521.0, "completions/mean_length": 573.390625, "completions/mean_terminated_length": 525.8225708007812, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.03571428571428571, "frac_reward_zero_std": 0.0, "grad_norm": 8.447587013244629, "learning_rate": 9.729086208503173e-07, "loss": -0.0, "num_tokens": 2191725.0, "reward": 0.3687499761581421, "reward_std": 0.4066651463508606, "rewards/format_reward/mean": 0.40625, "rewards/format_reward/std": 0.279951810836792, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "step": 16 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1512.0, "completions/mean_length": 557.953125, "completions/mean_terminated_length": 509.8870849609375, "completions/min_length": 10.0, "completions/min_terminated_length": 10.0, "epoch": 0.03794642857142857, "frac_reward_zero_std": 0.125, "grad_norm": 11.934052467346191, "learning_rate": 9.672822322997304e-07, "loss": -0.0, "num_tokens": 2320010.0, "reward": 0.2203124761581421, "reward_std": 0.30440255999565125, "rewards/format_reward/mean": 0.484375, "rewards/format_reward/std": 0.29504841566085815, "rewards/mcq_exact_match_reward/mean": 0.171875, "rewards/mcq_exact_match_reward/std": 0.38025420904159546, "step": 17 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1765.0, "completions/mean_length": 422.203125, "completions/mean_terminated_length": 396.3968505859375, "completions/min_length": 10.0, "completions/min_terminated_length": 10.0, "epoch": 0.04017857142857143, "frac_reward_zero_std": 0.0, "grad_norm": 5.943488597869873, "learning_rate": 9.611448774886923e-07, "loss": -0.0, "num_tokens": 2434295.0, "reward": 0.4687499701976776, "reward_std": 0.46519073843955994, "rewards/format_reward/mean": 0.46875, "rewards/format_reward/std": 0.279951810836792, "rewards/mcq_exact_match_reward/mean": 0.421875, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "step": 18 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1710.0, "completions/mean_length": 518.890625, "completions/mean_terminated_length": 494.61907958984375, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.04241071428571429, "frac_reward_zero_std": 0.0, "grad_norm": 9.51138973236084, "learning_rate": 9.545032675245813e-07, "loss": 0.0, "num_tokens": 2567016.0, "reward": 0.3218750059604645, "reward_std": 0.3252774775028229, "rewards/format_reward/mean": 0.5625, "rewards/format_reward/std": 0.3149704039096832, "rewards/mcq_exact_match_reward/mean": 0.265625, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "step": 19 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1747.0, "completions/mean_length": 612.84375, "completions/mean_terminated_length": 566.54833984375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.044642857142857144, "frac_reward_zero_std": 0.0, "grad_norm": 2.909266471862793, "learning_rate": 9.473646649103817e-07, "loss": -0.0, "num_tokens": 2696350.0, "reward": 0.22187498211860657, "reward_std": 0.34274256229400635, "rewards/format_reward/mean": 0.5, "rewards/format_reward/std": 0.29546841979026794, "rewards/mcq_exact_match_reward/mean": 0.171875, "rewards/mcq_exact_match_reward/std": 0.38025420904159546, "step": 20 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1607.0, "completions/mean_length": 535.65625, "completions/mean_terminated_length": 511.65081787109375, "completions/min_length": 10.0, "completions/min_terminated_length": 10.0, "epoch": 0.046875, "frac_reward_zero_std": 0.0, "grad_norm": 7.937094688415527, "learning_rate": 9.397368756032444e-07, "loss": 0.0, "num_tokens": 2838336.0, "reward": 0.44062501192092896, "reward_std": 0.3860820233821869, "rewards/format_reward/mean": 0.5, "rewards/format_reward/std": 0.3673154413700104, "rewards/mcq_exact_match_reward/mean": 0.390625, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "step": 21 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1839.0, "completions/max_terminated_length": 1839.0, "completions/mean_length": 493.40625, "completions/mean_terminated_length": 493.40625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.049107142857142856, "frac_reward_zero_std": 0.0, "grad_norm": 10.74514389038086, "learning_rate": 9.316282404787869e-07, "loss": -0.0, "num_tokens": 2973234.0, "reward": 0.19843748211860657, "reward_std": 0.30221039056777954, "rewards/format_reward/mean": 0.421875, "rewards/format_reward/std": 0.3238992393016815, "rewards/mcq_exact_match_reward/mean": 0.15625, "rewards/mcq_exact_match_reward/std": 0.36596253514289856, "step": 22 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1701.0, "completions/mean_length": 427.40625, "completions/mean_terminated_length": 401.68255615234375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.05133928571428571, "frac_reward_zero_std": 0.125, "grad_norm": 10.760703086853027, "learning_rate": 9.230476262104676e-07, "loss": -0.0, "num_tokens": 3102308.0, "reward": 0.21406251192092896, "reward_std": 0.2570106089115143, "rewards/format_reward/mean": 0.578125, "rewards/format_reward/std": 0.3475395441055298, "rewards/mcq_exact_match_reward/mean": 0.15625, "rewards/mcq_exact_match_reward/std": 0.36596253514289856, "step": 23 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1828.0, "completions/mean_length": 384.625, "completions/mean_terminated_length": 358.22222900390625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.05357142857142857, "frac_reward_zero_std": 0.0, "grad_norm": 12.165411949157715, "learning_rate": 9.1400441557401e-07, "loss": 0.0, "num_tokens": 3219100.0, "reward": 0.28437501192092896, "reward_std": 0.35062626004219055, "rewards/format_reward/mean": 0.65625, "rewards/format_reward/std": 0.3870421051979065, "rewards/mcq_exact_match_reward/mean": 0.21875, "rewards/mcq_exact_match_reward/std": 0.4166666865348816, "step": 24 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1937.0, "completions/mean_length": 338.5625, "completions/mean_terminated_length": 311.4285888671875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.05580357142857143, "frac_reward_zero_std": 0.0, "grad_norm": 17.14126205444336, "learning_rate": 9.045084971874737e-07, "loss": 0.0, "num_tokens": 3342312.0, "reward": 0.42265623807907104, "reward_std": 0.47679829597473145, "rewards/format_reward/mean": 0.6328125, "rewards/format_reward/std": 0.3699226677417755, "rewards/mcq_exact_match_reward/mean": 0.359375, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "step": 25 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1201.0, "completions/mean_length": 337.171875, "completions/mean_terminated_length": 310.0158996582031, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.05803571428571429, "frac_reward_zero_std": 0.0, "grad_norm": 19.676746368408203, "learning_rate": 8.945702546981968e-07, "loss": -0.0, "num_tokens": 3476507.0, "reward": 0.24843749403953552, "reward_std": 0.30121952295303345, "rewards/format_reward/mean": 0.609375, "rewards/format_reward/std": 0.40305137634277344, "rewards/mcq_exact_match_reward/mean": 0.1875, "rewards/mcq_exact_match_reward/std": 0.39339789748191833, "step": 26 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1115.0, "completions/mean_length": 344.84375, "completions/mean_terminated_length": 317.8095397949219, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.060267857142857144, "frac_reward_zero_std": 0.0, "grad_norm": 10.468759536743164, "learning_rate": 8.842005554284295e-07, "loss": 0.0, "num_tokens": 3603041.0, "reward": 0.3265625238418579, "reward_std": 0.4080768823623657, "rewards/format_reward/mean": 0.765625, "rewards/format_reward/std": 0.3329611122608185, "rewards/mcq_exact_match_reward/mean": 0.25, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "step": 27 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1426.0, "completions/mean_length": 212.5625, "completions/mean_terminated_length": 183.4285888671875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0625, "frac_reward_zero_std": 0.25, "grad_norm": 10.699164390563965, "learning_rate": 8.734107384920769e-07, "loss": 0.0, "num_tokens": 3704645.0, "reward": 0.28984373807907104, "reward_std": 0.2218824028968811, "rewards/format_reward/mean": 0.7109375, "rewards/format_reward/std": 0.3863205313682556, "rewards/mcq_exact_match_reward/mean": 0.21875, "rewards/mcq_exact_match_reward/std": 0.4166666865348816, "step": 28 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1619.0, "completions/max_terminated_length": 1619.0, "completions/mean_length": 228.25, "completions/mean_terminated_length": 228.25, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.06473214285714286, "frac_reward_zero_std": 0.125, "grad_norm": 11.572940826416016, "learning_rate": 8.622126023955445e-07, "loss": -0.0, "num_tokens": 3817429.0, "reward": 0.26484376192092896, "reward_std": 0.3309459090232849, "rewards/format_reward/mean": 0.7734375, "rewards/format_reward/std": 0.3077012598514557, "rewards/mcq_exact_match_reward/mean": 0.1875, "rewards/mcq_exact_match_reward/std": 0.39339789748191833, "step": 29 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 902.0, "completions/max_terminated_length": 902.0, "completions/mean_length": 73.796875, "completions/mean_terminated_length": 73.796875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.06696428571428571, "frac_reward_zero_std": 0.125, "grad_norm": 19.748498916625977, "learning_rate": 8.506183921362442e-07, "loss": 0.0, "num_tokens": 3925720.0, "reward": 0.4007812440395355, "reward_std": 0.2942570447921753, "rewards/format_reward/mean": 0.8828125, "rewards/format_reward/std": 0.2478829026222229, "rewards/mcq_exact_match_reward/mean": 0.3125, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "step": 30 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1586.0, "completions/max_terminated_length": 1586.0, "completions/mean_length": 194.546875, "completions/mean_terminated_length": 194.546875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.06919642857142858, "frac_reward_zero_std": 0.25, "grad_norm": 17.607955932617188, "learning_rate": 8.386407858128706e-07, "loss": -0.0, "num_tokens": 4068131.0, "reward": 0.47734373807907104, "reward_std": 0.36437416076660156, "rewards/format_reward/mean": 0.7109375, "rewards/format_reward/std": 0.2643453776836395, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "step": 31 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 618.0, "completions/max_terminated_length": 618.0, "completions/mean_length": 24.046875, "completions/mean_terminated_length": 24.046875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.07142857142857142, "frac_reward_zero_std": 0.375, "grad_norm": 15.571666717529297, "learning_rate": 8.262928807620843e-07, "loss": -0.0, "num_tokens": 4152806.0, "reward": 0.5843750238418579, "reward_std": 0.2688094973564148, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.484375, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "step": 32 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1241.0, "completions/max_terminated_length": 1241.0, "completions/mean_length": 49.3125, "completions/mean_terminated_length": 49.3125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.07366071428571429, "frac_reward_zero_std": 0.0, "grad_norm": 20.355253219604492, "learning_rate": 8.135881792367685e-07, "loss": -0.0, "num_tokens": 4275074.0, "reward": 0.43203121423721313, "reward_std": 0.2896382212638855, "rewards/format_reward/mean": 0.8828125, "rewards/format_reward/std": 0.2478829026222229, "rewards/mcq_exact_match_reward/mean": 0.34375, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "step": 33 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 108.0, "completions/max_terminated_length": 108.0, "completions/mean_length": 14.640625, "completions/mean_terminated_length": 14.640625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.07589285714285714, "frac_reward_zero_std": 0.75, "grad_norm": 6.9838972091674805, "learning_rate": 8.005405736415125e-07, "loss": 0.0, "num_tokens": 4356523.0, "reward": 0.44218751788139343, "reward_std": 0.06228317320346832, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.34375, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "step": 34 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.890625, "completions/mean_terminated_length": 12.890625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.078125, "frac_reward_zero_std": 0.625, "grad_norm": 12.993794441223145, "learning_rate": 7.871643313414718e-07, "loss": 0.0, "num_tokens": 4460820.0, "reward": 0.4281250238418579, "reward_std": 0.15992169082164764, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "step": 35 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 217.0, "completions/max_terminated_length": 217.0, "completions/mean_length": 20.46875, "completions/mean_terminated_length": 20.46875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08035714285714286, "frac_reward_zero_std": 0.5, "grad_norm": 15.35739517211914, "learning_rate": 7.734740790612136e-07, "loss": 0.0, "num_tokens": 4561778.0, "reward": 0.41796875, "reward_std": 0.1846834421157837, "rewards/format_reward/mean": 0.8984375, "rewards/format_reward/std": 0.20275264978408813, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "step": 36 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 479.0, "completions/max_terminated_length": 479.0, "completions/mean_length": 20.609375, "completions/mean_terminated_length": 20.609375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.08258928571428571, "frac_reward_zero_std": 0.75, "grad_norm": 8.591170310974121, "learning_rate": 7.594847868906076e-07, "loss": 0.0, "num_tokens": 4657305.0, "reward": 0.265625, "reward_std": 0.10499105602502823, "rewards/format_reward/mean": 0.9375, "rewards/format_reward/std": 0.24397502839565277, "rewards/mcq_exact_match_reward/mean": 0.171875, "rewards/mcq_exact_match_reward/std": 0.38025420904159546, "step": 37 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.0625, "completions/mean_terminated_length": 12.0625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08482142857142858, "frac_reward_zero_std": 0.75, "grad_norm": 9.901998519897461, "learning_rate": 7.452117519152541e-07, "loss": 0.0, "num_tokens": 4774925.0, "reward": 0.2187500149011612, "reward_std": 0.1293872892856598, "rewards/format_reward/mean": 0.9375, "rewards/format_reward/std": 0.1666666716337204, "rewards/mcq_exact_match_reward/mean": 0.125, "rewards/mcq_exact_match_reward/std": 0.3333333432674408, "step": 38 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.859375, "completions/mean_terminated_length": 12.859375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.08705357142857142, "frac_reward_zero_std": 0.625, "grad_norm": 10.899642944335938, "learning_rate": 7.306705814893439e-07, "loss": 0.0, "num_tokens": 4882996.0, "reward": 0.4125000238418579, "reward_std": 0.1552036553621292, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.3125, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "step": 39 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 320.0, "completions/max_terminated_length": 320.0, "completions/mean_length": 16.28125, "completions/mean_terminated_length": 16.28125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08928571428571429, "frac_reward_zero_std": 0.5, "grad_norm": 15.603341102600098, "learning_rate": 7.158771761692464e-07, "loss": 0.0, "num_tokens": 4981982.0, "reward": 0.3226562738418579, "reward_std": 0.15529169142246246, "rewards/format_reward/mean": 0.8828125, "rewards/format_reward/std": 0.21347814798355103, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "step": 40 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 11.328125, "completions/mean_terminated_length": 11.328125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.09151785714285714, "frac_reward_zero_std": 0.75, "grad_norm": 8.335301399230957, "learning_rate": 7.008477123264847e-07, "loss": -0.0, "num_tokens": 5110307.0, "reward": 0.3226562738418579, "reward_std": 0.1017967164516449, "rewards/format_reward/mean": 0.8828125, "rewards/format_reward/std": 0.21347814798355103, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "step": 41 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 13.0, "completions/mean_terminated_length": 13.0, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.09375, "frac_reward_zero_std": 0.5, "grad_norm": 12.084622383117676, "learning_rate": 6.855986244591103e-07, "loss": -0.0, "num_tokens": 5197755.0, "reward": 0.4750000238418579, "reward_std": 0.22461533546447754, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.375, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "step": 42 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.953125, "completions/mean_terminated_length": 12.953125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.09598214285714286, "frac_reward_zero_std": 0.75, "grad_norm": 6.176346302032471, "learning_rate": 6.701465872208216e-07, "loss": -0.0, "num_tokens": 5283112.0, "reward": 0.5531250238418579, "reward_std": 0.11100947856903076, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 43 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 13.0, "completions/mean_terminated_length": 13.0, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.09821428571428571, "frac_reward_zero_std": 0.5, "grad_norm": 20.082441329956055, "learning_rate": 6.545084971874736e-07, "loss": 0.0, "num_tokens": 5383936.0, "reward": 0.6937500238418579, "reward_std": 0.22461533546447754, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.59375, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "step": 44 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.9375, "completions/mean_terminated_length": 12.9375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.10044642857142858, "frac_reward_zero_std": 0.75, "grad_norm": 7.762583255767822, "learning_rate": 6.387014543809223e-07, "loss": -0.0, "num_tokens": 5480140.0, "reward": 0.7562500238418579, "reward_std": 0.0883883461356163, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.65625, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "step": 45 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 248.0, "completions/max_terminated_length": 248.0, "completions/mean_length": 16.65625, "completions/mean_terminated_length": 16.65625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.10267857142857142, "frac_reward_zero_std": 0.5, "grad_norm": 11.080392837524414, "learning_rate": 6.227427435703995e-07, "loss": 0.0, "num_tokens": 5594078.0, "reward": 0.6765625476837158, "reward_std": 0.22852949798107147, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.578125, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "step": 46 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 14.0, "completions/max_terminated_length": 14.0, "completions/mean_length": 12.921875, "completions/mean_terminated_length": 12.921875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.10491071428571429, "frac_reward_zero_std": 0.625, "grad_norm": 9.526235580444336, "learning_rate": 6.066498153718734e-07, "loss": -0.0, "num_tokens": 5713409.0, "reward": 0.45859378576278687, "reward_std": 0.18893101811408997, "rewards/format_reward/mean": 0.9921875, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.359375, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "step": 47 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 148.0, "completions/max_terminated_length": 148.0, "completions/mean_length": 16.296875, "completions/mean_terminated_length": 16.296875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.10714285714285714, "frac_reward_zero_std": 0.125, "grad_norm": 19.795419692993164, "learning_rate": 5.90440267166055e-07, "loss": 0.0, "num_tokens": 5822916.0, "reward": 0.4414062798023224, "reward_std": 0.3522634506225586, "rewards/format_reward/mean": 0.9765625, "rewards/format_reward/std": 0.13886408507823944, "rewards/mcq_exact_match_reward/mean": 0.34375, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "step": 48 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 45.0, "completions/max_terminated_length": 45.0, "completions/mean_length": 13.5, "completions/mean_terminated_length": 13.5, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.109375, "frac_reward_zero_std": 0.75, "grad_norm": 9.227639198303223, "learning_rate": 5.741318238559209e-07, "loss": 0.0, "num_tokens": 5905756.0, "reward": 0.5531250238418579, "reward_std": 0.12255740165710449, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 49 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 13.0, "completions/mean_terminated_length": 13.0, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.11160714285714286, "frac_reward_zero_std": 0.875, "grad_norm": 10.652731895446777, "learning_rate": 5.577423184847931e-07, "loss": 0.0, "num_tokens": 6007676.0, "reward": 0.5218750238418579, "reward_std": 0.0646936446428299, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.421875, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "step": 50 } ], "logging_steps": 1, "max_steps": 100, "num_input_tokens_seen": 6007676, "num_train_epochs": 1, "save_steps": 50, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": false }, "attributes": {} } }, "total_flos": 0.0, "train_batch_size": 1, "trial_name": null, "trial_params": null }