{ "best_global_step": null, "best_metric": null, "best_model_checkpoint": null, "epoch": 0.22321428571428573, "eval_steps": 500, "global_step": 100, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1990.0, "completions/mean_length": 538.546875, "completions/mean_terminated_length": 514.5873413085938, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.002232142857142857, "frac_reward_zero_std": 0.0, "grad_norm": 5.357117652893066, "learning_rate": 0.0, "loss": -0.0, "num_tokens": 112787.0, "reward": 0.30156248807907104, "reward_std": 0.23618578910827637, "rewards/format_reward/mean": 0.359375, "rewards/format_reward/std": 0.22658175230026245, "rewards/mcq_exact_match_reward/mean": 0.265625, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "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": 1528.0, "completions/max_terminated_length": 1528.0, "completions/mean_length": 472.5625, "completions/mean_terminated_length": 472.5625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.004464285714285714, "frac_reward_zero_std": 0.0, "grad_norm": 6.3122878074646, "learning_rate": 2e-07, "loss": -0.0, "num_tokens": 229703.0, "reward": 0.2749999761581421, "reward_std": 0.3567870259284973, "rewards/format_reward/mean": 0.25, "rewards/format_reward/std": 0.2519763112068176, "rewards/mcq_exact_match_reward/mean": 0.25, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "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": 0.0, "completions/max_length": 1485.0, "completions/max_terminated_length": 1485.0, "completions/mean_length": 529.59375, "completions/mean_terminated_length": 529.59375, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "epoch": 0.006696428571428571, "frac_reward_zero_std": 0.0, "grad_norm": 4.27625846862793, "learning_rate": 4e-07, "loss": 0.0, "num_tokens": 334789.0, "reward": 0.2554687261581421, "reward_std": 0.40990781784057617, "rewards/format_reward/mean": 0.3671875, "rewards/format_reward/std": 0.28510910272598267, "rewards/mcq_exact_match_reward/mean": 0.21875, "rewards/mcq_exact_match_reward/std": 0.4166666865348816, "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": 1175.0, "completions/mean_length": 357.5625, "completions/mean_terminated_length": 330.73016357421875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.008928571428571428, "frac_reward_zero_std": 0.125, "grad_norm": 23.33127212524414, "learning_rate": 6e-07, "loss": 0.0, "num_tokens": 458113.0, "reward": 0.27265623211860657, "reward_std": 0.284196674823761, "rewards/format_reward/mean": 0.3828125, "rewards/format_reward/std": 0.2780628502368927, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1768.0, "completions/mean_length": 566.59375, "completions/mean_terminated_length": 518.8064575195312, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.011160714285714286, "frac_reward_zero_std": 0.0, "grad_norm": 6.540904998779297, "learning_rate": 8e-07, "loss": 0.0, "num_tokens": 592039.0, "reward": 0.3968749940395355, "reward_std": 0.4285091459751129, "rewards/format_reward/mean": 0.375, "rewards/format_reward/std": 0.2357022762298584, "rewards/mcq_exact_match_reward/mean": 0.359375, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "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": 0.0, "completions/max_length": 1872.0, "completions/max_terminated_length": 1872.0, "completions/mean_length": 469.515625, "completions/mean_terminated_length": 469.515625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.013392857142857142, "frac_reward_zero_std": 0.0, "grad_norm": 11.142348289489746, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 708496.0, "reward": 0.36328125, "reward_std": 0.4375847578048706, "rewards/format_reward/mean": 0.3515625, "rewards/format_reward/std": 0.29113471508026123, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1686.0, "completions/mean_length": 506.09375, "completions/mean_terminated_length": 481.61907958984375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.015625, "frac_reward_zero_std": 0.0, "grad_norm": 7.097582817077637, "learning_rate": 9.99726628670463e-07, "loss": 0.0, "num_tokens": 826550.0, "reward": 0.46562498807907104, "reward_std": 0.361045241355896, "rewards/format_reward/mean": 0.4375, "rewards/format_reward/std": 0.2597312331199646, "rewards/mcq_exact_match_reward/mean": 0.421875, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1335.0, "completions/mean_length": 493.46875, "completions/mean_terminated_length": 468.7936706542969, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.017857142857142856, "frac_reward_zero_std": 0.0, "grad_norm": 10.78849983215332, "learning_rate": 9.989068136093872e-07, "loss": -0.0, "num_tokens": 950020.0, "reward": 0.4062499701976776, "reward_std": 0.43662554025650024, "rewards/format_reward/mean": 0.3125, "rewards/format_reward/std": 0.24397502839565277, "rewards/mcq_exact_match_reward/mean": 0.375, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 0.0625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1590.0, "completions/mean_length": 573.0, "completions/mean_terminated_length": 474.66668701171875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.020089285714285716, "frac_reward_zero_std": 0.0, "grad_norm": 15.594640731811523, "learning_rate": 9.975414512725056e-07, "loss": -0.0, "num_tokens": 1076804.0, "reward": 0.5062499642372131, "reward_std": 0.4312324523925781, "rewards/format_reward/mean": 0.375, "rewards/format_reward/std": 0.2357022762298584, "rewards/mcq_exact_match_reward/mean": 0.46875, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "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": 1214.0, "completions/mean_length": 438.484375, "completions/mean_terminated_length": 412.9365234375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.022321428571428572, "frac_reward_zero_std": 0.125, "grad_norm": 13.5233736038208, "learning_rate": 9.956320346634875e-07, "loss": -0.0, "num_tokens": 1191307.0, "reward": 0.3671874701976776, "reward_std": 0.3545979857444763, "rewards/format_reward/mean": 0.390625, "rewards/format_reward/std": 0.2083333432674408, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "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": 0.0, "completions/max_length": 1973.0, "completions/max_terminated_length": 1973.0, "completions/mean_length": 430.59375, "completions/mean_terminated_length": 430.59375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.024553571428571428, "frac_reward_zero_std": 0.0, "grad_norm": 6.149971008300781, "learning_rate": 9.931806517013612e-07, "loss": 0.0, "num_tokens": 1319913.0, "reward": 0.28203123807907104, "reward_std": 0.3983161449432373, "rewards/format_reward/mean": 0.4765625, "rewards/format_reward/std": 0.25865477323532104, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "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": 1246.0, "completions/max_terminated_length": 1246.0, "completions/mean_length": 423.96875, "completions/mean_terminated_length": 423.96875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.026785714285714284, "frac_reward_zero_std": 0.0, "grad_norm": 14.51206111907959, "learning_rate": 9.901899829374047e-07, "loss": -0.0, "num_tokens": 1450607.0, "reward": 0.2164062261581421, "reward_std": 0.32999470829963684, "rewards/format_reward/mean": 0.4453125, "rewards/format_reward/std": 0.20152568817138672, "rewards/mcq_exact_match_reward/mean": 0.171875, "rewards/mcq_exact_match_reward/std": 0.38025420904159546, "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": 0.0625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1615.0, "completions/mean_length": 610.75, "completions/mean_terminated_length": 514.933349609375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.029017857142857144, "frac_reward_zero_std": 0.0, "grad_norm": 15.69182014465332, "learning_rate": 9.866632986240029e-07, "loss": -0.0, "num_tokens": 1610687.0, "reward": 0.21015623211860657, "reward_std": 0.2368287742137909, "rewards/format_reward/mean": 0.3828125, "rewards/format_reward/std": 0.2634054720401764, "rewards/mcq_exact_match_reward/mean": 0.171875, "rewards/mcq_exact_match_reward/std": 0.38025420904159546, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1177.0, "completions/mean_length": 271.390625, "completions/mean_terminated_length": 243.19049072265625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.03125, "frac_reward_zero_std": 0.125, "grad_norm": 17.6346492767334, "learning_rate": 9.826044551386742e-07, "loss": 0.0, "num_tokens": 1715944.0, "reward": 0.3101562261581421, "reward_std": 0.3299850821495056, "rewards/format_reward/mean": 0.4453125, "rewards/format_reward/std": 0.26899561285972595, "rewards/mcq_exact_match_reward/mean": 0.265625, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1551.0, "completions/mean_length": 314.53125, "completions/mean_terminated_length": 287.0158996582031, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.033482142857142856, "frac_reward_zero_std": 0.375, "grad_norm": 9.082239151000977, "learning_rate": 9.780178907671788e-07, "loss": -0.0, "num_tokens": 1837930.0, "reward": 0.3453124761581421, "reward_std": 0.20426981151103973, "rewards/format_reward/mean": 0.484375, "rewards/format_reward/std": 0.17747680842876434, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "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": 0.0, "completions/max_length": 749.0, "completions/max_terminated_length": 749.0, "completions/mean_length": 70.03125, "completions/mean_terminated_length": 70.03125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.03571428571428571, "frac_reward_zero_std": 0.375, "grad_norm": 15.018674850463867, "learning_rate": 9.729086208503173e-07, "loss": -0.0, "num_tokens": 1933756.0, "reward": 0.6445311903953552, "reward_std": 0.27968671917915344, "rewards/format_reward/mean": 0.5078125, "rewards/format_reward/std": 0.1406387835741043, "rewards/mcq_exact_match_reward/mean": 0.59375, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "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": 0.0, "completions/max_length": 1400.0, "completions/max_terminated_length": 1400.0, "completions/mean_length": 109.015625, "completions/mean_terminated_length": 109.015625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.03794642857142857, "frac_reward_zero_std": 0.625, "grad_norm": 11.6762056350708, "learning_rate": 9.672822322997304e-07, "loss": 0.0, "num_tokens": 2044469.0, "reward": 0.14296874403953552, "reward_std": 0.11109752207994461, "rewards/format_reward/mean": 0.4921875, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.09375, "rewards/mcq_exact_match_reward/std": 0.29378482699394226, "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": 0.0, "completions/max_length": 6.0, "completions/max_terminated_length": 6.0, "completions/mean_length": 6.0, "completions/mean_terminated_length": 6.0, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.04017857142857143, "frac_reward_zero_std": 0.625, "grad_norm": 14.160598754882812, "learning_rate": 9.611448774886923e-07, "loss": -0.0, "num_tokens": 2139605.0, "reward": 0.4249999523162842, "reward_std": 0.17570310831069946, "rewards/format_reward/mean": 0.5, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.375, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 0.0, "completions/max_length": 721.0, "completions/max_terminated_length": 721.0, "completions/mean_length": 27.453125, "completions/mean_terminated_length": 27.453125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.04241071428571429, "frac_reward_zero_std": 0.375, "grad_norm": 23.57636833190918, "learning_rate": 9.545032675245813e-07, "loss": 0.0, "num_tokens": 2234258.0, "reward": 0.5679687261581421, "reward_std": 0.23467989265918732, "rewards/format_reward/mean": 0.5234375, "rewards/format_reward/std": 0.10652101784944534, "rewards/mcq_exact_match_reward/mean": 0.515625, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "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": 0.0, "completions/max_length": 579.0, "completions/max_terminated_length": 579.0, "completions/mean_length": 33.765625, "completions/mean_terminated_length": 33.765625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.044642857142857144, "frac_reward_zero_std": 0.5, "grad_norm": 20.334789276123047, "learning_rate": 9.473646649103817e-07, "loss": 0.0, "num_tokens": 2350395.0, "reward": 0.28749996423721313, "reward_std": 0.17692223191261292, "rewards/format_reward/mean": 0.53125, "rewards/format_reward/std": 0.12198751419782639, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "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": 0.0, "completions/max_length": 977.0, "completions/max_terminated_length": 977.0, "completions/mean_length": 23.625, "completions/mean_terminated_length": 23.625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.046875, "frac_reward_zero_std": 0.625, "grad_norm": 10.189783096313477, "learning_rate": 9.397368756032444e-07, "loss": 0.0, "num_tokens": 2440331.0, "reward": 0.4874999523162842, "reward_std": 0.1462520956993103, "rewards/format_reward/mean": 0.5, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.4375, "rewards/mcq_exact_match_reward/std": 0.5, "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": 448.0, "completions/max_terminated_length": 448.0, "completions/mean_length": 21.625, "completions/mean_terminated_length": 21.625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.049107142857142856, "frac_reward_zero_std": 0.625, "grad_norm": 10.06070327758789, "learning_rate": 9.316282404787869e-07, "loss": -0.0, "num_tokens": 2555947.0, "reward": 0.3304687440395355, "reward_std": 0.18777307868003845, "rewards/format_reward/mean": 0.4921875, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.28125, "rewards/mcq_exact_match_reward/std": 0.4531635046005249, "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": 0.0, "completions/max_length": 550.0, "completions/max_terminated_length": 550.0, "completions/mean_length": 40.859375, "completions/mean_terminated_length": 40.859375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.05133928571428571, "frac_reward_zero_std": 0.625, "grad_norm": 9.419411659240723, "learning_rate": 9.230476262104676e-07, "loss": -0.0, "num_tokens": 2655266.0, "reward": 0.34687498211860657, "reward_std": 0.1530819833278656, "rewards/format_reward/mean": 0.5, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "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": 0.0, "completions/max_length": 115.0, "completions/max_terminated_length": 115.0, "completions/mean_length": 7.703125, "completions/mean_terminated_length": 7.703125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.05357142857142857, "frac_reward_zero_std": 0.375, "grad_norm": 23.631738662719727, "learning_rate": 9.1400441557401e-07, "loss": -0.0, "num_tokens": 2753767.0, "reward": 0.5187499523162842, "reward_std": 0.2961388826370239, "rewards/format_reward/mean": 0.5, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.46875, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "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": 0.0, "completions/max_length": 331.0, "completions/max_terminated_length": 331.0, "completions/mean_length": 21.5625, "completions/mean_terminated_length": 21.5625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.05580357142857143, "frac_reward_zero_std": 0.625, "grad_norm": 32.62712478637695, "learning_rate": 9.045084971874737e-07, "loss": -0.0, "num_tokens": 2854067.0, "reward": 0.7687499523162842, "reward_std": 0.16675157845020294, "rewards/format_reward/mean": 0.5, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.71875, "rewards/mcq_exact_match_reward/std": 0.4531635046005249, "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": 0.0, "completions/max_length": 749.0, "completions/max_terminated_length": 749.0, "completions/mean_length": 46.765625, "completions/mean_terminated_length": 46.765625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.05803571428571429, "frac_reward_zero_std": 0.625, "grad_norm": 14.9358491897583, "learning_rate": 8.945702546981968e-07, "loss": -0.0, "num_tokens": 2941572.0, "reward": 0.4273437261581421, "reward_std": 0.16922587156295776, "rewards/format_reward/mean": 0.5234375, "rewards/format_reward/std": 0.13886408507823944, "rewards/mcq_exact_match_reward/mean": 0.375, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 473.0, "completions/mean_length": 55.65625, "completions/mean_terminated_length": 24.031747817993164, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.060267857142857144, "frac_reward_zero_std": 0.125, "grad_norm": 17.296993255615234, "learning_rate": 8.842005554284295e-07, "loss": -0.0, "num_tokens": 3036318.0, "reward": 0.45546871423721313, "reward_std": 0.3812890648841858, "rewards/format_reward/mean": 0.4921875, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "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": 0.0, "completions/max_length": 118.0, "completions/max_terminated_length": 118.0, "completions/mean_length": 9.265625, "completions/mean_terminated_length": 9.265625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0625, "frac_reward_zero_std": 0.375, "grad_norm": 22.46074676513672, "learning_rate": 8.734107384920769e-07, "loss": 0.0, "num_tokens": 3126295.0, "reward": 0.7382811903953552, "reward_std": 0.240484818816185, "rewards/format_reward/mean": 0.5078125, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.6875, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "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": 919.0, "completions/max_terminated_length": 919.0, "completions/mean_length": 34.875, "completions/mean_terminated_length": 34.875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.06473214285714286, "frac_reward_zero_std": 0.625, "grad_norm": 26.231388092041016, "learning_rate": 8.622126023955445e-07, "loss": 0.0, "num_tokens": 3218455.0, "reward": 0.524218738079071, "reward_std": 0.06296665221452713, "rewards/format_reward/mean": 0.5546875, "rewards/format_reward/std": 0.15728822350502014, "rewards/mcq_exact_match_reward/mean": 0.46875, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "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": 267.0, "completions/max_terminated_length": 267.0, "completions/mean_length": 27.46875, "completions/mean_terminated_length": 27.46875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.06696428571428571, "frac_reward_zero_std": 0.625, "grad_norm": 12.698409080505371, "learning_rate": 8.506183921362442e-07, "loss": -0.0, "num_tokens": 3300133.0, "reward": 0.42578125, "reward_std": 0.09059804677963257, "rewards/format_reward/mean": 0.5078125, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.375, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 295.0, "completions/max_terminated_length": 295.0, "completions/mean_length": 20.296875, "completions/mean_terminated_length": 20.296875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.06919642857142858, "frac_reward_zero_std": 0.5, "grad_norm": 25.770362854003906, "learning_rate": 8.386407858128706e-07, "loss": -0.0, "num_tokens": 3392480.0, "reward": 0.6601561903953552, "reward_std": 0.2199951410293579, "rewards/format_reward/mean": 0.5078125, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.609375, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "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": 58.0, "completions/max_terminated_length": 58.0, "completions/mean_length": 9.203125, "completions/mean_terminated_length": 9.203125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.07142857142857142, "frac_reward_zero_std": 0.125, "grad_norm": 25.56088638305664, "learning_rate": 8.262928807620843e-07, "loss": -0.0, "num_tokens": 3487341.0, "reward": 0.4843749701976776, "reward_std": 0.35943299531936646, "rewards/format_reward/mean": 0.625, "rewards/format_reward/std": 0.2182178944349289, "rewards/mcq_exact_match_reward/mean": 0.421875, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "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": 1187.0, "completions/max_terminated_length": 1187.0, "completions/mean_length": 74.703125, "completions/mean_terminated_length": 74.703125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.07366071428571429, "frac_reward_zero_std": 0.375, "grad_norm": 16.140928268432617, "learning_rate": 8.135881792367685e-07, "loss": -0.0, "num_tokens": 3597818.0, "reward": 0.6210936903953552, "reward_std": 0.15935926139354706, "rewards/format_reward/mean": 0.5859375, "rewards/format_reward/std": 0.19012710452079773, "rewards/mcq_exact_match_reward/mean": 0.5625, "rewards/mcq_exact_match_reward/std": 0.5, "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": 521.0, "completions/max_terminated_length": 521.0, "completions/mean_length": 78.921875, "completions/mean_terminated_length": 78.921875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.07589285714285714, "frac_reward_zero_std": 0.5, "grad_norm": 36.14888381958008, "learning_rate": 8.005405736415125e-07, "loss": 0.0, "num_tokens": 3699197.0, "reward": 1.0531249046325684, "reward_std": 0.052852895110845566, "rewards/format_reward/mean": 0.6875, "rewards/format_reward/std": 0.24397502839565277, "rewards/mcq_exact_match_reward/mean": 0.984375, "rewards/mcq_exact_match_reward/std": 0.125, "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": 701.0, "completions/max_terminated_length": 701.0, "completions/mean_length": 129.09375, "completions/mean_terminated_length": 129.09375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.078125, "frac_reward_zero_std": 0.0, "grad_norm": 38.96165084838867, "learning_rate": 7.871643313414718e-07, "loss": 0.0, "num_tokens": 3785907.0, "reward": 0.581250011920929, "reward_std": 0.19127598404884338, "rewards/format_reward/mean": 0.8125, "rewards/format_reward/std": 0.24397502839565277, "rewards/mcq_exact_match_reward/mean": 0.5, "rewards/mcq_exact_match_reward/std": 0.5039526224136353, "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": 1201.0, "completions/max_terminated_length": 1201.0, "completions/mean_length": 78.640625, "completions/mean_terminated_length": 78.640625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08035714285714286, "frac_reward_zero_std": 0.25, "grad_norm": 23.0406436920166, "learning_rate": 7.734740790612136e-07, "loss": 0.0, "num_tokens": 3888292.0, "reward": 0.43828126788139343, "reward_std": 0.23249360918998718, "rewards/format_reward/mean": 0.9453125, "rewards/format_reward/std": 0.180765300989151, "rewards/mcq_exact_match_reward/mean": 0.34375, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "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": 1375.0, "completions/max_terminated_length": 1375.0, "completions/mean_length": 107.90625, "completions/mean_terminated_length": 107.90625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08258928571428571, "frac_reward_zero_std": 0.0, "grad_norm": 29.743480682373047, "learning_rate": 7.594847868906076e-07, "loss": 0.0, "num_tokens": 3997214.0, "reward": 0.43437501788139343, "reward_std": 0.3827434182167053, "rewards/format_reward/mean": 0.90625, "rewards/format_reward/std": 0.19669894874095917, "rewards/mcq_exact_match_reward/mean": 0.34375, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "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": 594.0, "completions/max_terminated_length": 594.0, "completions/mean_length": 65.421875, "completions/mean_terminated_length": 65.421875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08482142857142858, "frac_reward_zero_std": 0.25, "grad_norm": 22.222890853881836, "learning_rate": 7.452117519152541e-07, "loss": -0.0, "num_tokens": 4083289.0, "reward": 0.8304687738418579, "reward_std": 0.27762293815612793, "rewards/format_reward/mean": 0.9609375, "rewards/format_reward/std": 0.16194961965084076, "rewards/mcq_exact_match_reward/mean": 0.734375, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "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": 612.0, "completions/max_terminated_length": 612.0, "completions/mean_length": 104.859375, "completions/mean_terminated_length": 104.859375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08705357142857142, "frac_reward_zero_std": 0.25, "grad_norm": 15.55250358581543, "learning_rate": 7.306705814893439e-07, "loss": 0.0, "num_tokens": 4180168.0, "reward": 0.4671874940395355, "reward_std": 0.22475478053092957, "rewards/format_reward/mean": 0.921875, "rewards/format_reward/std": 0.20351573824882507, "rewards/mcq_exact_match_reward/mean": 0.375, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 968.0, "completions/max_terminated_length": 968.0, "completions/mean_length": 113.1875, "completions/mean_terminated_length": 113.1875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08928571428571429, "frac_reward_zero_std": 0.375, "grad_norm": 5.488610744476318, "learning_rate": 7.158771761692464e-07, "loss": 0.0, "num_tokens": 4310372.0, "reward": 0.48359376192092896, "reward_std": 0.19265598058700562, "rewards/format_reward/mean": 0.9296875, "rewards/format_reward/std": 0.1751912236213684, "rewards/mcq_exact_match_reward/mean": 0.390625, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "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": 1017.0, "completions/max_terminated_length": 1017.0, "completions/mean_length": 122.84375, "completions/mean_terminated_length": 122.84375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.09151785714285714, "frac_reward_zero_std": 0.375, "grad_norm": 7.906015396118164, "learning_rate": 7.008477123264847e-07, "loss": -0.0, "num_tokens": 4400802.0, "reward": 0.34921878576278687, "reward_std": 0.17898640036582947, "rewards/format_reward/mean": 0.9921875, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.25, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 104.0, "completions/mean_length": 47.65625, "completions/mean_terminated_length": 15.904763221740723, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.09375, "frac_reward_zero_std": 0.5, "grad_norm": 18.783111572265625, "learning_rate": 6.855986244591103e-07, "loss": 0.0, "num_tokens": 4490844.0, "reward": 0.8484375476837158, "reward_std": 0.15067152678966522, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.75, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "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": 386.0, "completions/max_terminated_length": 386.0, "completions/mean_length": 44.71875, "completions/mean_terminated_length": 44.71875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.09598214285714286, "frac_reward_zero_std": 0.75, "grad_norm": 10.034747123718262, "learning_rate": 6.701465872208216e-07, "loss": -0.0, "num_tokens": 4581682.0, "reward": 0.6312500238418579, "reward_std": 0.0883883461356163, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.53125, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "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": 169.0, "completions/max_terminated_length": 169.0, "completions/mean_length": 28.828125, "completions/mean_terminated_length": 28.828125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.09821428571428571, "frac_reward_zero_std": 0.625, "grad_norm": 14.294143676757812, "learning_rate": 6.545084971874736e-07, "loss": -0.0, "num_tokens": 4687551.0, "reward": 0.4593750238418579, "reward_std": 0.1530819982290268, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.359375, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "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": 138.0, "completions/max_terminated_length": 138.0, "completions/mean_length": 20.34375, "completions/mean_terminated_length": 20.34375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.10044642857142858, "frac_reward_zero_std": 0.5, "grad_norm": 19.66748046875, "learning_rate": 6.387014543809223e-07, "loss": 0.0, "num_tokens": 4780245.0, "reward": 0.5515625476837158, "reward_std": 0.21536517143249512, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 288.0, "completions/mean_length": 63.03125, "completions/mean_terminated_length": 31.52381134033203, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.10267857142857142, "frac_reward_zero_std": 0.75, "grad_norm": 8.040264129638672, "learning_rate": 6.227427435703995e-07, "loss": 0.0, "num_tokens": 4885047.0, "reward": 0.7250000238418579, "reward_std": 0.0883883461356163, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.625, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 560.0, "completions/mean_length": 54.53125, "completions/mean_terminated_length": 22.88888931274414, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.10491071428571429, "frac_reward_zero_std": 0.75, "grad_norm": 4.303212642669678, "learning_rate": 6.066498153718734e-07, "loss": 0.0, "num_tokens": 4981921.0, "reward": 0.6140625476837158, "reward_std": 0.04861358925700188, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.515625, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "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": 423.0, "completions/max_terminated_length": 423.0, "completions/mean_length": 36.078125, "completions/mean_terminated_length": 36.078125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.10714285714285714, "frac_reward_zero_std": 0.5, "grad_norm": 14.959633827209473, "learning_rate": 5.90440267166055e-07, "loss": -0.0, "num_tokens": 5097710.0, "reward": 0.7867187261581421, "reward_std": 0.14846181869506836, "rewards/format_reward/mean": 0.9921875, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.6875, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "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": 1036.0, "completions/max_terminated_length": 1036.0, "completions/mean_length": 39.046875, "completions/mean_terminated_length": 39.046875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.109375, "frac_reward_zero_std": 0.625, "grad_norm": 13.03551197052002, "learning_rate": 5.741318238559209e-07, "loss": 0.0, "num_tokens": 5197737.0, "reward": 0.6468750238418579, "reward_std": 0.15992169082164764, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.546875, "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": 240.0, "completions/max_terminated_length": 240.0, "completions/mean_length": 22.109375, "completions/mean_terminated_length": 22.109375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.11160714285714286, "frac_reward_zero_std": 0.625, "grad_norm": 7.772871017456055, "learning_rate": 5.577423184847931e-07, "loss": 0.0, "num_tokens": 5273552.0, "reward": 0.5062500238418579, "reward_std": 0.1462521106004715, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "step": 50 }, { "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": 733.0, "completions/max_terminated_length": 733.0, "completions/mean_length": 27.578125, "completions/mean_terminated_length": 27.578125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.11383928571428571, "frac_reward_zero_std": 0.5, "grad_norm": 23.071674346923828, "learning_rate": 5.412896727361662e-07, "loss": 0.0, "num_tokens": 5372205.0, "reward": 0.6296875476837158, "reward_std": 0.2201814204454422, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.08768405020236969, "rewards/mcq_exact_match_reward/mean": 0.53125, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "step": 51 }, { "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": 527.0, "completions/max_terminated_length": 527.0, "completions/mean_length": 32.796875, "completions/mean_terminated_length": 32.796875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.11607142857142858, "frac_reward_zero_std": 0.5, "grad_norm": 22.4808349609375, "learning_rate": 5.247918773366111e-07, "loss": 0.0, "num_tokens": 5448896.0, "reward": 0.7703125476837158, "reward_std": 0.2490595281124115, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.671875, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "step": 52 }, { "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": 590.0, "completions/max_terminated_length": 590.0, "completions/mean_length": 38.53125, "completions/mean_terminated_length": 38.53125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.11830357142857142, "frac_reward_zero_std": 0.375, "grad_norm": 10.820849418640137, "learning_rate": 5.082669723831793e-07, "loss": 0.0, "num_tokens": 5528994.0, "reward": 0.7859375476837158, "reward_std": 0.23762542009353638, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.6875, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "step": 53 }, { "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": 18.0, "completions/max_terminated_length": 18.0, "completions/mean_length": 12.734375, "completions/mean_terminated_length": 12.734375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.12053571428571429, "frac_reward_zero_std": 0.875, "grad_norm": 5.189952373504639, "learning_rate": 4.917330276168208e-07, "loss": -0.0, "num_tokens": 5625761.0, "reward": 0.4906250238418579, "reward_std": 0.04419417306780815, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.390625, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "step": 54 }, { "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": 227.0, "completions/mean_length": 57.6875, "completions/mean_terminated_length": 26.095239639282227, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.12276785714285714, "frac_reward_zero_std": 0.625, "grad_norm": 19.476482391357422, "learning_rate": 4.752081226633888e-07, "loss": -0.0, "num_tokens": 5725301.0, "reward": 0.47187501192092896, "reward_std": 0.1898023784160614, "rewards/format_reward/mean": 0.96875, "rewards/format_reward/std": 0.17536810040473938, "rewards/mcq_exact_match_reward/mean": 0.375, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "step": 55 }, { "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": 75.0, "completions/max_terminated_length": 75.0, "completions/mean_length": 14.640625, "completions/mean_terminated_length": 14.640625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.125, "frac_reward_zero_std": 0.625, "grad_norm": 8.712614059448242, "learning_rate": 4.5871032726383385e-07, "loss": 0.0, "num_tokens": 5811646.0, "reward": 0.6375000476837158, "reward_std": 0.1167958527803421, "rewards/format_reward/mean": 0.90625, "rewards/format_reward/std": 0.29378482699394226, "rewards/mcq_exact_match_reward/mean": 0.546875, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 56 }, { "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": 373.0, "completions/max_terminated_length": 373.0, "completions/mean_length": 21.515625, "completions/mean_terminated_length": 21.515625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.12723214285714285, "frac_reward_zero_std": 0.375, "grad_norm": 31.22998046875, "learning_rate": 4.4225768151520694e-07, "loss": 0.0, "num_tokens": 5888127.0, "reward": 0.8031250238418579, "reward_std": 0.28930896520614624, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.703125, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "step": 57 }, { "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": 16.0, "completions/max_terminated_length": 16.0, "completions/mean_length": 12.0625, "completions/mean_terminated_length": 12.0625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.12946428571428573, "frac_reward_zero_std": 1.0, "grad_norm": 0.0, "learning_rate": 4.258681761440789e-07, "loss": 0.0, "num_tokens": 5981651.0, "reward": 0.3500000238418579, "reward_std": 0.0, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.25, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "step": 58 }, { "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": 16.0, "completions/max_terminated_length": 16.0, "completions/mean_length": 12.15625, "completions/mean_terminated_length": 12.15625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.13169642857142858, "frac_reward_zero_std": 0.75, "grad_norm": 12.346508026123047, "learning_rate": 4.095597328339452e-07, "loss": 0.0, "num_tokens": 6074389.0, "reward": 0.8500000238418579, "reward_std": 0.0883883461356163, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.75, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "step": 59 }, { "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": 758.0, "completions/max_terminated_length": 758.0, "completions/mean_length": 32.109375, "completions/mean_terminated_length": 32.109375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.13392857142857142, "frac_reward_zero_std": 0.625, "grad_norm": 11.00341510772705, "learning_rate": 3.9335018462812664e-07, "loss": 0.0, "num_tokens": 6165892.0, "reward": 0.6937500238418579, "reward_std": 0.16675157845020294, "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": 60 }, { "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": 159.0, "completions/max_terminated_length": 159.0, "completions/mean_length": 14.453125, "completions/mean_terminated_length": 14.453125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.13616071428571427, "frac_reward_zero_std": 0.75, "grad_norm": 10.805044174194336, "learning_rate": 3.772572564296004e-07, "loss": -0.0, "num_tokens": 6265945.0, "reward": 0.5687500238418579, "reward_std": 0.10888782143592834, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.46875, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "step": 61 }, { "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": 63.0, "completions/max_terminated_length": 63.0, "completions/mean_length": 14.328125, "completions/mean_terminated_length": 14.328125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.13839285714285715, "frac_reward_zero_std": 0.875, "grad_norm": 11.665460586547852, "learning_rate": 3.612985456190778e-07, "loss": 0.0, "num_tokens": 6332734.0, "reward": 0.6312500238418579, "reward_std": 0.0578637570142746, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.53125, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "step": 62 }, { "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": 12.0, "completions/max_terminated_length": 12.0, "completions/mean_length": 12.0, "completions/mean_terminated_length": 12.0, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.140625, "frac_reward_zero_std": 0.875, "grad_norm": 13.824284553527832, "learning_rate": 3.454915028125263e-07, "loss": -0.0, "num_tokens": 6401782.0, "reward": 0.7093750238418579, "reward_std": 0.04419417306780815, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.609375, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "step": 63 }, { "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": 103.0, "completions/max_terminated_length": 103.0, "completions/mean_length": 13.671875, "completions/mean_terminated_length": 13.671875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.14285714285714285, "frac_reward_zero_std": 0.875, "grad_norm": 6.85416316986084, "learning_rate": 3.2985341277917846e-07, "loss": 0.0, "num_tokens": 6497601.0, "reward": 0.7562500238418579, "reward_std": 0.0578637570142746, "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": 64 }, { "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": 167.0, "completions/max_terminated_length": 167.0, "completions/mean_length": 15.609375, "completions/mean_terminated_length": 15.609375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.14508928571428573, "frac_reward_zero_std": 0.875, "grad_norm": 8.13846206665039, "learning_rate": 3.1440137554088953e-07, "loss": 0.0, "num_tokens": 6598784.0, "reward": 0.6781250238418579, "reward_std": 0.0646936446428299, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.578125, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "step": 65 }, { "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": 17.0, "completions/max_terminated_length": 17.0, "completions/mean_length": 12.078125, "completions/mean_terminated_length": 12.078125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.14732142857142858, "frac_reward_zero_std": 0.75, "grad_norm": 10.682966232299805, "learning_rate": 2.9915228767351535e-07, "loss": 0.0, "num_tokens": 6675989.0, "reward": 0.5531250238418579, "reward_std": 0.10205793380737305, "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": 66 }, { "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": 130.0, "completions/max_terminated_length": 130.0, "completions/mean_length": 13.84375, "completions/mean_terminated_length": 13.84375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.14955357142857142, "frac_reward_zero_std": 0.875, "grad_norm": 7.651659965515137, "learning_rate": 2.841228238307536e-07, "loss": 0.0, "num_tokens": 6784259.0, "reward": 0.6937500238418579, "reward_std": 0.0578637570142746, "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": 67 }, { "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": 84.0, "completions/max_terminated_length": 84.0, "completions/mean_length": 13.125, "completions/mean_terminated_length": 13.125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.15178571428571427, "frac_reward_zero_std": 0.875, "grad_norm": 13.613445281982422, "learning_rate": 2.6932941851065615e-07, "loss": -0.0, "num_tokens": 6875019.0, "reward": 0.6156250238418579, "reward_std": 0.04419417306780815, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.515625, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "step": 68 }, { "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": 12.0, "completions/max_terminated_length": 12.0, "completions/mean_length": 12.0, "completions/mean_terminated_length": 12.0, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.15401785714285715, "frac_reward_zero_std": 0.875, "grad_norm": 10.362603187561035, "learning_rate": 2.547882480847461e-07, "loss": 0.0, "num_tokens": 6974267.0, "reward": 0.6625000238418579, "reward_std": 0.06681530922651291, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.5625, "rewards/mcq_exact_match_reward/std": 0.5, "step": 69 }, { "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": 175.0, "completions/max_terminated_length": 175.0, "completions/mean_length": 14.546875, "completions/mean_terminated_length": 14.546875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.15625, "frac_reward_zero_std": 0.75, "grad_norm": 16.298009872436523, "learning_rate": 2.4051521310939254e-07, "loss": 0.0, "num_tokens": 7092542.0, "reward": 0.7093750238418579, "reward_std": 0.12255740165710449, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.609375, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "step": 70 }, { "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": 757.0, "completions/max_terminated_length": 757.0, "completions/mean_length": 34.875, "completions/mean_terminated_length": 34.875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.15848214285714285, "frac_reward_zero_std": 1.0, "grad_norm": 0.0, "learning_rate": 2.2652592093878665e-07, "loss": 0.0, "num_tokens": 7222646.0, "reward": 0.6000000238418579, "reward_std": 0.0, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.5, "rewards/mcq_exact_match_reward/std": 0.5039526224136353, "step": 71 }, { "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": 116.0, "completions/max_terminated_length": 116.0, "completions/mean_length": 13.625, "completions/mean_terminated_length": 13.625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.16071428571428573, "frac_reward_zero_std": 0.875, "grad_norm": 4.760712623596191, "learning_rate": 2.128356686585282e-07, "loss": 0.0, "num_tokens": 7298638.0, "reward": 0.7562500238418579, "reward_std": 0.0578637570142746, "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": 72 }, { "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": 29.0, "completions/max_terminated_length": 29.0, "completions/mean_length": 13.46875, "completions/mean_terminated_length": 13.46875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.16294642857142858, "frac_reward_zero_std": 0.875, "grad_norm": 4.456165313720703, "learning_rate": 1.9945942635848745e-07, "loss": -0.0, "num_tokens": 7370340.0, "reward": 0.7281250357627869, "reward_std": 0.04419417306780815, "rewards/format_reward/mean": 0.875, "rewards/format_reward/std": 0.3333333432674408, "rewards/mcq_exact_match_reward/mean": 0.640625, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "step": 73 }, { "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": 18.0, "completions/max_terminated_length": 18.0, "completions/mean_length": 12.09375, "completions/mean_terminated_length": 12.09375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.16517857142857142, "frac_reward_zero_std": 0.75, "grad_norm": 28.935243606567383, "learning_rate": 1.8641182076323148e-07, "loss": 0.0, "num_tokens": 7443554.0, "reward": 0.6468750238418579, "reward_std": 0.10205793380737305, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.546875, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 74 }, { "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": 17.0, "completions/max_terminated_length": 17.0, "completions/mean_length": 12.28125, "completions/mean_terminated_length": 12.28125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.16741071428571427, "frac_reward_zero_std": 1.0, "grad_norm": 0.0, "learning_rate": 1.7370711923791564e-07, "loss": 0.0, "num_tokens": 7576996.0, "reward": 0.6000000238418579, "reward_std": 0.0, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.5, "rewards/mcq_exact_match_reward/std": 0.5039526224136353, "step": 75 }, { "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": 12.015625, "completions/mean_terminated_length": 12.015625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.16964285714285715, "frac_reward_zero_std": 0.5, "grad_norm": 29.626203536987305, "learning_rate": 1.6135921418712955e-07, "loss": 0.0, "num_tokens": 7681661.0, "reward": 0.7093750238418579, "reward_std": 0.23144522309303284, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.609375, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "step": 76 }, { "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": 12.0, "completions/max_terminated_length": 12.0, "completions/mean_length": 12.0, "completions/mean_terminated_length": 12.0, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.171875, "frac_reward_zero_std": 0.875, "grad_norm": 9.431557655334473, "learning_rate": 1.493816078637557e-07, "loss": -0.0, "num_tokens": 7757197.0, "reward": 0.8656250238418579, "reward_std": 0.04419417306780815, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.765625, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "step": 77 }, { "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": 12.0, "completions/max_terminated_length": 12.0, "completions/mean_length": 12.0, "completions/mean_terminated_length": 12.0, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.17410714285714285, "frac_reward_zero_std": 0.875, "grad_norm": 34.71562576293945, "learning_rate": 1.3778739760445552e-07, "loss": 0.0, "num_tokens": 7832973.0, "reward": 0.6468750238418579, "reward_std": 0.0646936446428299, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.546875, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 78 }, { "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": 17.0, "completions/max_terminated_length": 17.0, "completions/mean_length": 12.078125, "completions/mean_terminated_length": 12.078125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.17633928571428573, "frac_reward_zero_std": 1.0, "grad_norm": 0.0, "learning_rate": 1.2658926150792322e-07, "loss": 0.0, "num_tokens": 7919842.0, "reward": 1.100000023841858, "reward_std": 0.0, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 1.0, "rewards/mcq_exact_match_reward/std": 0.0, "step": 79 }, { "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": 405.0, "completions/max_terminated_length": 405.0, "completions/mean_length": 21.828125, "completions/mean_terminated_length": 21.828125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.17857142857142858, "frac_reward_zero_std": 0.875, "grad_norm": 4.398412704467773, "learning_rate": 1.1579944457157059e-07, "loss": 0.0, "num_tokens": 8028367.0, "reward": 0.4906250238418579, "reward_std": 0.04419417306780815, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.390625, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "step": 80 }, { "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": 12.0, "completions/max_terminated_length": 12.0, "completions/mean_length": 12.0, "completions/mean_terminated_length": 12.0, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.18080357142857142, "frac_reward_zero_std": 1.0, "grad_norm": 0.0, "learning_rate": 1.0542974530180327e-07, "loss": 0.0, "num_tokens": 8112679.0, "reward": 0.8500000238418579, "reward_std": 0.0, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.75, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "step": 81 }, { "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": 233.0, "completions/max_terminated_length": 233.0, "completions/mean_length": 15.640625, "completions/mean_terminated_length": 15.640625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.18303571428571427, "frac_reward_zero_std": 0.75, "grad_norm": 11.36860179901123, "learning_rate": 9.549150281252632e-08, "loss": 0.0, "num_tokens": 8215616.0, "reward": 0.6781250238418579, "reward_std": 0.12255740165710449, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.578125, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "step": 82 }, { "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": 12.0, "completions/max_terminated_length": 12.0, "completions/mean_length": 12.0, "completions/mean_terminated_length": 12.0, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.18526785714285715, "frac_reward_zero_std": 0.875, "grad_norm": 9.782796859741211, "learning_rate": 8.599558442598998e-08, "loss": 0.0, "num_tokens": 8295304.0, "reward": 0.8968750238418579, "reward_std": 0.0646936446428299, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.796875, "rewards/mcq_exact_match_reward/std": 0.40550529956817627, "step": 83 }, { "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": 12.0, "completions/max_terminated_length": 12.0, "completions/mean_length": 12.0, "completions/mean_terminated_length": 12.0, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1875, "frac_reward_zero_std": 0.75, "grad_norm": 11.062796592712402, "learning_rate": 7.695237378953224e-08, "loss": 0.0, "num_tokens": 8389848.0, "reward": 0.6156250238418579, "reward_std": 0.10205793380737305, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.515625, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "step": 84 }, { "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": 97.0, "completions/max_terminated_length": 97.0, "completions/mean_length": 13.015625, "completions/mean_terminated_length": 13.015625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.18973214285714285, "frac_reward_zero_std": 0.875, "grad_norm": 12.023233413696289, "learning_rate": 6.837175952121304e-08, "loss": 0.0, "num_tokens": 8494849.0, "reward": 0.8460937738418579, "reward_std": 0.003234688425436616, "rewards/format_reward/mean": 0.9609375, "rewards/format_reward/std": 0.13524486124515533, "rewards/mcq_exact_match_reward/mean": 0.75, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "step": 85 }, { "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": 12.0, "completions/max_terminated_length": 12.0, "completions/mean_length": 12.0, "completions/mean_terminated_length": 12.0, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.19196428571428573, "frac_reward_zero_std": 0.75, "grad_norm": 17.247573852539062, "learning_rate": 6.026312439675551e-08, "loss": 0.0, "num_tokens": 8574217.0, "reward": 0.5687500238418579, "reward_std": 0.10888782143592834, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.46875, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "step": 86 }, { "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": 21.0, "completions/max_terminated_length": 21.0, "completions/mean_length": 12.78125, "completions/mean_terminated_length": 12.78125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.19419642857142858, "frac_reward_zero_std": 0.75, "grad_norm": 9.152361869812012, "learning_rate": 5.263533508961826e-08, "loss": 0.0, "num_tokens": 8654035.0, "reward": 0.596875011920929, "reward_std": 0.12151388823986053, "rewards/format_reward/mean": 0.96875, "rewards/format_reward/std": 0.17536810040473938, "rewards/mcq_exact_match_reward/mean": 0.5, "rewards/mcq_exact_match_reward/std": 0.5039526224136353, "step": 87 }, { "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": 12.0, "completions/max_terminated_length": 12.0, "completions/mean_length": 12.0, "completions/mean_terminated_length": 12.0, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.19642857142857142, "frac_reward_zero_std": 0.75, "grad_norm": 12.168911933898926, "learning_rate": 4.549673247541874e-08, "loss": 0.0, "num_tokens": 8733459.0, "reward": 0.6937500238418579, "reward_std": 0.1246790662407875, "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": 88 }, { "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": 62.0, "completions/max_terminated_length": 62.0, "completions/mean_length": 12.78125, "completions/mean_terminated_length": 12.78125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.19866071428571427, "frac_reward_zero_std": 0.875, "grad_norm": 9.789281845092773, "learning_rate": 3.8855122511307626e-08, "loss": 0.0, "num_tokens": 8832629.0, "reward": 0.5843750238418579, "reward_std": 0.04419417306780815, "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": 89 }, { "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": 12.0, "completions/max_terminated_length": 12.0, "completions/mean_length": 12.0, "completions/mean_terminated_length": 12.0, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.20089285714285715, "frac_reward_zero_std": 0.875, "grad_norm": 10.417126655578613, "learning_rate": 3.271776770026963e-08, "loss": -0.0, "num_tokens": 8909925.0, "reward": 0.6156250238418579, "reward_std": 0.04419417306780815, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.515625, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "step": 90 }, { "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": 12.0, "completions/max_terminated_length": 12.0, "completions/mean_length": 12.0, "completions/mean_terminated_length": 12.0, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.203125, "frac_reward_zero_std": 0.625, "grad_norm": 20.833892822265625, "learning_rate": 2.7091379149682682e-08, "loss": 0.0, "num_tokens": 9013213.0, "reward": 0.7718750238418579, "reward_std": 0.1804211586713791, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.671875, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "step": 91 }, { "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": 12.0, "completions/max_terminated_length": 12.0, "completions/mean_length": 12.0, "completions/mean_terminated_length": 12.0, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.20535714285714285, "frac_reward_zero_std": 0.75, "grad_norm": 16.622377395629883, "learning_rate": 2.1982109232821176e-08, "loss": 0.0, "num_tokens": 9105005.0, "reward": 0.6156250238418579, "reward_std": 0.12255740165710449, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.515625, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "step": 92 }, { "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": 60.0, "completions/max_terminated_length": 60.0, "completions/mean_length": 12.75, "completions/mean_terminated_length": 12.75, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.20758928571428573, "frac_reward_zero_std": 0.875, "grad_norm": 8.70872974395752, "learning_rate": 1.7395544861325718e-08, "loss": -0.0, "num_tokens": 9185501.0, "reward": 0.7406250238418579, "reward_std": 0.04419417306780815, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.640625, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "step": 93 }, { "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": 16.0, "completions/max_terminated_length": 16.0, "completions/mean_length": 12.0625, "completions/mean_terminated_length": 12.0625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.20982142857142858, "frac_reward_zero_std": 0.75, "grad_norm": 14.33220386505127, "learning_rate": 1.3336701375997127e-08, "loss": 0.0, "num_tokens": 9280585.0, "reward": 0.7875000238418579, "reward_std": 0.10888782143592834, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.6875, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "step": 94 }, { "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": 128.0, "completions/max_terminated_length": 128.0, "completions/mean_length": 13.8125, "completions/mean_terminated_length": 13.8125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.21205357142857142, "frac_reward_zero_std": 1.0, "grad_norm": 0.0, "learning_rate": 9.810017062595321e-09, "loss": 0.0, "num_tokens": 9362941.0, "reward": 0.9750000238418579, "reward_std": 0.0, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.875, "rewards/mcq_exact_match_reward/std": 0.3333333432674408, "step": 95 }, { "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": 18.0, "completions/max_terminated_length": 18.0, "completions/mean_length": 12.796875, "completions/mean_terminated_length": 12.796875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.21428571428571427, "frac_reward_zero_std": 1.0, "grad_norm": 0.0, "learning_rate": 6.819348298638839e-09, "loss": 0.0, "num_tokens": 9469168.0, "reward": 0.5875000357627869, "reward_std": 0.0, "rewards/format_reward/mean": 0.875, "rewards/format_reward/std": 0.3333333432674408, "rewards/mcq_exact_match_reward/mean": 0.5, "rewards/mcq_exact_match_reward/std": 0.5039526224136353, "step": 96 }, { "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": 583.0, "completions/max_terminated_length": 583.0, "completions/mean_length": 22.390625, "completions/mean_terminated_length": 22.390625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.21651785714285715, "frac_reward_zero_std": 0.875, "grad_norm": 3.0157196521759033, "learning_rate": 4.367965336512403e-09, "loss": 0.0, "num_tokens": 9584393.0, "reward": 0.4906250238418579, "reward_std": 0.04419417306780815, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.390625, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "step": 97 }, { "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": 449.0, "completions/max_terminated_length": 449.0, "completions/mean_length": 18.921875, "completions/mean_terminated_length": 18.921875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.21875, "frac_reward_zero_std": 0.875, "grad_norm": 11.126572608947754, "learning_rate": 2.458548727494292e-09, "loss": 0.0, "num_tokens": 9677436.0, "reward": 0.6781250238418579, "reward_std": 0.0646936446428299, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.578125, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "step": 98 }, { "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": 96.0, "completions/max_terminated_length": 96.0, "completions/mean_length": 13.3125, "completions/mean_terminated_length": 13.3125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.22098214285714285, "frac_reward_zero_std": 1.0, "grad_norm": 0.0, "learning_rate": 1.0931863906127325e-09, "loss": 0.0, "num_tokens": 9763336.0, "reward": 0.9750000238418579, "reward_std": 0.0, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.875, "rewards/mcq_exact_match_reward/std": 0.3333333432674408, "step": 99 }, { "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": 25.0, "completions/max_terminated_length": 25.0, "completions/mean_length": 13.671875, "completions/mean_terminated_length": 13.671875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.22321428571428573, "frac_reward_zero_std": 0.875, "grad_norm": 8.683713912963867, "learning_rate": 2.733713295369755e-10, "loss": 0.0, "num_tokens": 9859443.0, "reward": 0.9312500357627869, "reward_std": 0.0578637570142746, "rewards/format_reward/mean": 0.875, "rewards/format_reward/std": 0.3333333432674408, "rewards/mcq_exact_match_reward/mean": 0.84375, "rewards/mcq_exact_match_reward/std": 0.36596253514289856, "step": 100 } ], "logging_steps": 1, "max_steps": 100, "num_input_tokens_seen": 9859443, "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": true }, "attributes": {} } }, "total_flos": 0.0, "train_batch_size": 1, "trial_name": null, "trial_params": null }