{ "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": 0.0, "completions/max_length": 1847.0, "completions/max_terminated_length": 1847.0, "completions/mean_length": 618.921875, "completions/mean_terminated_length": 618.921875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.002232142857142857, "frac_reward_zero_std": 0.0, "grad_norm": 3.1935155391693115, "learning_rate": 0.0, "loss": -0.0, "num_tokens": 155971.0, "reward": 0.30156245827674866, "reward_std": 0.3944985270500183, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1534.0, "completions/mean_length": 621.78125, "completions/mean_terminated_length": 575.774169921875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.004464285714285714, "frac_reward_zero_std": 0.0, "grad_norm": 3.723464012145996, "learning_rate": 2e-07, "loss": -0.0, "num_tokens": 298325.0, "reward": 0.36796873807907104, "reward_std": 0.40465426445007324, "rewards/format_reward/mean": 0.3984375, "rewards/format_reward/std": 0.2387082874774933, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1761.0, "completions/mean_length": 497.125, "completions/mean_terminated_length": 472.5079650878906, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.006696428571428571, "frac_reward_zero_std": 0.125, "grad_norm": 5.9803361892700195, "learning_rate": 4e-07, "loss": 0.0, "num_tokens": 428061.0, "reward": 0.17656248807907104, "reward_std": 0.1757061928510666, "rewards/format_reward/mean": 0.359375, "rewards/format_reward/std": 0.24346621334552765, "rewards/mcq_exact_match_reward/mean": 0.140625, "rewards/mcq_exact_match_reward/std": 0.3503824472427368, "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": 0.046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1712.0, "completions/mean_length": 621.515625, "completions/mean_terminated_length": 551.360595703125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.008928571428571428, "frac_reward_zero_std": 0.0, "grad_norm": 8.650753021240234, "learning_rate": 6e-07, "loss": 0.0, "num_tokens": 582222.0, "reward": 0.2890624701976776, "reward_std": 0.3460876941680908, "rewards/format_reward/mean": 0.390625, "rewards/format_reward/std": 0.2083333432674408, "rewards/mcq_exact_match_reward/mean": 0.25, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "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": 1926.0, "completions/max_terminated_length": 1926.0, "completions/mean_length": 519.265625, "completions/mean_terminated_length": 519.265625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.011160714285714286, "frac_reward_zero_std": 0.25, "grad_norm": 5.728817462921143, "learning_rate": 8e-07, "loss": 0.0, "num_tokens": 727527.0, "reward": 0.4023437201976776, "reward_std": 0.3245325982570648, "rewards/format_reward/mean": 0.4296875, "rewards/format_reward/std": 0.1751912236213684, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1278.0, "completions/mean_length": 400.5625, "completions/mean_terminated_length": 374.4127197265625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.013392857142857142, "frac_reward_zero_std": 0.125, "grad_norm": 8.398402214050293, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 838107.0, "reward": 0.4453125, "reward_std": 0.25657644867897034, "rewards/format_reward/mean": 0.390625, "rewards/format_reward/std": 0.28824523091316223, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "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": 0.0, "completions/max_length": 1415.0, "completions/max_terminated_length": 1415.0, "completions/mean_length": 522.875, "completions/mean_terminated_length": 522.875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.015625, "frac_reward_zero_std": 0.0, "grad_norm": 5.590754508972168, "learning_rate": 9.99726628670463e-07, "loss": 0.0, "num_tokens": 951107.0, "reward": 0.30390623211860657, "reward_std": 0.35596728324890137, "rewards/format_reward/mean": 0.3828125, "rewards/format_reward/std": 0.2313210517168045, "rewards/mcq_exact_match_reward/mean": 0.265625, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "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": 0.0625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1859.0, "completions/mean_length": 618.671875, "completions/mean_terminated_length": 523.3833618164062, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.017857142857142856, "frac_reward_zero_std": 0.0, "grad_norm": 4.199244022369385, "learning_rate": 9.989068136093872e-07, "loss": -0.0, "num_tokens": 1088646.0, "reward": 0.508593738079071, "reward_std": 0.3985562324523926, "rewards/format_reward/mean": 0.3984375, "rewards/format_reward/std": 0.25479042530059814, "rewards/mcq_exact_match_reward/mean": 0.46875, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "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": 0.0, "completions/max_length": 1359.0, "completions/max_terminated_length": 1359.0, "completions/mean_length": 427.140625, "completions/mean_terminated_length": 427.140625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.020089285714285716, "frac_reward_zero_std": 0.0, "grad_norm": 8.815071105957031, "learning_rate": 9.975414512725056e-07, "loss": 0.0, "num_tokens": 1201455.0, "reward": 0.2640624940395355, "reward_std": 0.36402058601379395, "rewards/format_reward/mean": 0.453125, "rewards/format_reward/std": 0.2916666865348816, "rewards/mcq_exact_match_reward/mean": 0.21875, "rewards/mcq_exact_match_reward/std": 0.4166666865348816, "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": 0.0, "completions/max_length": 1958.0, "completions/max_terminated_length": 1958.0, "completions/mean_length": 503.765625, "completions/mean_terminated_length": 503.765625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.022321428571428572, "frac_reward_zero_std": 0.25, "grad_norm": 7.586217403411865, "learning_rate": 9.956320346634875e-07, "loss": -0.0, "num_tokens": 1331104.0, "reward": 0.32890623807907104, "reward_std": 0.3252153694629669, "rewards/format_reward/mean": 0.4765625, "rewards/format_reward/std": 0.10652101784944534, "rewards/mcq_exact_match_reward/mean": 0.28125, "rewards/mcq_exact_match_reward/std": 0.4531635046005249, "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": 1293.0, "completions/max_terminated_length": 1293.0, "completions/mean_length": 407.5625, "completions/mean_terminated_length": 407.5625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.024553571428571428, "frac_reward_zero_std": 0.25, "grad_norm": 16.40622329711914, "learning_rate": 9.931806517013612e-07, "loss": -0.0, "num_tokens": 1446412.0, "reward": 0.296875, "reward_std": 0.281208336353302, "rewards/format_reward/mean": 0.46875, "rewards/format_reward/std": 0.21593283116817474, "rewards/mcq_exact_match_reward/mean": 0.25, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1098.0, "completions/mean_length": 290.3125, "completions/mean_terminated_length": 262.4127197265625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.026785714285714284, "frac_reward_zero_std": 0.125, "grad_norm": 27.468236923217773, "learning_rate": 9.901899829374047e-07, "loss": 0.0, "num_tokens": 1586944.0, "reward": 0.29374998807907104, "reward_std": 0.27264729142189026, "rewards/format_reward/mean": 0.4375, "rewards/format_reward/std": 0.18898223340511322, "rewards/mcq_exact_match_reward/mean": 0.25, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "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": 0.0, "completions/max_length": 1656.0, "completions/max_terminated_length": 1656.0, "completions/mean_length": 256.328125, "completions/mean_terminated_length": 256.328125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.029017857142857144, "frac_reward_zero_std": 0.25, "grad_norm": 9.994261741638184, "learning_rate": 9.866632986240029e-07, "loss": 0.0, "num_tokens": 1704029.0, "reward": 0.18593749403953552, "reward_std": 0.19405311346054077, "rewards/format_reward/mean": 0.453125, "rewards/format_reward/std": 0.14689241349697113, "rewards/mcq_exact_match_reward/mean": 0.140625, "rewards/mcq_exact_match_reward/std": 0.3503824472427368, "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": 0.0, "completions/max_length": 1780.0, "completions/max_terminated_length": 1780.0, "completions/mean_length": 232.90625, "completions/mean_terminated_length": 232.90625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.03125, "frac_reward_zero_std": 0.25, "grad_norm": 11.842218399047852, "learning_rate": 9.826044551386742e-07, "loss": -0.0, "num_tokens": 1807007.0, "reward": 0.21171873807907104, "reward_std": 0.25511986017227173, "rewards/format_reward/mean": 0.5546875, "rewards/format_reward/std": 0.2203386276960373, "rewards/mcq_exact_match_reward/mean": 0.15625, "rewards/mcq_exact_match_reward/std": 0.36596253514289856, "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": 933.0, "completions/max_terminated_length": 933.0, "completions/mean_length": 73.234375, "completions/mean_terminated_length": 73.234375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.033482142857142856, "frac_reward_zero_std": 0.25, "grad_norm": 15.33292007446289, "learning_rate": 9.780178907671788e-07, "loss": -0.0, "num_tokens": 1903646.0, "reward": 0.6117187142372131, "reward_std": 0.33579903841018677, "rewards/format_reward/mean": 0.4921875, "rewards/format_reward/std": 0.10882443934679031, "rewards/mcq_exact_match_reward/mean": 0.5625, "rewards/mcq_exact_match_reward/std": 0.5, "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": 572.0, "completions/max_terminated_length": 572.0, "completions/mean_length": 135.515625, "completions/mean_terminated_length": 135.515625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.03571428571428571, "frac_reward_zero_std": 0.625, "grad_norm": 1.767112135887146, "learning_rate": 9.729086208503173e-07, "loss": -0.0, "num_tokens": 2000023.0, "reward": 0.22578124701976776, "reward_std": 0.10782171785831451, "rewards/format_reward/mean": 0.5390625, "rewards/format_reward/std": 0.20518454909324646, "rewards/mcq_exact_match_reward/mean": 0.171875, "rewards/mcq_exact_match_reward/std": 0.38025420904159546, "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": 741.0, "completions/max_terminated_length": 741.0, "completions/mean_length": 43.453125, "completions/mean_terminated_length": 43.453125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.03794642857142857, "frac_reward_zero_std": 0.75, "grad_norm": 14.93751335144043, "learning_rate": 9.672822322997304e-07, "loss": -0.0, "num_tokens": 2100004.0, "reward": 0.26953125, "reward_std": 0.06007346138358116, "rewards/format_reward/mean": 0.5078125, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.21875, "rewards/mcq_exact_match_reward/std": 0.4166666865348816, "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": 959.0, "completions/max_terminated_length": 959.0, "completions/mean_length": 79.984375, "completions/mean_terminated_length": 79.984375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.04017857142857143, "frac_reward_zero_std": 0.5, "grad_norm": 35.31517028808594, "learning_rate": 9.611448774886923e-07, "loss": 0.0, "num_tokens": 2218307.0, "reward": 0.19765624403953552, "reward_std": 0.053563401103019714, "rewards/format_reward/mean": 0.5703125, "rewards/format_reward/std": 0.23345555365085602, "rewards/mcq_exact_match_reward/mean": 0.140625, "rewards/mcq_exact_match_reward/std": 0.3503824472427368, "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": 358.0, "completions/max_terminated_length": 358.0, "completions/mean_length": 38.046875, "completions/mean_terminated_length": 38.046875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.04241071428571429, "frac_reward_zero_std": 0.625, "grad_norm": 9.192914962768555, "learning_rate": 9.545032675245813e-07, "loss": 0.0, "num_tokens": 2315190.0, "reward": 0.20859374105930328, "reward_std": 0.09303848445415497, "rewards/format_reward/mean": 0.5234375, "rewards/format_reward/std": 0.13886408507823944, "rewards/mcq_exact_match_reward/mean": 0.15625, "rewards/mcq_exact_match_reward/std": 0.36596253514289856, "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": 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": 0.044642857142857144, "frac_reward_zero_std": 0.75, "grad_norm": 10.627936363220215, "learning_rate": 9.473646649103817e-07, "loss": -0.0, "num_tokens": 2406830.0, "reward": 0.31562498211860657, "reward_std": 0.1315089464187622, "rewards/format_reward/mean": 0.5, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.265625, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "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": 954.0, "completions/max_terminated_length": 954.0, "completions/mean_length": 64.6875, "completions/mean_terminated_length": 64.6875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.046875, "frac_reward_zero_std": 0.625, "grad_norm": 15.555487632751465, "learning_rate": 9.397368756032444e-07, "loss": -0.0, "num_tokens": 2540074.0, "reward": 0.53515625, "reward_std": 0.10684418678283691, "rewards/format_reward/mean": 0.5078125, "rewards/format_reward/std": 0.10882443934679031, "rewards/mcq_exact_match_reward/mean": 0.484375, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "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": 351.0, "completions/max_terminated_length": 351.0, "completions/mean_length": 42.421875, "completions/mean_terminated_length": 42.421875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.049107142857142856, "frac_reward_zero_std": 0.625, "grad_norm": 23.398914337158203, "learning_rate": 9.316282404787869e-07, "loss": 0.0, "num_tokens": 2633949.0, "reward": 0.5703125, "reward_std": 0.1606418639421463, "rewards/format_reward/mean": 0.546875, "rewards/format_reward/std": 0.14689241349697113, "rewards/mcq_exact_match_reward/mean": 0.515625, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "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": 543.0, "completions/max_terminated_length": 543.0, "completions/mean_length": 60.734375, "completions/mean_terminated_length": 60.734375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.05133928571428571, "frac_reward_zero_std": 0.25, "grad_norm": 18.460739135742188, "learning_rate": 9.230476262104676e-07, "loss": 0.0, "num_tokens": 2729492.0, "reward": 0.29765623807907104, "reward_std": 0.21675951778888702, "rewards/format_reward/mean": 0.6328125, "rewards/format_reward/std": 0.23974503576755524, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "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": 441.0, "completions/max_terminated_length": 441.0, "completions/mean_length": 24.171875, "completions/mean_terminated_length": 24.171875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.05357142857142857, "frac_reward_zero_std": 0.5, "grad_norm": 13.40369987487793, "learning_rate": 9.1400441557401e-07, "loss": 0.0, "num_tokens": 2820175.0, "reward": 0.2476562261581421, "reward_std": 0.14756274223327637, "rewards/format_reward/mean": 0.6015625, "rewards/format_reward/std": 0.20275264978408813, "rewards/mcq_exact_match_reward/mean": 0.1875, "rewards/mcq_exact_match_reward/std": 0.39339789748191833, "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": 558.0, "completions/max_terminated_length": 558.0, "completions/mean_length": 32.125, "completions/mean_terminated_length": 32.125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.05580357142857143, "frac_reward_zero_std": 0.375, "grad_norm": 22.2802791595459, "learning_rate": 9.045084971874737e-07, "loss": -0.0, "num_tokens": 2898175.0, "reward": 0.46015623211860657, "reward_std": 0.2491144984960556, "rewards/format_reward/mean": 0.5390625, "rewards/format_reward/std": 0.18483558297157288, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "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": 376.0, "completions/max_terminated_length": 376.0, "completions/mean_length": 11.78125, "completions/mean_terminated_length": 11.78125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.05803571428571429, "frac_reward_zero_std": 0.375, "grad_norm": 23.949974060058594, "learning_rate": 8.945702546981968e-07, "loss": -0.0, "num_tokens": 3004137.0, "reward": 0.5656249523162842, "reward_std": 0.25726157426834106, "rewards/format_reward/mean": 0.5, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.515625, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "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": 0.0, "completions/max_length": 261.0, "completions/max_terminated_length": 261.0, "completions/mean_length": 18.078125, "completions/mean_terminated_length": 18.078125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.060267857142857144, "frac_reward_zero_std": 0.375, "grad_norm": 13.89456558227539, "learning_rate": 8.842005554284295e-07, "loss": 0.0, "num_tokens": 3087174.0, "reward": 0.5843750238418579, "reward_std": 0.24831002950668335, "rewards/format_reward/mean": 0.6875, "rewards/format_reward/std": 0.24397502839565277, "rewards/mcq_exact_match_reward/mean": 0.515625, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "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": 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": 0.0625, "frac_reward_zero_std": 0.75, "grad_norm": 13.80541706085205, "learning_rate": 8.734107384920769e-07, "loss": 0.0, "num_tokens": 3198614.0, "reward": 0.5812499523162842, "reward_std": 0.0883883386850357, "rewards/format_reward/mean": 0.5, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.53125, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 7.171875, "completions/mean_terminated_length": 7.171875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.06473214285714286, "frac_reward_zero_std": 0.5, "grad_norm": 22.059465408325195, "learning_rate": 8.622126023955445e-07, "loss": -0.0, "num_tokens": 3318513.0, "reward": 0.49609375, "reward_std": 0.1584383249282837, "rewards/format_reward/mean": 0.5859375, "rewards/format_reward/std": 0.19012710452079773, "rewards/mcq_exact_match_reward/mean": 0.4375, "rewards/mcq_exact_match_reward/std": 0.5, "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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 7.078125, "completions/mean_terminated_length": 7.078125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.06696428571428571, "frac_reward_zero_std": 0.375, "grad_norm": 24.122072219848633, "learning_rate": 8.506183921362442e-07, "loss": -0.0, "num_tokens": 3401286.0, "reward": 0.3070312440395355, "reward_std": 0.2200935184955597, "rewards/format_reward/mean": 0.5703125, "rewards/format_reward/std": 0.1751912236213684, "rewards/mcq_exact_match_reward/mean": 0.25, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "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": 147.0, "completions/max_terminated_length": 147.0, "completions/mean_length": 10.75, "completions/mean_terminated_length": 10.75, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.06919642857142858, "frac_reward_zero_std": 0.625, "grad_norm": 20.449478149414062, "learning_rate": 8.386407858128706e-07, "loss": 0.0, "num_tokens": 3514062.0, "reward": 0.33515626192092896, "reward_std": 0.19047920405864716, "rewards/format_reward/mean": 0.6953125, "rewards/format_reward/std": 0.24587368965148926, "rewards/mcq_exact_match_reward/mean": 0.265625, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 6.21875, "completions/mean_terminated_length": 6.21875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.07142857142857142, "frac_reward_zero_std": 0.75, "grad_norm": 17.937305450439453, "learning_rate": 8.262928807620843e-07, "loss": -0.0, "num_tokens": 3610476.0, "reward": 0.5671874284744263, "reward_std": 0.04708735644817352, "rewards/format_reward/mean": 0.515625, "rewards/format_reward/std": 0.08768405020236969, "rewards/mcq_exact_match_reward/mean": 0.515625, "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": 632.0, "completions/max_terminated_length": 632.0, "completions/mean_length": 29.0625, "completions/mean_terminated_length": 29.0625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.07366071428571429, "frac_reward_zero_std": 0.25, "grad_norm": 36.576087951660156, "learning_rate": 8.135881792367685e-07, "loss": 0.0, "num_tokens": 3674536.0, "reward": 0.6148437261581421, "reward_std": 0.29603251814842224, "rewards/format_reward/mean": 0.8359375, "rewards/format_reward/std": 0.2366211861371994, "rewards/mcq_exact_match_reward/mean": 0.53125, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "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": 218.0, "completions/max_terminated_length": 218.0, "completions/mean_length": 13.375, "completions/mean_terminated_length": 13.375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.07589285714285714, "frac_reward_zero_std": 0.125, "grad_norm": 39.33657455444336, "learning_rate": 8.005405736415125e-07, "loss": 0.0, "num_tokens": 3771560.0, "reward": 0.49531251192092896, "reward_std": 0.3159250020980835, "rewards/format_reward/mean": 0.734375, "rewards/format_reward/std": 0.2514837086200714, "rewards/mcq_exact_match_reward/mean": 0.421875, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "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": 220.0, "completions/max_terminated_length": 220.0, "completions/mean_length": 15.25, "completions/mean_terminated_length": 15.25, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.078125, "frac_reward_zero_std": 0.125, "grad_norm": 41.215240478515625, "learning_rate": 7.871643313414718e-07, "loss": 0.0, "num_tokens": 3869144.0, "reward": 0.663281261920929, "reward_std": 0.27762800455093384, "rewards/format_reward/mean": 0.8515625, "rewards/format_reward/std": 0.24688033759593964, "rewards/mcq_exact_match_reward/mean": 0.578125, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "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": 673.0, "completions/max_terminated_length": 673.0, "completions/mean_length": 28.1875, "completions/mean_terminated_length": 28.1875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08035714285714286, "frac_reward_zero_std": 0.0, "grad_norm": 29.907503128051758, "learning_rate": 7.734740790612136e-07, "loss": 0.0, "num_tokens": 3972940.0, "reward": 0.41718751192092896, "reward_std": 0.38989460468292236, "rewards/format_reward/mean": 0.890625, "rewards/format_reward/std": 0.2083333432674408, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 730.0, "completions/mean_length": 72.578125, "completions/mean_terminated_length": 41.222225189208984, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08258928571428571, "frac_reward_zero_std": 0.0, "grad_norm": 16.981098175048828, "learning_rate": 7.594847868906076e-07, "loss": 0.0, "num_tokens": 4090801.0, "reward": 0.582812488079071, "reward_std": 0.3880528211593628, "rewards/format_reward/mean": 0.828125, "rewards/format_reward/std": 0.27048972249031067, "rewards/mcq_exact_match_reward/mean": 0.5, "rewards/mcq_exact_match_reward/std": 0.5039526224136353, "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": 733.0, "completions/max_terminated_length": 733.0, "completions/mean_length": 62.90625, "completions/mean_terminated_length": 62.90625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08482142857142858, "frac_reward_zero_std": 0.25, "grad_norm": 27.475454330444336, "learning_rate": 7.452117519152541e-07, "loss": -0.0, "num_tokens": 4213259.0, "reward": 0.23359374701976776, "reward_std": 0.23486268520355225, "rewards/format_reward/mean": 0.7734375, "rewards/format_reward/std": 0.2662152051925659, "rewards/mcq_exact_match_reward/mean": 0.15625, "rewards/mcq_exact_match_reward/std": 0.36596253514289856, "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": 224.0, "completions/max_terminated_length": 224.0, "completions/mean_length": 21.359375, "completions/mean_terminated_length": 21.359375, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.08705357142857142, "frac_reward_zero_std": 0.25, "grad_norm": 14.140779495239258, "learning_rate": 7.306705814893439e-07, "loss": 0.0, "num_tokens": 4305690.0, "reward": 0.5359375476837158, "reward_std": 0.2853722870349884, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.08768405020236969, "rewards/mcq_exact_match_reward/mean": 0.4375, "rewards/mcq_exact_match_reward/std": 0.5, "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": 792.0, "completions/max_terminated_length": 792.0, "completions/mean_length": 46.15625, "completions/mean_terminated_length": 46.15625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08928571428571429, "frac_reward_zero_std": 0.125, "grad_norm": 22.493703842163086, "learning_rate": 7.158771761692464e-07, "loss": 0.0, "num_tokens": 4415060.0, "reward": 0.6328125, "reward_std": 0.39019253849983215, "rewards/format_reward/mean": 0.859375, "rewards/format_reward/std": 0.301698237657547, "rewards/mcq_exact_match_reward/mean": 0.546875, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "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": 247.0, "completions/max_terminated_length": 247.0, "completions/mean_length": 18.703125, "completions/mean_terminated_length": 18.703125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.09151785714285714, "frac_reward_zero_std": 0.25, "grad_norm": 18.93410301208496, "learning_rate": 7.008477123264847e-07, "loss": -0.0, "num_tokens": 4520105.0, "reward": 0.4703125059604645, "reward_std": 0.28388339281082153, "rewards/format_reward/mean": 0.953125, "rewards/format_reward/std": 0.1717960685491562, "rewards/mcq_exact_match_reward/mean": 0.375, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 672.0, "completions/max_terminated_length": 672.0, "completions/mean_length": 35.28125, "completions/mean_terminated_length": 35.28125, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.09375, "frac_reward_zero_std": 0.375, "grad_norm": 28.909137725830078, "learning_rate": 6.855986244591103e-07, "loss": -0.0, "num_tokens": 4620715.0, "reward": 0.5960937738418579, "reward_std": 0.1878252476453781, "rewards/format_reward/mean": 0.9609375, "rewards/format_reward/std": 0.18483558297157288, "rewards/mcq_exact_match_reward/mean": 0.5, "rewards/mcq_exact_match_reward/std": 0.5039526224136353, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 327.0, "completions/mean_length": 50.90625, "completions/mean_terminated_length": 19.206350326538086, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.09598214285714286, "frac_reward_zero_std": 0.625, "grad_norm": 8.222065925598145, "learning_rate": 6.701465872208216e-07, "loss": 0.0, "num_tokens": 4718005.0, "reward": 0.37968751788139343, "reward_std": 0.15067152678966522, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.28125, "rewards/mcq_exact_match_reward/std": 0.4531635046005249, "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": 777.0, "completions/max_terminated_length": 777.0, "completions/mean_length": 25.9375, "completions/mean_terminated_length": 25.9375, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.09821428571428571, "frac_reward_zero_std": 0.625, "grad_norm": 14.55408763885498, "learning_rate": 6.545084971874736e-07, "loss": 0.0, "num_tokens": 4805409.0, "reward": 0.8171875476837158, "reward_std": 0.11330723762512207, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.71875, "rewards/mcq_exact_match_reward/std": 0.4531635046005249, "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": 20.0, "completions/max_terminated_length": 20.0, "completions/mean_length": 13.4375, "completions/mean_terminated_length": 13.4375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.10044642857142858, "frac_reward_zero_std": 0.75, "grad_norm": 10.875714302062988, "learning_rate": 6.387014543809223e-07, "loss": -0.0, "num_tokens": 4868893.0, "reward": 0.47343751788139343, "reward_std": 0.08910852670669556, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.08768405020236969, "rewards/mcq_exact_match_reward/mean": 0.375, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 1421.0, "completions/max_terminated_length": 1421.0, "completions/mean_length": 59.859375, "completions/mean_terminated_length": 59.859375, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.10267857142857142, "frac_reward_zero_std": 0.5, "grad_norm": 7.904510021209717, "learning_rate": 6.227427435703995e-07, "loss": 0.0, "num_tokens": 4949860.0, "reward": 0.30000001192092896, "reward_std": 0.15740421414375305, "rewards/format_reward/mean": 0.96875, "rewards/format_reward/std": 0.1510545015335083, "rewards/mcq_exact_match_reward/mean": 0.203125, "rewards/mcq_exact_match_reward/std": 0.40550529956817627, "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": 377.0, "completions/max_terminated_length": 377.0, "completions/mean_length": 51.984375, "completions/mean_terminated_length": 51.984375, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.10491071428571429, "frac_reward_zero_std": 0.5, "grad_norm": 8.806200981140137, "learning_rate": 6.066498153718734e-07, "loss": -0.0, "num_tokens": 5035171.0, "reward": 0.61328125, "reward_std": 0.13581719994544983, "rewards/format_reward/mean": 0.9765625, "rewards/format_reward/std": 0.10652101784944534, "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": 855.0, "completions/max_terminated_length": 855.0, "completions/mean_length": 35.125, "completions/mean_terminated_length": 35.125, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.10714285714285714, "frac_reward_zero_std": 0.5, "grad_norm": 8.751934051513672, "learning_rate": 5.90440267166055e-07, "loss": -0.0, "num_tokens": 5148395.0, "reward": 0.6929687857627869, "reward_std": 0.14846181869506836, "rewards/format_reward/mean": 0.9921875, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.59375, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "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": 943.0, "completions/max_terminated_length": 943.0, "completions/mean_length": 27.53125, "completions/mean_terminated_length": 27.53125, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.109375, "frac_reward_zero_std": 0.625, "grad_norm": 5.283928871154785, "learning_rate": 5.741318238559209e-07, "loss": 0.0, "num_tokens": 5272181.0, "reward": 0.39609378576278687, "reward_std": 0.10426764190196991, "rewards/format_reward/mean": 0.9921875, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "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": 326.0, "completions/max_terminated_length": 326.0, "completions/mean_length": 19.34375, "completions/mean_terminated_length": 19.34375, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.11160714285714286, "frac_reward_zero_std": 0.375, "grad_norm": 16.974990844726562, "learning_rate": 5.577423184847931e-07, "loss": 0.0, "num_tokens": 5360851.0, "reward": 0.42656251788139343, "reward_std": 0.21536517143249512, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "step": 50 } ], "logging_steps": 1, "max_steps": 100, "num_input_tokens_seen": 5360851, "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 }