{ "best_global_step": null, "best_metric": null, "best_model_checkpoint": null, "epoch": 0.25, "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": 0.0, "completions/max_length": 972.0, "completions/max_terminated_length": 972.0, "completions/mean_length": 432.421875, "completions/mean_terminated_length": 432.421875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0025, "frac_reward_zero_std": 0.125, "grad_norm": 6.885483741760254, "learning_rate": 0.0, "loss": -0.0, "num_tokens": 127395.0, "reward": 0.2382812350988388, "reward_std": 0.3547627329826355, "rewards/format_reward/mean": 0.3515625, "rewards/format_reward/std": 0.2302463799715042, "rewards/mcq_exact_match_reward/mean": 0.203125, "rewards/mcq_exact_match_reward/std": 0.40550529956817627, "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": 1513.0, "completions/max_terminated_length": 1513.0, "completions/mean_length": 507.921875, "completions/mean_terminated_length": 507.921875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.005, "frac_reward_zero_std": 0.125, "grad_norm": 1.3625750541687012, "learning_rate": 2e-07, "loss": 0.0, "num_tokens": 266894.0, "reward": 0.23906247317790985, "reward_std": 0.307381272315979, "rewards/format_reward/mean": 0.359375, "rewards/format_reward/std": 0.22658175230026245, "rewards/mcq_exact_match_reward/mean": 0.203125, "rewards/mcq_exact_match_reward/std": 0.40550529956817627, "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": 1351.0, "completions/mean_length": 531.109375, "completions/mean_terminated_length": 507.0317687988281, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0075, "frac_reward_zero_std": 0.0, "grad_norm": 4.586404323577881, "learning_rate": 4e-07, "loss": 0.0, "num_tokens": 368589.0, "reward": 0.2671874761581421, "reward_std": 0.3730037212371826, "rewards/format_reward/mean": 0.328125, "rewards/format_reward/std": 0.2847827076911926, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "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": 0.0, "completions/max_length": 1940.0, "completions/max_terminated_length": 1940.0, "completions/mean_length": 511.15625, "completions/mean_terminated_length": 511.15625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.01, "frac_reward_zero_std": 0.0, "grad_norm": 5.185728073120117, "learning_rate": 6e-07, "loss": -0.0, "num_tokens": 486911.0, "reward": 0.3593749701976776, "reward_std": 0.34273526072502136, "rewards/format_reward/mean": 0.3125, "rewards/format_reward/std": 0.28867512941360474, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1585.0, "completions/mean_length": 433.125, "completions/mean_terminated_length": 407.4920959472656, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.0125, "frac_reward_zero_std": 0.125, "grad_norm": 6.260329723358154, "learning_rate": 8e-07, "loss": -0.0, "num_tokens": 603519.0, "reward": 0.16484373807907104, "reward_std": 0.27282217144966125, "rewards/format_reward/mean": 0.2421875, "rewards/format_reward/std": 0.25185325741767883, "rewards/mcq_exact_match_reward/mean": 0.140625, "rewards/mcq_exact_match_reward/std": 0.3503824472427368, "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": 1663.0, "completions/mean_length": 475.421875, "completions/mean_terminated_length": 450.4603576660156, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "epoch": 0.015, "frac_reward_zero_std": 0.0, "grad_norm": 6.672026634216309, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 726482.0, "reward": 0.16796875, "reward_std": 0.23057740926742554, "rewards/format_reward/mean": 0.2734375, "rewards/format_reward/std": 0.250866562128067, "rewards/mcq_exact_match_reward/mean": 0.140625, "rewards/mcq_exact_match_reward/std": 0.3503824472427368, "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": 1156.0, "completions/mean_length": 512.171875, "completions/mean_terminated_length": 487.7936706542969, "completions/min_length": 17.0, "completions/min_terminated_length": 17.0, "epoch": 0.0175, "frac_reward_zero_std": 0.0, "grad_norm": 2.1586880683898926, "learning_rate": 9.99726628670463e-07, "loss": -0.0, "num_tokens": 855861.0, "reward": 0.32109373807907104, "reward_std": 0.34560394287109375, "rewards/format_reward/mean": 0.3984375, "rewards/format_reward/std": 0.25479042530059814, "rewards/mcq_exact_match_reward/mean": 0.28125, "rewards/mcq_exact_match_reward/std": 0.4531635046005249, "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": 1871.0, "completions/mean_length": 604.96875, "completions/mean_terminated_length": 582.0635375976562, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.02, "frac_reward_zero_std": 0.0, "grad_norm": 2.5194599628448486, "learning_rate": 9.989068136093872e-07, "loss": 0.0, "num_tokens": 1000971.0, "reward": 0.28984373807907104, "reward_std": 0.26004695892333984, "rewards/format_reward/mean": 0.3984375, "rewards/format_reward/std": 0.20275264978408813, "rewards/mcq_exact_match_reward/mean": 0.25, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "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": 0.046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1630.0, "completions/mean_length": 737.5625, "completions/mean_terminated_length": 673.11474609375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0225, "frac_reward_zero_std": 0.0, "grad_norm": 3.3560500144958496, "learning_rate": 9.975414512725056e-07, "loss": 0.0, "num_tokens": 1158791.0, "reward": 0.1812499761581421, "reward_std": 0.3075662851333618, "rewards/format_reward/mean": 0.40625, "rewards/format_reward/std": 0.21593283116817474, "rewards/mcq_exact_match_reward/mean": 0.140625, "rewards/mcq_exact_match_reward/std": 0.3503824472427368, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1701.0, "completions/mean_length": 546.578125, "completions/mean_terminated_length": 498.1451416015625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.025, "frac_reward_zero_std": 0.0, "grad_norm": 3.1816318035125732, "learning_rate": 9.956320346634875e-07, "loss": -0.0, "num_tokens": 1285484.0, "reward": 0.24531248211860657, "reward_std": 0.3608049750328064, "rewards/format_reward/mean": 0.421875, "rewards/format_reward/std": 0.2847827076911926, "rewards/mcq_exact_match_reward/mean": 0.203125, "rewards/mcq_exact_match_reward/std": 0.40550529956817627, "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": 1215.0, "completions/max_terminated_length": 1215.0, "completions/mean_length": 379.984375, "completions/mean_terminated_length": 379.984375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0275, "frac_reward_zero_std": 0.125, "grad_norm": 7.100859642028809, "learning_rate": 9.931806517013612e-07, "loss": 0.0, "num_tokens": 1386251.0, "reward": 0.4843749701976776, "reward_std": 0.3645934760570526, "rewards/format_reward/mean": 0.46875, "rewards/format_reward/std": 0.279951810836792, "rewards/mcq_exact_match_reward/mean": 0.4375, "rewards/mcq_exact_match_reward/std": 0.5, "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": 1224.0, "completions/mean_length": 561.59375, "completions/mean_terminated_length": 538.0000610351562, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.03, "frac_reward_zero_std": 0.125, "grad_norm": 3.26202130317688, "learning_rate": 9.901899829374047e-07, "loss": 0.0, "num_tokens": 1521185.0, "reward": 0.21328125894069672, "reward_std": 0.329146146774292, "rewards/format_reward/mean": 0.4140625, "rewards/format_reward/std": 0.28942593932151794, "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": 0.0, "completions/max_length": 1572.0, "completions/max_terminated_length": 1572.0, "completions/mean_length": 550.875, "completions/mean_terminated_length": 550.875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.0325, "frac_reward_zero_std": 0.0, "grad_norm": 7.087386131286621, "learning_rate": 9.866632986240029e-07, "loss": 0.0, "num_tokens": 1649577.0, "reward": 0.36406248807907104, "reward_std": 0.43930676579475403, "rewards/format_reward/mean": 0.359375, "rewards/format_reward/std": 0.2592533528804779, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "step": 13 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2013.0, "completions/mean_length": 763.84375, "completions/mean_terminated_length": 722.4193115234375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.035, "frac_reward_zero_std": 0.0, "grad_norm": 2.967664957046509, "learning_rate": 9.826044551386742e-07, "loss": 0.0, "num_tokens": 1810239.0, "reward": 0.23906248807907104, "reward_std": 0.3521445095539093, "rewards/format_reward/mean": 0.359375, "rewards/format_reward/std": 0.22658175230026245, "rewards/mcq_exact_match_reward/mean": 0.203125, "rewards/mcq_exact_match_reward/std": 0.40550529956817627, "step": 14 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1914.0, "completions/mean_length": 658.796875, "completions/mean_terminated_length": 636.74609375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.0375, "frac_reward_zero_std": 0.25, "grad_norm": 18.77581214904785, "learning_rate": 9.780178907671788e-07, "loss": 0.0, "num_tokens": 1949426.0, "reward": 0.2718749940395355, "reward_std": 0.3570863604545593, "rewards/format_reward/mean": 0.375, "rewards/format_reward/std": 0.2182178944349289, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1361.0, "completions/mean_length": 340.0625, "completions/mean_terminated_length": 312.952392578125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.04, "frac_reward_zero_std": 0.0, "grad_norm": 23.68616485595703, "learning_rate": 9.729086208503173e-07, "loss": -0.0, "num_tokens": 2048270.0, "reward": 0.18906249105930328, "reward_std": 0.20243585109710693, "rewards/format_reward/mean": 0.484375, "rewards/format_reward/std": 0.30820462107658386, "rewards/mcq_exact_match_reward/mean": 0.140625, "rewards/mcq_exact_match_reward/std": 0.3503824472427368, "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": 1144.0, "completions/max_terminated_length": 1144.0, "completions/mean_length": 434.234375, "completions/mean_terminated_length": 434.234375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0425, "frac_reward_zero_std": 0.125, "grad_norm": 8.685861587524414, "learning_rate": 9.672822322997304e-07, "loss": 0.0, "num_tokens": 2161405.0, "reward": 0.29609373211860657, "reward_std": 0.23511800169944763, "rewards/format_reward/mean": 0.4609375, "rewards/format_reward/std": 0.18483558297157288, "rewards/mcq_exact_match_reward/mean": 0.25, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "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": 0.0625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1305.0, "completions/mean_length": 709.125, "completions/mean_terminated_length": 619.86669921875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.045, "frac_reward_zero_std": 0.25, "grad_norm": 4.618074893951416, "learning_rate": 9.611448774886923e-07, "loss": 0.0, "num_tokens": 2324109.0, "reward": 0.16953124105930328, "reward_std": 0.17864511907100677, "rewards/format_reward/mean": 0.4453125, "rewards/format_reward/std": 0.20152568817138672, "rewards/mcq_exact_match_reward/mean": 0.125, "rewards/mcq_exact_match_reward/std": 0.3333333432674408, "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": 1713.0, "completions/max_terminated_length": 1713.0, "completions/mean_length": 314.46875, "completions/mean_terminated_length": 314.46875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0475, "frac_reward_zero_std": 0.375, "grad_norm": 9.724517822265625, "learning_rate": 9.545032675245813e-07, "loss": 0.0, "num_tokens": 2455171.0, "reward": 0.27031248807907104, "reward_std": 0.2369493544101715, "rewards/format_reward/mean": 0.515625, "rewards/format_reward/std": 0.1534975916147232, "rewards/mcq_exact_match_reward/mean": 0.21875, "rewards/mcq_exact_match_reward/std": 0.4166666865348816, "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": 1605.0, "completions/max_terminated_length": 1605.0, "completions/mean_length": 495.53125, "completions/mean_terminated_length": 495.53125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.05, "frac_reward_zero_std": 0.0, "grad_norm": 10.205577850341797, "learning_rate": 9.473646649103817e-07, "loss": -0.0, "num_tokens": 2586061.0, "reward": 0.2992187440395355, "reward_std": 0.40997689962387085, "rewards/format_reward/mean": 0.4921875, "rewards/format_reward/std": 0.2744719088077545, "rewards/mcq_exact_match_reward/mean": 0.25, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "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": 1628.0, "completions/max_terminated_length": 1628.0, "completions/mean_length": 203.703125, "completions/mean_terminated_length": 203.703125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0525, "frac_reward_zero_std": 0.125, "grad_norm": 21.00765609741211, "learning_rate": 9.397368756032444e-07, "loss": 0.0, "num_tokens": 2716834.0, "reward": 0.36249998211860657, "reward_std": 0.3577616214752197, "rewards/format_reward/mean": 0.5, "rewards/format_reward/std": 0.1259881556034088, "rewards/mcq_exact_match_reward/mean": 0.3125, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "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": 958.0, "completions/max_terminated_length": 958.0, "completions/mean_length": 246.671875, "completions/mean_terminated_length": 246.671875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.055, "frac_reward_zero_std": 0.25, "grad_norm": 14.727567672729492, "learning_rate": 9.316282404787869e-07, "loss": -0.0, "num_tokens": 2813885.0, "reward": 0.3828125, "reward_std": 0.35105907917022705, "rewards/format_reward/mean": 0.546875, "rewards/format_reward/std": 0.23091863095760345, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "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": 1318.0, "completions/max_terminated_length": 1318.0, "completions/mean_length": 243.296875, "completions/mean_terminated_length": 243.296875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0575, "frac_reward_zero_std": 0.125, "grad_norm": 13.531010627746582, "learning_rate": 9.230476262104676e-07, "loss": -0.0, "num_tokens": 2917120.0, "reward": 0.29765623807907104, "reward_std": 0.3336710035800934, "rewards/format_reward/mean": 0.6328125, "rewards/format_reward/std": 0.28510910272598267, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1232.0, "completions/mean_length": 266.390625, "completions/mean_terminated_length": 238.1111297607422, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.06, "frac_reward_zero_std": 0.0, "grad_norm": 22.170225143432617, "learning_rate": 9.1400441557401e-07, "loss": -0.0, "num_tokens": 3028249.0, "reward": 0.55859375, "reward_std": 0.42872872948646545, "rewards/format_reward/mean": 0.5859375, "rewards/format_reward/std": 0.2280818521976471, "rewards/mcq_exact_match_reward/mean": 0.5, "rewards/mcq_exact_match_reward/std": 0.5039526224136353, "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": 1675.0, "completions/max_terminated_length": 1675.0, "completions/mean_length": 120.84375, "completions/mean_terminated_length": 120.84375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0625, "frac_reward_zero_std": 0.125, "grad_norm": 27.930810928344727, "learning_rate": 9.045084971874737e-07, "loss": -0.0, "num_tokens": 3130519.0, "reward": 0.3492187261581421, "reward_std": 0.2351623922586441, "rewards/format_reward/mean": 0.5234375, "rewards/format_reward/std": 0.1875000149011612, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "step": 25 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 934.0, "completions/mean_length": 150.71875, "completions/mean_terminated_length": 120.60317993164062, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.065, "frac_reward_zero_std": 0.25, "grad_norm": 20.356454849243164, "learning_rate": 8.945702546981968e-07, "loss": -0.0, "num_tokens": 3227205.0, "reward": 0.31640625, "reward_std": 0.2614574134349823, "rewards/format_reward/mean": 0.6640625, "rewards/format_reward/std": 0.30930912494659424, "rewards/mcq_exact_match_reward/mean": 0.25, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "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": 889.0, "completions/max_terminated_length": 889.0, "completions/mean_length": 54.734375, "completions/mean_terminated_length": 54.734375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0675, "frac_reward_zero_std": 0.375, "grad_norm": 21.18809700012207, "learning_rate": 8.842005554284295e-07, "loss": 0.0, "num_tokens": 3323188.0, "reward": 0.4710937738418579, "reward_std": 0.261618435382843, "rewards/format_reward/mean": 0.6484375, "rewards/format_reward/std": 0.2302463799715042, "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": 476.0, "completions/max_terminated_length": 476.0, "completions/mean_length": 20.765625, "completions/mean_terminated_length": 20.765625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.07, "frac_reward_zero_std": 0.5, "grad_norm": 25.11161231994629, "learning_rate": 8.734107384920769e-07, "loss": -0.0, "num_tokens": 3451333.0, "reward": 0.39374998211860657, "reward_std": 0.22461532056331635, "rewards/format_reward/mean": 0.5, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.34375, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "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": 659.0, "completions/max_terminated_length": 659.0, "completions/mean_length": 29.296875, "completions/mean_terminated_length": 29.296875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0725, "frac_reward_zero_std": 0.5, "grad_norm": 22.686452865600586, "learning_rate": 8.622126023955445e-07, "loss": 0.0, "num_tokens": 3540248.0, "reward": 0.2671874761581421, "reward_std": 0.2404593825340271, "rewards/format_reward/mean": 0.640625, "rewards/format_reward/std": 0.22658175230026245, "rewards/mcq_exact_match_reward/mean": 0.203125, "rewards/mcq_exact_match_reward/std": 0.40550529956817627, "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": 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": 0.075, "frac_reward_zero_std": 0.5, "grad_norm": 26.69069480895996, "learning_rate": 8.506183921362442e-07, "loss": -0.0, "num_tokens": 3657816.0, "reward": 0.4562499523162842, "reward_std": 0.23827511072158813, "rewards/format_reward/mean": 0.5, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "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": 19.0, "completions/max_terminated_length": 19.0, "completions/mean_length": 7.671875, "completions/mean_terminated_length": 7.671875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0775, "frac_reward_zero_std": 0.625, "grad_norm": 14.952986717224121, "learning_rate": 8.386407858128706e-07, "loss": 0.0, "num_tokens": 3734843.0, "reward": 0.7179687023162842, "reward_std": 0.14677414298057556, "rewards/format_reward/mean": 0.6171875, "rewards/format_reward/std": 0.21347814798355103, "rewards/mcq_exact_match_reward/mean": 0.65625, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "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": 494.0, "completions/max_terminated_length": 494.0, "completions/mean_length": 38.546875, "completions/mean_terminated_length": 38.546875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08, "frac_reward_zero_std": 0.375, "grad_norm": 21.555471420288086, "learning_rate": 8.262928807620843e-07, "loss": -0.0, "num_tokens": 3842926.0, "reward": 0.4749999940395355, "reward_std": 0.20376864075660706, "rewards/format_reward/mean": 0.53125, "rewards/format_reward/std": 0.12198751419782639, "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": 530.0, "completions/max_terminated_length": 530.0, "completions/mean_length": 53.96875, "completions/mean_terminated_length": 53.96875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0825, "frac_reward_zero_std": 0.375, "grad_norm": 23.959991455078125, "learning_rate": 8.135881792367685e-07, "loss": -0.0, "num_tokens": 3921060.0, "reward": 0.4624999761581421, "reward_std": 0.19796372950077057, "rewards/format_reward/mean": 0.5625, "rewards/format_reward/std": 0.1666666716337204, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "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": 268.0, "completions/max_terminated_length": 268.0, "completions/mean_length": 14.71875, "completions/mean_terminated_length": 14.71875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.085, "frac_reward_zero_std": 0.625, "grad_norm": 16.529582977294922, "learning_rate": 8.005405736415125e-07, "loss": 0.0, "num_tokens": 4004298.0, "reward": 0.22499999403953552, "reward_std": 0.10678248107433319, "rewards/format_reward/mean": 0.53125, "rewards/format_reward/std": 0.1510545015335083, "rewards/mcq_exact_match_reward/mean": 0.171875, "rewards/mcq_exact_match_reward/std": 0.38025420904159546, "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": 49.0, "completions/max_terminated_length": 49.0, "completions/mean_length": 7.78125, "completions/mean_terminated_length": 7.78125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0875, "frac_reward_zero_std": 0.375, "grad_norm": 33.520294189453125, "learning_rate": 7.871643313414718e-07, "loss": 0.0, "num_tokens": 4117796.0, "reward": 0.32421875, "reward_std": 0.1381329894065857, "rewards/format_reward/mean": 0.5859375, "rewards/format_reward/std": 0.19012710452079773, "rewards/mcq_exact_match_reward/mean": 0.265625, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "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": 75.0, "completions/max_terminated_length": 75.0, "completions/mean_length": 8.8125, "completions/mean_terminated_length": 8.8125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.09, "frac_reward_zero_std": 0.75, "grad_norm": 14.681447982788086, "learning_rate": 7.734740790612136e-07, "loss": -0.0, "num_tokens": 4232312.0, "reward": 0.296875, "reward_std": 0.1315089464187622, "rewards/format_reward/mean": 0.625, "rewards/format_reward/std": 0.2182178944349289, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "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": 263.0, "completions/max_terminated_length": 263.0, "completions/mean_length": 16.765625, "completions/mean_terminated_length": 16.765625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0925, "frac_reward_zero_std": 0.75, "grad_norm": 14.76207447052002, "learning_rate": 7.594847868906076e-07, "loss": 0.0, "num_tokens": 4338633.0, "reward": 0.6812499761581421, "reward_std": 0.0883883386850357, "rewards/format_reward/mean": 0.5625, "rewards/format_reward/std": 0.1666666716337204, "rewards/mcq_exact_match_reward/mean": 0.625, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 80.0, "completions/max_terminated_length": 80.0, "completions/mean_length": 8.96875, "completions/mean_terminated_length": 8.96875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.095, "frac_reward_zero_std": 0.5, "grad_norm": 15.640183448791504, "learning_rate": 7.452117519152541e-07, "loss": 0.0, "num_tokens": 4433351.0, "reward": 0.84375, "reward_std": 0.22461533546447754, "rewards/format_reward/mean": 0.625, "rewards/format_reward/std": 0.2182178944349289, "rewards/mcq_exact_match_reward/mean": 0.78125, "rewards/mcq_exact_match_reward/std": 0.4166666865348816, "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": 1192.0, "completions/max_terminated_length": 1192.0, "completions/mean_length": 37.171875, "completions/mean_terminated_length": 37.171875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0975, "frac_reward_zero_std": 0.75, "grad_norm": 17.66131591796875, "learning_rate": 7.306705814893439e-07, "loss": -0.0, "num_tokens": 4529722.0, "reward": 0.6390624642372131, "reward_std": 0.10308553278446198, "rewards/format_reward/mean": 0.609375, "rewards/format_reward/std": 0.2083333432674408, "rewards/mcq_exact_match_reward/mean": 0.578125, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "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": 343.0, "completions/max_terminated_length": 343.0, "completions/mean_length": 14.6875, "completions/mean_terminated_length": 14.6875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1, "frac_reward_zero_std": 0.625, "grad_norm": 15.048425674438477, "learning_rate": 7.158771761692464e-07, "loss": 0.0, "num_tokens": 4635278.0, "reward": 0.4625000059604645, "reward_std": 0.16200581192970276, "rewards/format_reward/mean": 0.71875, "rewards/format_reward/std": 0.265398770570755, "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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 6.984375, "completions/mean_terminated_length": 6.984375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1025, "frac_reward_zero_std": 0.625, "grad_norm": 15.095502853393555, "learning_rate": 7.008477123264847e-07, "loss": 0.0, "num_tokens": 4728541.0, "reward": 0.7445312142372131, "reward_std": 0.11793722957372665, "rewards/format_reward/mean": 0.5703125, "rewards/format_reward/std": 0.1751912236213684, "rewards/mcq_exact_match_reward/mean": 0.6875, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "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": 60.0, "completions/max_terminated_length": 60.0, "completions/mean_length": 8.4375, "completions/mean_terminated_length": 8.4375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.105, "frac_reward_zero_std": 0.625, "grad_norm": 18.993345260620117, "learning_rate": 6.855986244591103e-07, "loss": 0.0, "num_tokens": 4813697.0, "reward": 0.328125, "reward_std": 0.17358145117759705, "rewards/format_reward/mean": 0.625, "rewards/format_reward/std": 0.2182178944349289, "rewards/mcq_exact_match_reward/mean": 0.265625, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "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": 866.0, "completions/max_terminated_length": 866.0, "completions/mean_length": 20.3125, "completions/mean_terminated_length": 20.3125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1075, "frac_reward_zero_std": 0.375, "grad_norm": 25.09459114074707, "learning_rate": 6.701465872208216e-07, "loss": -0.0, "num_tokens": 4917501.0, "reward": 0.5093749761581421, "reward_std": 0.2961486577987671, "rewards/format_reward/mean": 0.5625, "rewards/format_reward/std": 0.1666666716337204, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 43 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 10.28125, "completions/mean_terminated_length": 10.28125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.11, "frac_reward_zero_std": 0.625, "grad_norm": 16.505359649658203, "learning_rate": 6.545084971874736e-07, "loss": -0.0, "num_tokens": 4987663.0, "reward": 0.596875011920929, "reward_std": 0.16887323558330536, "rewards/format_reward/mean": 0.8125, "rewards/format_reward/std": 0.24397502839565277, "rewards/mcq_exact_match_reward/mean": 0.515625, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "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": 274.0, "completions/max_terminated_length": 274.0, "completions/mean_length": 14.15625, "completions/mean_terminated_length": 14.15625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1125, "frac_reward_zero_std": 0.75, "grad_norm": 11.136392593383789, "learning_rate": 6.387014543809223e-07, "loss": 0.0, "num_tokens": 5076561.0, "reward": 0.6726561784744263, "reward_std": 0.04640388861298561, "rewards/format_reward/mean": 0.6328125, "rewards/format_reward/std": 0.22257846593856812, "rewards/mcq_exact_match_reward/mean": 0.609375, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "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": 69.0, "completions/max_terminated_length": 69.0, "completions/mean_length": 7.84375, "completions/mean_terminated_length": 7.84375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.115, "frac_reward_zero_std": 0.625, "grad_norm": 18.416553497314453, "learning_rate": 6.227427435703995e-07, "loss": 0.0, "num_tokens": 5176647.0, "reward": 0.4468749761581421, "reward_std": 0.13258251547813416, "rewards/format_reward/mean": 0.5625, "rewards/format_reward/std": 0.1666666716337204, "rewards/mcq_exact_match_reward/mean": 0.390625, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 7.46875, "completions/mean_terminated_length": 7.46875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1175, "frac_reward_zero_std": 0.375, "grad_norm": 45.77751541137695, "learning_rate": 6.066498153718734e-07, "loss": 0.0, "num_tokens": 5268189.0, "reward": 0.46718746423721313, "reward_std": 0.15169650316238403, "rewards/format_reward/mean": 0.609375, "rewards/format_reward/std": 0.2083333432674408, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 327.0, "completions/mean_length": 44.640625, "completions/mean_terminated_length": 12.841270446777344, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.12, "frac_reward_zero_std": 0.5, "grad_norm": 23.31832504272461, "learning_rate": 5.90440267166055e-07, "loss": 0.0, "num_tokens": 5375718.0, "reward": 0.51171875, "reward_std": 0.25713759660720825, "rewards/format_reward/mean": 0.5859375, "rewards/format_reward/std": 0.209963858127594, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "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": 197.0, "completions/max_terminated_length": 197.0, "completions/mean_length": 10.390625, "completions/mean_terminated_length": 10.390625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1225, "frac_reward_zero_std": 0.625, "grad_norm": 13.738724708557129, "learning_rate": 5.741318238559209e-07, "loss": 0.0, "num_tokens": 5495031.0, "reward": 0.2789062559604645, "reward_std": 0.10854631662368774, "rewards/format_reward/mean": 0.6015625, "rewards/format_reward/std": 0.20275264978408813, "rewards/mcq_exact_match_reward/mean": 0.21875, "rewards/mcq_exact_match_reward/std": 0.4166666865348816, "step": 49 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 6.921875, "completions/mean_terminated_length": 6.921875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.125, "frac_reward_zero_std": 0.5, "grad_norm": 19.474903106689453, "learning_rate": 5.577423184847931e-07, "loss": -0.0, "num_tokens": 5621498.0, "reward": 0.4945312440395355, "reward_std": 0.23424796760082245, "rewards/format_reward/mean": 0.5703125, "rewards/format_reward/std": 0.1751912236213684, "rewards/mcq_exact_match_reward/mean": 0.4375, "rewards/mcq_exact_match_reward/std": 0.5, "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": 377.0, "completions/max_terminated_length": 377.0, "completions/mean_length": 28.609375, "completions/mean_terminated_length": 28.609375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1275, "frac_reward_zero_std": 0.125, "grad_norm": 21.532567977905273, "learning_rate": 5.412896727361662e-07, "loss": 0.0, "num_tokens": 5719329.0, "reward": 0.63671875, "reward_std": 0.2972930073738098, "rewards/format_reward/mean": 0.7421875, "rewards/format_reward/std": 0.26714521646499634, "rewards/mcq_exact_match_reward/mean": 0.5625, "rewards/mcq_exact_match_reward/std": 0.5, "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": 249.0, "completions/max_terminated_length": 249.0, "completions/mean_length": 12.265625, "completions/mean_terminated_length": 12.265625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.13, "frac_reward_zero_std": 0.375, "grad_norm": 19.431655883789062, "learning_rate": 5.247918773366111e-07, "loss": -0.0, "num_tokens": 5849002.0, "reward": 0.6773437261581421, "reward_std": 0.24944616854190826, "rewards/format_reward/mean": 0.6796875, "rewards/format_reward/std": 0.24180518090724945, "rewards/mcq_exact_match_reward/mean": 0.609375, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "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": 469.0, "completions/max_terminated_length": 469.0, "completions/mean_length": 29.8125, "completions/mean_terminated_length": 29.8125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1325, "frac_reward_zero_std": 0.25, "grad_norm": 60.234066009521484, "learning_rate": 5.082669723831793e-07, "loss": 0.0, "num_tokens": 5971662.0, "reward": 0.46796876192092896, "reward_std": 0.1877097487449646, "rewards/format_reward/mean": 0.7734375, "rewards/format_reward/std": 0.250866562128067, "rewards/mcq_exact_match_reward/mean": 0.390625, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 611.0, "completions/mean_length": 93.984375, "completions/mean_terminated_length": 62.968257904052734, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.135, "frac_reward_zero_std": 0.375, "grad_norm": 7.178009510040283, "learning_rate": 4.917330276168208e-07, "loss": 0.0, "num_tokens": 6052661.0, "reward": 0.42578125, "reward_std": 0.20135706663131714, "rewards/format_reward/mean": 0.8203125, "rewards/format_reward/std": 0.27265870571136475, "rewards/mcq_exact_match_reward/mean": 0.34375, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "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": 0.0, "completions/max_length": 319.0, "completions/max_terminated_length": 319.0, "completions/mean_length": 26.71875, "completions/mean_terminated_length": 26.71875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1375, "frac_reward_zero_std": 0.5, "grad_norm": 41.9691162109375, "learning_rate": 4.752081226633888e-07, "loss": 0.0, "num_tokens": 6178699.0, "reward": 0.58984375, "reward_std": 0.18487374484539032, "rewards/format_reward/mean": 0.7421875, "rewards/format_reward/std": 0.26714521646499634, "rewards/mcq_exact_match_reward/mean": 0.515625, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1056.0, "completions/mean_length": 129.1875, "completions/mean_terminated_length": 67.29032135009766, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.14, "frac_reward_zero_std": 0.25, "grad_norm": 30.365982055664062, "learning_rate": 4.5871032726383385e-07, "loss": 0.0, "num_tokens": 6291911.0, "reward": 0.54296875, "reward_std": 0.23562288284301758, "rewards/format_reward/mean": 0.7421875, "rewards/format_reward/std": 0.2816080152988434, "rewards/mcq_exact_match_reward/mean": 0.46875, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "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": 457.0, "completions/max_terminated_length": 457.0, "completions/mean_length": 31.0, "completions/mean_terminated_length": 31.0, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1425, "frac_reward_zero_std": 0.375, "grad_norm": 26.010089874267578, "learning_rate": 4.4225768151520694e-07, "loss": 0.0, "num_tokens": 6399375.0, "reward": 0.6148437857627869, "reward_std": 0.2862638831138611, "rewards/format_reward/mean": 0.8359375, "rewards/format_reward/std": 0.25283610820770264, "rewards/mcq_exact_match_reward/mean": 0.53125, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 235.0, "completions/mean_length": 48.671875, "completions/mean_terminated_length": 16.936508178710938, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.145, "frac_reward_zero_std": 0.5, "grad_norm": 18.15598487854004, "learning_rate": 4.258681761440789e-07, "loss": 0.0, "num_tokens": 6500122.0, "reward": 0.5476562976837158, "reward_std": 0.15600594878196716, "rewards/format_reward/mean": 0.7890625, "rewards/format_reward/std": 0.2789533734321594, "rewards/mcq_exact_match_reward/mean": 0.46875, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "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": 670.0, "completions/max_terminated_length": 670.0, "completions/mean_length": 28.40625, "completions/mean_terminated_length": 28.40625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1475, "frac_reward_zero_std": 0.25, "grad_norm": 33.12482452392578, "learning_rate": 4.095597328339452e-07, "loss": 0.0, "num_tokens": 6612548.0, "reward": 0.2984375059604645, "reward_std": 0.176957368850708, "rewards/format_reward/mean": 0.796875, "rewards/format_reward/std": 0.2630521357059479, "rewards/mcq_exact_match_reward/mean": 0.21875, "rewards/mcq_exact_match_reward/std": 0.4166666865348816, "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": 59.078125, "completions/mean_terminated_length": 59.078125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.15, "frac_reward_zero_std": 0.5, "grad_norm": 13.72029972076416, "learning_rate": 3.9335018462812664e-07, "loss": -0.0, "num_tokens": 6700497.0, "reward": 0.7210937738418579, "reward_std": 0.12268893420696259, "rewards/format_reward/mean": 0.9609375, "rewards/format_reward/std": 0.16194961965084076, "rewards/mcq_exact_match_reward/mean": 0.625, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 916.0, "completions/max_terminated_length": 916.0, "completions/mean_length": 26.921875, "completions/mean_terminated_length": 26.921875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1525, "frac_reward_zero_std": 0.25, "grad_norm": 30.64706039428711, "learning_rate": 3.772572564296004e-07, "loss": -0.0, "num_tokens": 6804804.0, "reward": 0.6664062738418579, "reward_std": 0.21409152448177338, "rewards/format_reward/mean": 0.8828125, "rewards/format_reward/std": 0.2478829026222229, "rewards/mcq_exact_match_reward/mean": 0.578125, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "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": 294.0, "completions/max_terminated_length": 294.0, "completions/mean_length": 43.203125, "completions/mean_terminated_length": 43.203125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.155, "frac_reward_zero_std": 0.625, "grad_norm": 13.85999584197998, "learning_rate": 3.612985456190778e-07, "loss": -0.0, "num_tokens": 6897521.0, "reward": 0.6109375357627869, "reward_std": 0.13473878800868988, "rewards/format_reward/mean": 0.953125, "rewards/format_reward/std": 0.1717960685491562, "rewards/mcq_exact_match_reward/mean": 0.515625, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 546.0, "completions/mean_length": 58.984375, "completions/mean_terminated_length": 27.412700653076172, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1575, "frac_reward_zero_std": 0.375, "grad_norm": 11.780682563781738, "learning_rate": 3.454915028125263e-07, "loss": 0.0, "num_tokens": 7029224.0, "reward": 0.79296875, "reward_std": 0.26555266976356506, "rewards/format_reward/mean": 0.8984375, "rewards/format_reward/std": 0.22146137058734894, "rewards/mcq_exact_match_reward/mean": 0.703125, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "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": 236.0, "completions/max_terminated_length": 236.0, "completions/mean_length": 18.65625, "completions/mean_terminated_length": 18.65625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.16, "frac_reward_zero_std": 0.625, "grad_norm": 10.003776550292969, "learning_rate": 3.2985341277917846e-07, "loss": 0.0, "num_tokens": 7110578.0, "reward": 0.4906250238418579, "reward_std": 0.13258251547813416, "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": 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": 225.0, "completions/max_terminated_length": 225.0, "completions/mean_length": 15.703125, "completions/mean_terminated_length": 15.703125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1625, "frac_reward_zero_std": 0.875, "grad_norm": 15.267725944519043, "learning_rate": 3.1440137554088953e-07, "loss": -0.0, "num_tokens": 7199823.0, "reward": 0.3851562738418579, "reward_std": 0.06580014526844025, "rewards/format_reward/mean": 0.8828125, "rewards/format_reward/std": 0.21347814798355103, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "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": 622.0, "completions/max_terminated_length": 622.0, "completions/mean_length": 47.515625, "completions/mean_terminated_length": 47.515625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.165, "frac_reward_zero_std": 0.375, "grad_norm": 25.18468475341797, "learning_rate": 2.9915228767351535e-07, "loss": 0.0, "num_tokens": 7291152.0, "reward": 0.7789062261581421, "reward_std": 0.1763656735420227, "rewards/format_reward/mean": 0.9140625, "rewards/format_reward/std": 0.2280818521976471, "rewards/mcq_exact_match_reward/mean": 0.6875, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "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": 387.0, "completions/max_terminated_length": 387.0, "completions/mean_length": 29.0625, "completions/mean_terminated_length": 29.0625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1675, "frac_reward_zero_std": 0.5, "grad_norm": 19.555036544799805, "learning_rate": 2.841228238307536e-07, "loss": -0.0, "num_tokens": 7398340.0, "reward": 0.6765625476837158, "reward_std": 0.10647734999656677, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.08768405020236969, "rewards/mcq_exact_match_reward/mean": 0.578125, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "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": 284.0, "completions/max_terminated_length": 284.0, "completions/mean_length": 27.21875, "completions/mean_terminated_length": 27.21875, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "epoch": 0.17, "frac_reward_zero_std": 0.625, "grad_norm": 27.19438362121582, "learning_rate": 2.6932941851065615e-07, "loss": 0.0, "num_tokens": 7482682.0, "reward": 0.59765625, "reward_std": 0.09254656732082367, "rewards/format_reward/mean": 0.9765625, "rewards/format_reward/std": 0.13886408507823944, "rewards/mcq_exact_match_reward/mean": 0.5, "rewards/mcq_exact_match_reward/std": 0.5039526224136353, "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": 347.0, "completions/max_terminated_length": 347.0, "completions/mean_length": 23.203125, "completions/mean_terminated_length": 23.203125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1725, "frac_reward_zero_std": 0.75, "grad_norm": 12.474835395812988, "learning_rate": 2.547882480847461e-07, "loss": 0.0, "num_tokens": 7557887.0, "reward": 0.4750000238418579, "reward_std": 0.0883883461356163, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.375, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "step": 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": 368.0, "completions/max_terminated_length": 368.0, "completions/mean_length": 18.09375, "completions/mean_terminated_length": 18.09375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.175, "frac_reward_zero_std": 0.625, "grad_norm": 17.514318466186523, "learning_rate": 2.4051521310939254e-07, "loss": -0.0, "num_tokens": 7689621.0, "reward": 0.4429687559604645, "reward_std": 0.11109752953052521, "rewards/format_reward/mean": 0.9921875, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.34375, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "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": 297.0, "completions/max_terminated_length": 297.0, "completions/mean_length": 19.640625, "completions/mean_terminated_length": 19.640625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1775, "frac_reward_zero_std": 0.75, "grad_norm": 23.81156349182129, "learning_rate": 2.2652592093878665e-07, "loss": 0.0, "num_tokens": 7776670.0, "reward": 0.42656251788139343, "reward_std": 0.12398644536733627, "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": 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": 745.0, "completions/max_terminated_length": 745.0, "completions/mean_length": 56.96875, "completions/mean_terminated_length": 56.96875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.18, "frac_reward_zero_std": 0.375, "grad_norm": 35.66807556152344, "learning_rate": 2.128356686585282e-07, "loss": 0.0, "num_tokens": 7865628.0, "reward": 0.7078125476837158, "reward_std": 0.2680327594280243, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.08768405020236969, "rewards/mcq_exact_match_reward/mean": 0.609375, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 397.0, "completions/mean_length": 50.0625, "completions/mean_terminated_length": 18.349206924438477, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1825, "frac_reward_zero_std": 0.75, "grad_norm": 9.515981674194336, "learning_rate": 1.9945942635848745e-07, "loss": -0.0, "num_tokens": 7975904.0, "reward": 0.37968751788139343, "reward_std": 0.08923260122537613, "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": 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": 26.0, "completions/max_terminated_length": 26.0, "completions/mean_length": 12.8125, "completions/mean_terminated_length": 12.8125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.185, "frac_reward_zero_std": 0.75, "grad_norm": 9.167645454406738, "learning_rate": 1.8641182076323148e-07, "loss": 0.0, "num_tokens": 8071300.0, "reward": 0.6617187857627869, "reward_std": 0.11624954640865326, "rewards/format_reward/mean": 0.9921875, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.5625, "rewards/mcq_exact_match_reward/std": 0.5, "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": 589.0, "completions/max_terminated_length": 589.0, "completions/mean_length": 60.859375, "completions/mean_terminated_length": 60.859375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1875, "frac_reward_zero_std": 0.625, "grad_norm": 12.873812675476074, "learning_rate": 1.7370711923791564e-07, "loss": -0.0, "num_tokens": 8161003.0, "reward": 0.5015625357627869, "reward_std": 0.182172492146492, "rewards/format_reward/mean": 0.953125, "rewards/format_reward/std": 0.21304203569889069, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "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": 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": 0.19, "frac_reward_zero_std": 0.625, "grad_norm": 11.817203521728516, "learning_rate": 1.6135921418712955e-07, "loss": -0.0, "num_tokens": 8252747.0, "reward": 0.5843750238418579, "reward_std": 0.17358146607875824, "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": 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": 21.0, "completions/max_terminated_length": 21.0, "completions/mean_length": 12.765625, "completions/mean_terminated_length": 12.765625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1925, "frac_reward_zero_std": 0.625, "grad_norm": 18.291208267211914, "learning_rate": 1.493816078637557e-07, "loss": 0.0, "num_tokens": 8353548.0, "reward": 0.5015625357627869, "reward_std": 0.19132769107818604, "rewards/format_reward/mean": 0.953125, "rewards/format_reward/std": 0.21304203569889069, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "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": 20.0, "completions/max_terminated_length": 20.0, "completions/mean_length": 12.765625, "completions/mean_terminated_length": 12.765625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.195, "frac_reward_zero_std": 0.625, "grad_norm": 15.619884490966797, "learning_rate": 1.3778739760445552e-07, "loss": 0.0, "num_tokens": 8439149.0, "reward": 0.6000000238418579, "reward_std": 0.16675157845020294, "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": 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": 12.359375, "completions/mean_terminated_length": 12.359375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1975, "frac_reward_zero_std": 0.75, "grad_norm": 14.782150268554688, "learning_rate": 1.2658926150792322e-07, "loss": 0.0, "num_tokens": 8534740.0, "reward": 0.5515625476837158, "reward_std": 0.11185373365879059, "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": 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": 201.0, "completions/max_terminated_length": 201.0, "completions/mean_length": 19.265625, "completions/mean_terminated_length": 19.265625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2, "frac_reward_zero_std": 0.5, "grad_norm": 14.222006797790527, "learning_rate": 1.1579944457157059e-07, "loss": 0.0, "num_tokens": 8643861.0, "reward": 0.7093750238418579, "reward_std": 0.2109457552433014, "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": 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": 591.0, "completions/max_terminated_length": 591.0, "completions/mean_length": 37.140625, "completions/mean_terminated_length": 37.140625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2025, "frac_reward_zero_std": 0.75, "grad_norm": 7.147845268249512, "learning_rate": 1.0542974530180327e-07, "loss": 0.0, "num_tokens": 8745982.0, "reward": 0.6453125476837158, "reward_std": 0.11280198395252228, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.546875, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "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": 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": 0.205, "frac_reward_zero_std": 0.875, "grad_norm": 9.346044540405273, "learning_rate": 9.549150281252632e-08, "loss": 0.0, "num_tokens": 8850006.0, "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": 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": 274.0, "completions/max_terminated_length": 274.0, "completions/mean_length": 20.203125, "completions/mean_terminated_length": 20.203125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2075, "frac_reward_zero_std": 0.75, "grad_norm": 7.28822135925293, "learning_rate": 8.599558442598998e-08, "loss": -0.0, "num_tokens": 8934139.0, "reward": 0.5054687857627869, "reward_std": 0.06007346510887146, "rewards/format_reward/mean": 0.9921875, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "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": 115.0, "completions/max_terminated_length": 115.0, "completions/mean_length": 15.015625, "completions/mean_terminated_length": 15.015625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.21, "frac_reward_zero_std": 0.875, "grad_norm": 4.744446754455566, "learning_rate": 7.695237378953224e-08, "loss": 0.0, "num_tokens": 9034204.0, "reward": 0.44218751788139343, "reward_std": 0.060848724097013474, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.34375, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "step": 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": 16.0, "completions/max_terminated_length": 16.0, "completions/mean_length": 11.53125, "completions/mean_terminated_length": 11.53125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.2125, "frac_reward_zero_std": 0.5, "grad_norm": 24.570159912109375, "learning_rate": 6.837175952121304e-08, "loss": 0.0, "num_tokens": 9156518.0, "reward": 0.7203125357627869, "reward_std": 0.14914530515670776, "rewards/format_reward/mean": 0.953125, "rewards/format_reward/std": 0.14689241349697113, "rewards/mcq_exact_match_reward/mean": 0.625, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.03125, "completions/mean_terminated_length": 12.03125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.215, "frac_reward_zero_std": 0.875, "grad_norm": 12.611838340759277, "learning_rate": 6.026312439675551e-08, "loss": 0.0, "num_tokens": 9270768.0, "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": 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": 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": 0.2175, "frac_reward_zero_std": 0.75, "grad_norm": 10.183032989501953, "learning_rate": 5.263533508961826e-08, "loss": 0.0, "num_tokens": 9370269.0, "reward": 0.3968750238418579, "reward_std": 0.12255740165710449, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "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": 374.0, "completions/max_terminated_length": 374.0, "completions/mean_length": 17.78125, "completions/mean_terminated_length": 17.78125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.22, "frac_reward_zero_std": 0.625, "grad_norm": 14.991341590881348, "learning_rate": 4.549673247541874e-08, "loss": 0.0, "num_tokens": 9461007.0, "reward": 0.706250011920929, "reward_std": 0.18132445216178894, "rewards/format_reward/mean": 0.96875, "rewards/format_reward/std": 0.17536810040473938, "rewards/mcq_exact_match_reward/mean": 0.609375, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "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": 871.0, "completions/max_terminated_length": 871.0, "completions/mean_length": 28.859375, "completions/mean_terminated_length": 28.859375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2225, "frac_reward_zero_std": 0.875, "grad_norm": 10.161040306091309, "learning_rate": 3.8855122511307626e-08, "loss": 0.0, "num_tokens": 9550606.0, "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": 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": 108.0, "completions/max_terminated_length": 108.0, "completions/mean_length": 13.53125, "completions/mean_terminated_length": 13.53125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.225, "frac_reward_zero_std": 0.75, "grad_norm": 14.991464614868164, "learning_rate": 3.271776770026963e-08, "loss": -0.0, "num_tokens": 9655536.0, "reward": 0.9437500238418579, "reward_std": 0.0883883461356163, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.84375, "rewards/mcq_exact_match_reward/std": 0.36596253514289856, "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": 133.0, "completions/max_terminated_length": 133.0, "completions/mean_length": 15.265625, "completions/mean_terminated_length": 15.265625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2275, "frac_reward_zero_std": 0.875, "grad_norm": 1.2314852476119995, "learning_rate": 2.7091379149682682e-08, "loss": -0.0, "num_tokens": 9756273.0, "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": 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": 16.0, "completions/max_terminated_length": 16.0, "completions/mean_length": 12.109375, "completions/mean_terminated_length": 12.109375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.23, "frac_reward_zero_std": 0.875, "grad_norm": 12.691123962402344, "learning_rate": 2.1982109232821176e-08, "loss": 0.0, "num_tokens": 9838968.0, "reward": 0.47343751788139343, "reward_std": 0.0044194171205163, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.375, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 289.0, "completions/max_terminated_length": 289.0, "completions/mean_length": 20.671875, "completions/mean_terminated_length": 20.671875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2325, "frac_reward_zero_std": 0.875, "grad_norm": 5.805517673492432, "learning_rate": 1.7395544861325718e-08, "loss": -0.0, "num_tokens": 9931579.0, "reward": 0.8343750238418579, "reward_std": 0.04419417306780815, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.734375, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "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": 108.0, "completions/max_terminated_length": 108.0, "completions/mean_length": 13.515625, "completions/mean_terminated_length": 13.515625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.235, "frac_reward_zero_std": 0.625, "grad_norm": 15.913883209228516, "learning_rate": 1.3336701375997127e-08, "loss": 0.0, "num_tokens": 10042036.0, "reward": 0.8187500238418579, "reward_std": 0.1462521106004715, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.71875, "rewards/mcq_exact_match_reward/std": 0.4531635046005249, "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": 162.0, "completions/max_terminated_length": 162.0, "completions/mean_length": 15.90625, "completions/mean_terminated_length": 15.90625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2375, "frac_reward_zero_std": 0.625, "grad_norm": 16.905920028686523, "learning_rate": 9.810017062595321e-09, "loss": 0.0, "num_tokens": 10144830.0, "reward": 0.73046875, "reward_std": 0.11005401611328125, "rewards/format_reward/mean": 0.8984375, "rewards/format_reward/std": 0.2978719174861908, "rewards/mcq_exact_match_reward/mean": 0.640625, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "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": 625.0, "completions/max_terminated_length": 625.0, "completions/mean_length": 21.578125, "completions/mean_terminated_length": 21.578125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.24, "frac_reward_zero_std": 0.75, "grad_norm": 6.906428337097168, "learning_rate": 6.819348298638839e-09, "loss": -0.0, "num_tokens": 10268723.0, "reward": 0.6781250238418579, "reward_std": 0.11100947856903076, "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": 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": 894.0, "completions/max_terminated_length": 894.0, "completions/mean_length": 26.421875, "completions/mean_terminated_length": 26.421875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2425, "frac_reward_zero_std": 0.75, "grad_norm": 8.926799774169922, "learning_rate": 4.367965336512403e-09, "loss": 0.0, "num_tokens": 10355998.0, "reward": 0.47343751788139343, "reward_std": 0.09280776977539062, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.375, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 60.0, "completions/max_terminated_length": 60.0, "completions/mean_length": 12.765625, "completions/mean_terminated_length": 12.765625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.245, "frac_reward_zero_std": 0.875, "grad_norm": 5.698251724243164, "learning_rate": 2.458548727494292e-09, "loss": 0.0, "num_tokens": 10445535.0, "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": 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": 18.0, "completions/max_terminated_length": 18.0, "completions/mean_length": 11.578125, "completions/mean_terminated_length": 11.578125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.2475, "frac_reward_zero_std": 0.625, "grad_norm": 34.691612243652344, "learning_rate": 1.0931863906127325e-09, "loss": -0.0, "num_tokens": 10542772.0, "reward": 0.6882812976837158, "reward_std": 0.09102033078670502, "rewards/format_reward/mean": 0.9453125, "rewards/format_reward/std": 0.15728822350502014, "rewards/mcq_exact_match_reward/mean": 0.59375, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "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": 160.0, "completions/max_terminated_length": 160.0, "completions/mean_length": 16.46875, "completions/mean_terminated_length": 16.46875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.25, "frac_reward_zero_std": 0.875, "grad_norm": 4.237037658691406, "learning_rate": 2.733713295369755e-10, "loss": 0.0, "num_tokens": 10624106.0, "reward": 0.3343750238418579, "reward_std": 0.04419417306780815, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "step": 100 } ], "logging_steps": 1, "max_steps": 100, "num_input_tokens_seen": 10624106, "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 }