{ "best_global_step": null, "best_metric": null, "best_model_checkpoint": null, "epoch": 0.22321428571428573, "eval_steps": 500, "global_step": 100, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1821.0, "completions/mean_length": 478.25, "completions/mean_terminated_length": 453.3333740234375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.002232142857142857, "frac_reward_zero_std": 0.125, "grad_norm": 6.604314804077148, "learning_rate": 0.0, "loss": -0.0, "num_tokens": 131688.0, "reward": 0.2242187261581421, "reward_std": 0.33872127532958984, "rewards/format_reward/mean": 0.3671875, "rewards/format_reward/std": 0.255761981010437, "rewards/mcq_exact_match_reward/mean": 0.1875, "rewards/mcq_exact_match_reward/std": 0.39339789748191833, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1499.0, "completions/mean_length": 477.34375, "completions/mean_terminated_length": 452.4127197265625, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "epoch": 0.004464285714285714, "frac_reward_zero_std": 0.0, "grad_norm": 7.3943586349487305, "learning_rate": 2e-07, "loss": -0.0, "num_tokens": 243622.0, "reward": 0.31484371423721313, "reward_std": 0.41326600313186646, "rewards/format_reward/mean": 0.3359375, "rewards/format_reward/std": 0.2680720090866089, "rewards/mcq_exact_match_reward/mean": 0.28125, "rewards/mcq_exact_match_reward/std": 0.4531635046005249, "step": 2 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1671.0, "completions/max_terminated_length": 1671.0, "completions/mean_length": 349.5625, "completions/mean_terminated_length": 349.5625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.006696428571428571, "frac_reward_zero_std": 0.0, "grad_norm": 19.394031524658203, "learning_rate": 4e-07, "loss": -0.0, "num_tokens": 335706.0, "reward": 0.5656249523162842, "reward_std": 0.47859323024749756, "rewards/format_reward/mean": 0.34375, "rewards/format_reward/std": 0.279951810836792, "rewards/mcq_exact_match_reward/mean": 0.53125, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "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": 0.0625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1757.0, "completions/mean_length": 743.375, "completions/mean_terminated_length": 656.4000244140625, "completions/min_length": 184.0, "completions/min_terminated_length": 184.0, "epoch": 0.008928571428571428, "frac_reward_zero_std": 0.0, "grad_norm": 1.0879732370376587, "learning_rate": 6e-07, "loss": 0.0, "num_tokens": 486274.0, "reward": 0.18906250596046448, "reward_std": 0.24741396307945251, "rewards/format_reward/mean": 0.328125, "rewards/format_reward/std": 0.25539806485176086, "rewards/mcq_exact_match_reward/mean": 0.15625, "rewards/mcq_exact_match_reward/std": 0.36596253514289856, "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": 1615.0, "completions/max_terminated_length": 1615.0, "completions/mean_length": 463.59375, "completions/mean_terminated_length": 463.59375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.011160714285714286, "frac_reward_zero_std": 0.0, "grad_norm": 17.959299087524414, "learning_rate": 8e-07, "loss": 0.0, "num_tokens": 616736.0, "reward": 0.3179687261581421, "reward_std": 0.2955024838447571, "rewards/format_reward/mean": 0.3671875, "rewards/format_reward/std": 0.23974503576755524, "rewards/mcq_exact_match_reward/mean": 0.28125, "rewards/mcq_exact_match_reward/std": 0.4531635046005249, "step": 5 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1609.0, "completions/mean_length": 552.015625, "completions/mean_terminated_length": 528.2698974609375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.013392857142857142, "frac_reward_zero_std": 0.0, "grad_norm": 37.449737548828125, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 742465.0, "reward": 0.2593749761581421, "reward_std": 0.2996486723423004, "rewards/format_reward/mean": 0.40625, "rewards/format_reward/std": 0.29378482699394226, "rewards/mcq_exact_match_reward/mean": 0.21875, "rewards/mcq_exact_match_reward/std": 0.4166666865348816, "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": 0.046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 519.703125, "completions/mean_terminated_length": 444.54095458984375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.015625, "frac_reward_zero_std": 0.125, "grad_norm": 10.06786060333252, "learning_rate": 9.99726628670463e-07, "loss": 0.0, "num_tokens": 863350.0, "reward": 0.2734375, "reward_std": 0.3021301031112671, "rewards/format_reward/mean": 0.390625, "rewards/format_reward/std": 0.24346621334552765, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "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": 0.0, "completions/max_length": 1562.0, "completions/max_terminated_length": 1562.0, "completions/mean_length": 526.390625, "completions/mean_terminated_length": 526.390625, "completions/min_length": 56.0, "completions/min_terminated_length": 56.0, "epoch": 0.017857142857142856, "frac_reward_zero_std": 0.0, "grad_norm": 1.6308486461639404, "learning_rate": 9.989068136093872e-07, "loss": -0.0, "num_tokens": 967087.0, "reward": 0.4406249523162842, "reward_std": 0.424815833568573, "rewards/format_reward/mean": 0.34375, "rewards/format_reward/std": 0.3069951832294464, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "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": 1013.0, "completions/max_terminated_length": 1013.0, "completions/mean_length": 357.703125, "completions/mean_terminated_length": 357.703125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.020089285714285716, "frac_reward_zero_std": 0.0, "grad_norm": 16.63702964782715, "learning_rate": 9.975414512725056e-07, "loss": -0.0, "num_tokens": 1093596.0, "reward": 0.25703126192092896, "reward_std": 0.30407845973968506, "rewards/format_reward/mean": 0.3828125, "rewards/format_reward/std": 0.2919853627681732, "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": 1768.0, "completions/max_terminated_length": 1768.0, "completions/mean_length": 497.328125, "completions/mean_terminated_length": 497.328125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.022321428571428572, "frac_reward_zero_std": 0.0, "grad_norm": 6.847559928894043, "learning_rate": 9.956320346634875e-07, "loss": 0.0, "num_tokens": 1194129.0, "reward": 0.22968748211860657, "reward_std": 0.28835001587867737, "rewards/format_reward/mean": 0.421875, "rewards/format_reward/std": 0.2221602201461792, "rewards/mcq_exact_match_reward/mean": 0.1875, "rewards/mcq_exact_match_reward/std": 0.39339789748191833, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1022.0, "completions/mean_length": 497.40625, "completions/mean_terminated_length": 447.3870849609375, "completions/min_length": 10.0, "completions/min_terminated_length": 10.0, "epoch": 0.024553571428571428, "frac_reward_zero_std": 0.25, "grad_norm": 1.0336768627166748, "learning_rate": 9.931806517013612e-07, "loss": -0.0, "num_tokens": 1314011.0, "reward": 0.20703125, "reward_std": 0.2851349711418152, "rewards/format_reward/mean": 0.3515625, "rewards/format_reward/std": 0.24688033759593964, "rewards/mcq_exact_match_reward/mean": 0.171875, "rewards/mcq_exact_match_reward/std": 0.38025420904159546, "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": 1426.0, "completions/mean_length": 583.375, "completions/mean_terminated_length": 560.1270141601562, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.026785714285714284, "frac_reward_zero_std": 0.0, "grad_norm": 2.6043832302093506, "learning_rate": 9.901899829374047e-07, "loss": -0.0, "num_tokens": 1440035.0, "reward": 0.35859373211860657, "reward_std": 0.36114513874053955, "rewards/format_reward/mean": 0.3046875, "rewards/format_reward/std": 0.2615155577659607, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "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": 1847.0, "completions/max_terminated_length": 1847.0, "completions/mean_length": 547.359375, "completions/mean_terminated_length": 547.359375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.029017857142857144, "frac_reward_zero_std": 0.0, "grad_norm": 4.645428657531738, "learning_rate": 9.866632986240029e-07, "loss": 0.0, "num_tokens": 1578162.0, "reward": 0.25078126788139343, "reward_std": 0.319469690322876, "rewards/format_reward/mean": 0.4765625, "rewards/format_reward/std": 0.3011839985847473, "rewards/mcq_exact_match_reward/mean": 0.203125, "rewards/mcq_exact_match_reward/std": 0.40550529956817627, "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": 1896.0, "completions/max_terminated_length": 1896.0, "completions/mean_length": 576.359375, "completions/mean_terminated_length": 576.359375, "completions/min_length": 133.0, "completions/min_terminated_length": 133.0, "epoch": 0.03125, "frac_reward_zero_std": 0.25, "grad_norm": 0.9529601335525513, "learning_rate": 9.826044551386742e-07, "loss": 0.0, "num_tokens": 1677417.0, "reward": 0.29296875, "reward_std": 0.18695229291915894, "rewards/format_reward/mean": 0.4296875, "rewards/format_reward/std": 0.2498759627342224, "rewards/mcq_exact_match_reward/mean": 0.25, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "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": 0.046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1744.0, "completions/mean_length": 629.296875, "completions/mean_terminated_length": 559.5245361328125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.033482142857142856, "frac_reward_zero_std": 0.0, "grad_norm": 3.4656269550323486, "learning_rate": 9.780178907671788e-07, "loss": 0.0, "num_tokens": 1806836.0, "reward": 0.32265621423721313, "reward_std": 0.36032652854919434, "rewards/format_reward/mean": 0.4140625, "rewards/format_reward/std": 0.2605654299259186, "rewards/mcq_exact_match_reward/mean": 0.28125, "rewards/mcq_exact_match_reward/std": 0.4531635046005249, "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": 1514.0, "completions/mean_length": 500.453125, "completions/mean_terminated_length": 475.888916015625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.03571428571428571, "frac_reward_zero_std": 0.0, "grad_norm": 6.499881267547607, "learning_rate": 9.729086208503173e-07, "loss": -0.0, "num_tokens": 1946145.0, "reward": 0.19453124701976776, "reward_std": 0.3073563873767853, "rewards/format_reward/mean": 0.5390625, "rewards/format_reward/std": 0.2567298710346222, "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": 1560.0, "completions/max_terminated_length": 1560.0, "completions/mean_length": 588.59375, "completions/mean_terminated_length": 588.59375, "completions/min_length": 11.0, "completions/min_terminated_length": 11.0, "epoch": 0.03794642857142857, "frac_reward_zero_std": 0.0, "grad_norm": 2.1308276653289795, "learning_rate": 9.672822322997304e-07, "loss": -0.0, "num_tokens": 2072791.0, "reward": 0.3390624523162842, "reward_std": 0.3752942681312561, "rewards/format_reward/mean": 0.421875, "rewards/format_reward/std": 0.18298126757144928, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1790.0, "completions/mean_length": 588.03125, "completions/mean_terminated_length": 540.9354858398438, "completions/min_length": 20.0, "completions/min_terminated_length": 20.0, "epoch": 0.04017857142857143, "frac_reward_zero_std": 0.0, "grad_norm": 4.073460102081299, "learning_rate": 9.611448774886923e-07, "loss": -0.0, "num_tokens": 2205801.0, "reward": 0.2890625, "reward_std": 0.40764081478118896, "rewards/format_reward/mean": 0.546875, "rewards/format_reward/std": 0.30496877431869507, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "step": 18 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1826.0, "completions/mean_length": 688.0625, "completions/mean_terminated_length": 666.4762573242188, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.04241071428571429, "frac_reward_zero_std": 0.0, "grad_norm": 2.2295756340026855, "learning_rate": 9.545032675245813e-07, "loss": 0.0, "num_tokens": 2357477.0, "reward": 0.49843746423721313, "reward_std": 0.3448905348777771, "rewards/format_reward/mean": 0.453125, "rewards/format_reward/std": 0.1717960685491562, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2025.0, "completions/mean_length": 543.28125, "completions/mean_terminated_length": 519.3968505859375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.044642857142857144, "frac_reward_zero_std": 0.0, "grad_norm": 8.84579086303711, "learning_rate": 9.473646649103817e-07, "loss": 0.0, "num_tokens": 2507007.0, "reward": 0.3828125, "reward_std": 0.38536137342453003, "rewards/format_reward/mean": 0.390625, "rewards/format_reward/std": 0.28824523091316223, "rewards/mcq_exact_match_reward/mean": 0.34375, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "step": 20 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1227.0, "completions/mean_length": 490.15625, "completions/mean_terminated_length": 465.4285888671875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.046875, "frac_reward_zero_std": 0.0, "grad_norm": 4.950660705566406, "learning_rate": 9.397368756032444e-07, "loss": 0.0, "num_tokens": 2638873.0, "reward": 0.29609373211860657, "reward_std": 0.30609115958213806, "rewards/format_reward/mean": 0.4609375, "rewards/format_reward/std": 0.2236899733543396, "rewards/mcq_exact_match_reward/mean": 0.25, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "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": 0.0625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1645.0, "completions/mean_length": 654.34375, "completions/mean_terminated_length": 561.433349609375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.049107142857142856, "frac_reward_zero_std": 0.0, "grad_norm": 3.7237157821655273, "learning_rate": 9.316282404787869e-07, "loss": 0.0, "num_tokens": 2773527.0, "reward": 0.3023437261581421, "reward_std": 0.2951734960079193, "rewards/format_reward/mean": 0.5234375, "rewards/format_reward/std": 0.25865477323532104, "rewards/mcq_exact_match_reward/mean": 0.25, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2003.0, "completions/mean_length": 572.84375, "completions/mean_terminated_length": 525.258056640625, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.05133928571428571, "frac_reward_zero_std": 0.0, "grad_norm": 2.378859758377075, "learning_rate": 9.230476262104676e-07, "loss": 0.0, "num_tokens": 2896093.0, "reward": 0.19999998807907104, "reward_std": 0.31976455450057983, "rewards/format_reward/mean": 0.59375, "rewards/format_reward/std": 0.3196600377559662, "rewards/mcq_exact_match_reward/mean": 0.140625, "rewards/mcq_exact_match_reward/std": 0.3503824472427368, "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": 1945.0, "completions/mean_length": 526.09375, "completions/mean_terminated_length": 501.9365234375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.05357142857142857, "frac_reward_zero_std": 0.0, "grad_norm": 8.608967781066895, "learning_rate": 9.1400441557401e-07, "loss": 0.0, "num_tokens": 3021931.0, "reward": 0.45156246423721313, "reward_std": 0.4028390347957611, "rewards/format_reward/mean": 0.609375, "rewards/format_reward/std": 0.2592533528804779, "rewards/mcq_exact_match_reward/mean": 0.390625, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "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": 1521.0, "completions/max_terminated_length": 1521.0, "completions/mean_length": 499.296875, "completions/mean_terminated_length": 499.296875, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.05580357142857143, "frac_reward_zero_std": 0.0, "grad_norm": 1.4608433246612549, "learning_rate": 9.045084971874737e-07, "loss": 0.0, "num_tokens": 3148094.0, "reward": 0.26171875, "reward_std": 0.28530025482177734, "rewards/format_reward/mean": 0.5859375, "rewards/format_reward/std": 0.3398726284503937, "rewards/mcq_exact_match_reward/mean": 0.203125, "rewards/mcq_exact_match_reward/std": 0.40550529956817627, "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": 1995.0, "completions/max_terminated_length": 1995.0, "completions/mean_length": 606.828125, "completions/mean_terminated_length": 606.828125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.05803571428571429, "frac_reward_zero_std": 0.0, "grad_norm": 4.1999640464782715, "learning_rate": 8.945702546981968e-07, "loss": 0.0, "num_tokens": 3289635.0, "reward": 0.30781251192092896, "reward_std": 0.3465702533721924, "rewards/format_reward/mean": 0.578125, "rewards/format_reward/std": 0.25539806485176086, "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": 1643.0, "completions/max_terminated_length": 1643.0, "completions/mean_length": 315.59375, "completions/mean_terminated_length": 315.59375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.060267857142857144, "frac_reward_zero_std": 0.0, "grad_norm": 9.185786247253418, "learning_rate": 8.842005554284295e-07, "loss": -0.0, "num_tokens": 3397177.0, "reward": 0.40625, "reward_std": 0.3422674238681793, "rewards/format_reward/mean": 0.625, "rewards/format_reward/std": 0.28171807527542114, "rewards/mcq_exact_match_reward/mean": 0.34375, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "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": 1005.0, "completions/max_terminated_length": 1005.0, "completions/mean_length": 236.4375, "completions/mean_terminated_length": 236.4375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0625, "frac_reward_zero_std": 0.125, "grad_norm": 15.800006866455078, "learning_rate": 8.734107384920769e-07, "loss": -0.0, "num_tokens": 3505253.0, "reward": 0.5593749284744263, "reward_std": 0.40181463956832886, "rewards/format_reward/mean": 0.59375, "rewards/format_reward/std": 0.25, "rewards/mcq_exact_match_reward/mean": 0.5, "rewards/mcq_exact_match_reward/std": 0.5039526224136353, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1942.0, "completions/mean_length": 559.359375, "completions/mean_terminated_length": 535.7301635742188, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.06473214285714286, "frac_reward_zero_std": 0.125, "grad_norm": 11.225300788879395, "learning_rate": 8.622126023955445e-07, "loss": -0.0, "num_tokens": 3644956.0, "reward": 0.27968746423721313, "reward_std": 0.28882110118865967, "rewards/format_reward/mean": 0.453125, "rewards/format_reward/std": 0.24750742316246033, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "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": 801.0, "completions/max_terminated_length": 801.0, "completions/mean_length": 205.65625, "completions/mean_terminated_length": 205.65625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.06696428571428571, "frac_reward_zero_std": 0.125, "grad_norm": 14.722556114196777, "learning_rate": 8.506183921362442e-07, "loss": 0.0, "num_tokens": 3740878.0, "reward": 0.6921875476837158, "reward_std": 0.2401888072490692, "rewards/format_reward/mean": 0.671875, "rewards/format_reward/std": 0.23935678601264954, "rewards/mcq_exact_match_reward/mean": 0.625, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "step": 30 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1640.0, "completions/max_terminated_length": 1640.0, "completions/mean_length": 274.78125, "completions/mean_terminated_length": 274.78125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.06919642857142858, "frac_reward_zero_std": 0.0, "grad_norm": 17.6087703704834, "learning_rate": 8.386407858128706e-07, "loss": -0.0, "num_tokens": 3848656.0, "reward": 0.24140623211860657, "reward_std": 0.3355621099472046, "rewards/format_reward/mean": 0.6953125, "rewards/format_reward/std": 0.2762732207775116, "rewards/mcq_exact_match_reward/mean": 0.171875, "rewards/mcq_exact_match_reward/std": 0.38025420904159546, "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": 1539.0, "completions/max_terminated_length": 1539.0, "completions/mean_length": 463.765625, "completions/mean_terminated_length": 463.765625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.07142857142857142, "frac_reward_zero_std": 0.0, "grad_norm": 22.964920043945312, "learning_rate": 8.262928807620843e-07, "loss": 0.0, "num_tokens": 3963865.0, "reward": 0.31953126192092896, "reward_std": 0.32839491963386536, "rewards/format_reward/mean": 0.5390625, "rewards/format_reward/std": 0.3125, "rewards/mcq_exact_match_reward/mean": 0.265625, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 780.0, "completions/mean_length": 287.796875, "completions/mean_terminated_length": 231.01612854003906, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.07366071428571429, "frac_reward_zero_std": 0.0, "grad_norm": 11.87487506866455, "learning_rate": 8.135881792367685e-07, "loss": -0.0, "num_tokens": 4084156.0, "reward": 0.4976562261581421, "reward_std": 0.44011881947517395, "rewards/format_reward/mean": 0.6015625, "rewards/format_reward/std": 0.3583437204360962, "rewards/mcq_exact_match_reward/mean": 0.4375, "rewards/mcq_exact_match_reward/std": 0.5, "step": 33 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1871.0, "completions/mean_length": 337.640625, "completions/mean_terminated_length": 282.4677429199219, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.07589285714285714, "frac_reward_zero_std": 0.125, "grad_norm": 10.959040641784668, "learning_rate": 8.005405736415125e-07, "loss": -0.0, "num_tokens": 4230133.0, "reward": 0.41718748211860657, "reward_std": 0.27033573389053345, "rewards/format_reward/mean": 0.578125, "rewards/format_reward/std": 0.29839184880256653, "rewards/mcq_exact_match_reward/mean": 0.359375, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "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": 1117.0, "completions/max_terminated_length": 1117.0, "completions/mean_length": 156.953125, "completions/mean_terminated_length": 156.953125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.078125, "frac_reward_zero_std": 0.0, "grad_norm": 18.629175186157227, "learning_rate": 7.871643313414718e-07, "loss": 0.0, "num_tokens": 4350706.0, "reward": 0.34453123807907104, "reward_std": 0.333055317401886, "rewards/format_reward/mean": 0.7890625, "rewards/format_reward/std": 0.3759083151817322, "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": 0.109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 973.0, "completions/mean_length": 355.03125, "completions/mean_terminated_length": 147.122802734375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08035714285714286, "frac_reward_zero_std": 0.25, "grad_norm": 14.444250106811523, "learning_rate": 7.734740790612136e-07, "loss": 0.0, "num_tokens": 4455580.0, "reward": 0.33281251788139343, "reward_std": 0.2278469204902649, "rewards/format_reward/mean": 0.671875, "rewards/format_reward/std": 0.43842023611068726, "rewards/mcq_exact_match_reward/mean": 0.265625, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "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": 494.0, "completions/mean_length": 81.59375, "completions/mean_terminated_length": 50.38095474243164, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08258928571428571, "frac_reward_zero_std": 0.25, "grad_norm": 12.015493392944336, "learning_rate": 7.594847868906076e-07, "loss": -0.0, "num_tokens": 4555386.0, "reward": 0.58203125, "reward_std": 0.2816838026046753, "rewards/format_reward/mean": 0.8203125, "rewards/format_reward/std": 0.27265870571136475, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 931.0, "completions/mean_length": 94.375, "completions/mean_terminated_length": 63.365081787109375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08482142857142858, "frac_reward_zero_std": 0.25, "grad_norm": 21.68492889404297, "learning_rate": 7.452117519152541e-07, "loss": 0.0, "num_tokens": 4654522.0, "reward": 0.3843749761581421, "reward_std": 0.32893675565719604, "rewards/format_reward/mean": 0.71875, "rewards/format_reward/std": 0.279951810836792, "rewards/mcq_exact_match_reward/mean": 0.3125, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "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": 768.0, "completions/max_terminated_length": 768.0, "completions/mean_length": 103.640625, "completions/mean_terminated_length": 103.640625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08705357142857142, "frac_reward_zero_std": 0.375, "grad_norm": 8.8945894241333, "learning_rate": 7.306705814893439e-07, "loss": -0.0, "num_tokens": 4728763.0, "reward": 0.51953125, "reward_std": 0.15724021196365356, "rewards/format_reward/mean": 0.9765625, "rewards/format_reward/std": 0.10652101784944534, "rewards/mcq_exact_match_reward/mean": 0.421875, "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": 1689.0, "completions/max_terminated_length": 1689.0, "completions/mean_length": 109.890625, "completions/mean_terminated_length": 109.890625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08928571428571429, "frac_reward_zero_std": 0.125, "grad_norm": 33.39426040649414, "learning_rate": 7.158771761692464e-07, "loss": -0.0, "num_tokens": 4842300.0, "reward": 0.5140625238418579, "reward_std": 0.35162004828453064, "rewards/format_reward/mean": 0.921875, "rewards/format_reward/std": 0.18298126757144928, "rewards/mcq_exact_match_reward/mean": 0.421875, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 835.0, "completions/mean_length": 125.484375, "completions/mean_terminated_length": 94.96826171875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.09151785714285714, "frac_reward_zero_std": 0.125, "grad_norm": 29.336641311645508, "learning_rate": 7.008477123264847e-07, "loss": 0.0, "num_tokens": 4959531.0, "reward": 0.606249988079071, "reward_std": 0.36156216263771057, "rewards/format_reward/mean": 0.75, "rewards/format_reward/std": 0.26726123690605164, "rewards/mcq_exact_match_reward/mean": 0.53125, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "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": 1656.0, "completions/max_terminated_length": 1656.0, "completions/mean_length": 97.90625, "completions/mean_terminated_length": 97.90625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.09375, "frac_reward_zero_std": 0.25, "grad_norm": 18.094688415527344, "learning_rate": 6.855986244591103e-07, "loss": 0.0, "num_tokens": 5079949.0, "reward": 0.34843748807907104, "reward_std": 0.3434920907020569, "rewards/format_reward/mean": 0.828125, "rewards/format_reward/std": 0.23935678601264954, "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": 184.0, "completions/max_terminated_length": 184.0, "completions/mean_length": 13.34375, "completions/mean_terminated_length": 13.34375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.09598214285714286, "frac_reward_zero_std": 0.25, "grad_norm": 29.087217330932617, "learning_rate": 6.701465872208216e-07, "loss": -0.0, "num_tokens": 5194635.0, "reward": 0.4085937440395355, "reward_std": 0.2798510491847992, "rewards/format_reward/mean": 0.8046875, "rewards/format_reward/std": 0.2615155577659607, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "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": 651.0, "completions/max_terminated_length": 651.0, "completions/mean_length": 44.703125, "completions/mean_terminated_length": 44.703125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.09821428571428571, "frac_reward_zero_std": 0.5, "grad_norm": 21.405025482177734, "learning_rate": 6.545084971874736e-07, "loss": 0.0, "num_tokens": 5283472.0, "reward": 0.63671875, "reward_std": 0.19260740280151367, "rewards/format_reward/mean": 0.8984375, "rewards/format_reward/std": 0.20275264978408813, "rewards/mcq_exact_match_reward/mean": 0.546875, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "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": 1221.0, "completions/max_terminated_length": 1221.0, "completions/mean_length": 39.0625, "completions/mean_terminated_length": 39.0625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.10044642857142858, "frac_reward_zero_std": 0.75, "grad_norm": 3.6213607788085938, "learning_rate": 6.387014543809223e-07, "loss": -0.0, "num_tokens": 5376116.0, "reward": 0.23984375596046448, "reward_std": 0.046403881162405014, "rewards/format_reward/mean": 0.9921875, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.140625, "rewards/mcq_exact_match_reward/std": 0.3503824472427368, "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": 425.0, "completions/max_terminated_length": 425.0, "completions/mean_length": 21.953125, "completions/mean_terminated_length": 21.953125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.10267857142857142, "frac_reward_zero_std": 0.25, "grad_norm": 44.87046813964844, "learning_rate": 6.227427435703995e-07, "loss": 0.0, "num_tokens": 5484241.0, "reward": 0.612500011920929, "reward_std": 0.27364858984947205, "rewards/format_reward/mean": 0.96875, "rewards/format_reward/std": 0.12198751419782639, "rewards/mcq_exact_match_reward/mean": 0.515625, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "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": 116.0, "completions/max_terminated_length": 116.0, "completions/mean_length": 13.78125, "completions/mean_terminated_length": 13.78125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.10491071428571429, "frac_reward_zero_std": 0.625, "grad_norm": 21.078166961669922, "learning_rate": 6.066498153718734e-07, "loss": 0.0, "num_tokens": 5542659.0, "reward": 0.3929687738418579, "reward_std": 0.10529261827468872, "rewards/format_reward/mean": 0.9609375, "rewards/format_reward/std": 0.13524486124515533, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "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": 1428.0, "completions/max_terminated_length": 1428.0, "completions/mean_length": 39.84375, "completions/mean_terminated_length": 39.84375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.10714285714285714, "frac_reward_zero_std": 0.75, "grad_norm": 6.818737506866455, "learning_rate": 5.90440267166055e-07, "loss": -0.0, "num_tokens": 5633841.0, "reward": 0.4906250238418579, "reward_std": 0.12255740165710449, "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": 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": 760.0, "completions/max_terminated_length": 760.0, "completions/mean_length": 29.453125, "completions/mean_terminated_length": 29.453125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.109375, "frac_reward_zero_std": 0.5, "grad_norm": 11.917037010192871, "learning_rate": 5.741318238559209e-07, "loss": -0.0, "num_tokens": 5717382.0, "reward": 0.23359376192092896, "reward_std": 0.18003450334072113, "rewards/format_reward/mean": 0.9296875, "rewards/format_reward/std": 0.1751912236213684, "rewards/mcq_exact_match_reward/mean": 0.140625, "rewards/mcq_exact_match_reward/std": 0.3503824472427368, "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": 366.0, "completions/max_terminated_length": 366.0, "completions/mean_length": 29.3125, "completions/mean_terminated_length": 29.3125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.11160714285714286, "frac_reward_zero_std": 0.375, "grad_norm": 13.875815391540527, "learning_rate": 5.577423184847931e-07, "loss": 0.0, "num_tokens": 5825506.0, "reward": 0.612500011920929, "reward_std": 0.30738940834999084, "rewards/format_reward/mean": 0.96875, "rewards/format_reward/std": 0.17536810040473938, "rewards/mcq_exact_match_reward/mean": 0.515625, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "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": 340.0, "completions/max_terminated_length": 340.0, "completions/mean_length": 21.484375, "completions/mean_terminated_length": 21.484375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.11383928571428571, "frac_reward_zero_std": 0.75, "grad_norm": 11.78389835357666, "learning_rate": 5.412896727361662e-07, "loss": 0.0, "num_tokens": 5900009.0, "reward": 0.5953125357627869, "reward_std": 0.11684979498386383, "rewards/format_reward/mean": 0.953125, "rewards/format_reward/std": 0.1717960685491562, "rewards/mcq_exact_match_reward/mean": 0.5, "rewards/mcq_exact_match_reward/std": 0.5039526224136353, "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": 124.0, "completions/max_terminated_length": 124.0, "completions/mean_length": 14.515625, "completions/mean_terminated_length": 14.515625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.11607142857142858, "frac_reward_zero_std": 0.5, "grad_norm": 17.35973358154297, "learning_rate": 5.247918773366111e-07, "loss": -0.0, "num_tokens": 5985602.0, "reward": 0.7250000238418579, "reward_std": 0.22461533546447754, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.625, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "step": 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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 11.65625, "completions/mean_terminated_length": 11.65625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.11830357142857142, "frac_reward_zero_std": 0.5, "grad_norm": 45.946311950683594, "learning_rate": 5.082669723831793e-07, "loss": 0.0, "num_tokens": 6095956.0, "reward": 0.828125, "reward_std": 0.18517296016216278, "rewards/format_reward/mean": 0.9375, "rewards/format_reward/std": 0.1666666716337204, "rewards/mcq_exact_match_reward/mean": 0.734375, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "step": 53 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 154.0, "completions/max_terminated_length": 154.0, "completions/mean_length": 15.078125, "completions/mean_terminated_length": 15.078125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.12053571428571429, "frac_reward_zero_std": 0.875, "grad_norm": 5.150336742401123, "learning_rate": 4.917330276168208e-07, "loss": -0.0, "num_tokens": 6201345.0, "reward": 0.47812503576278687, "reward_std": 0.04419417306780815, "rewards/format_reward/mean": 0.875, "rewards/format_reward/std": 0.3333333432674408, "rewards/mcq_exact_match_reward/mean": 0.390625, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "step": 54 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.734375, "completions/mean_terminated_length": 12.734375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.12276785714285714, "frac_reward_zero_std": 0.875, "grad_norm": 5.761120796203613, "learning_rate": 4.752081226633888e-07, "loss": 0.0, "num_tokens": 6279080.0, "reward": 0.5843750238418579, "reward_std": 0.04419417306780815, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.484375, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "step": 55 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 365.0, "completions/max_terminated_length": 365.0, "completions/mean_length": 18.28125, "completions/mean_terminated_length": 18.28125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.125, "frac_reward_zero_std": 0.625, "grad_norm": 9.205299377441406, "learning_rate": 4.5871032726383385e-07, "loss": -0.0, "num_tokens": 6365402.0, "reward": 0.5523437261581421, "reward_std": 0.10426764190196991, "rewards/format_reward/mean": 0.9921875, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 56 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 86.0, "completions/max_terminated_length": 86.0, "completions/mean_length": 13.890625, "completions/mean_terminated_length": 13.890625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.12723214285714285, "frac_reward_zero_std": 0.875, "grad_norm": 4.692060470581055, "learning_rate": 4.4225768151520694e-07, "loss": 0.0, "num_tokens": 6469867.0, "reward": 0.4437500238418579, "reward_std": 0.0578637570142746, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.34375, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "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": 367.0, "completions/mean_length": 50.109375, "completions/mean_terminated_length": 18.396825790405273, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.12946428571428573, "frac_reward_zero_std": 0.75, "grad_norm": 8.68276309967041, "learning_rate": 4.258681761440789e-07, "loss": 0.0, "num_tokens": 6559138.0, "reward": 0.5046875476837158, "reward_std": 0.08923259377479553, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "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": 24.0, "completions/max_terminated_length": 24.0, "completions/mean_length": 13.109375, "completions/mean_terminated_length": 13.109375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.13169642857142858, "frac_reward_zero_std": 0.625, "grad_norm": 7.954644203186035, "learning_rate": 4.095597328339452e-07, "loss": 0.0, "num_tokens": 6663777.0, "reward": 0.550000011920929, "reward_std": 0.07353247702121735, "rewards/format_reward/mean": 0.96875, "rewards/format_reward/std": 0.17536810040473938, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "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": 15.0, "completions/max_terminated_length": 15.0, "completions/mean_length": 12.65625, "completions/mean_terminated_length": 12.65625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.13392857142857142, "frac_reward_zero_std": 0.5, "grad_norm": 13.709247589111328, "learning_rate": 3.9335018462812664e-07, "loss": 0.0, "num_tokens": 6749379.0, "reward": 0.6468750238418579, "reward_std": 0.2109457552433014, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.546875, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.796875, "completions/mean_terminated_length": 12.796875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.13616071428571427, "frac_reward_zero_std": 0.75, "grad_norm": 8.133370399475098, "learning_rate": 3.772572564296004e-07, "loss": 0.0, "num_tokens": 6858742.0, "reward": 0.6312500238418579, "reward_std": 0.10888782143592834, "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": 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": 271.0, "completions/max_terminated_length": 271.0, "completions/mean_length": 17.046875, "completions/mean_terminated_length": 17.046875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.13839285714285715, "frac_reward_zero_std": 0.5, "grad_norm": 16.2270450592041, "learning_rate": 3.612985456190778e-07, "loss": 0.0, "num_tokens": 6971881.0, "reward": 0.6781250238418579, "reward_std": 0.2198973000049591, "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": 62 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 123.0, "completions/max_terminated_length": 123.0, "completions/mean_length": 14.6875, "completions/mean_terminated_length": 14.6875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.140625, "frac_reward_zero_std": 0.5, "grad_norm": 37.52165222167969, "learning_rate": 3.454915028125263e-07, "loss": 0.0, "num_tokens": 7066509.0, "reward": 0.3554687798023224, "reward_std": 0.1327619105577469, "rewards/format_reward/mean": 0.8984375, "rewards/format_reward/std": 0.25479042530059814, "rewards/mcq_exact_match_reward/mean": 0.265625, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.84375, "completions/mean_terminated_length": 12.84375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.14285714285714285, "frac_reward_zero_std": 0.5, "grad_norm": 15.275923728942871, "learning_rate": 3.2985341277917846e-07, "loss": 0.0, "num_tokens": 7157859.0, "reward": 0.7546875476837158, "reward_std": 0.15809683501720428, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.08768405020236969, "rewards/mcq_exact_match_reward/mean": 0.65625, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "step": 64 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 138.0, "completions/max_terminated_length": 138.0, "completions/mean_length": 14.75, "completions/mean_terminated_length": 14.75, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.14508928571428573, "frac_reward_zero_std": 0.75, "grad_norm": 13.033477783203125, "learning_rate": 3.1440137554088953e-07, "loss": 0.0, "num_tokens": 7244707.0, "reward": 0.4281250238418579, "reward_std": 0.10205793380737305, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "step": 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": 359.0, "completions/max_terminated_length": 359.0, "completions/mean_length": 18.046875, "completions/mean_terminated_length": 18.046875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.14732142857142858, "frac_reward_zero_std": 0.625, "grad_norm": 8.87108039855957, "learning_rate": 2.9915228767351535e-07, "loss": 0.0, "num_tokens": 7353542.0, "reward": 0.5531250238418579, "reward_std": 0.1530819982290268, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 66 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 264.0, "completions/max_terminated_length": 264.0, "completions/mean_length": 16.484375, "completions/mean_terminated_length": 16.484375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.14955357142857142, "frac_reward_zero_std": 0.75, "grad_norm": 7.64149284362793, "learning_rate": 2.841228238307536e-07, "loss": -0.0, "num_tokens": 7447069.0, "reward": 0.6468750238418579, "reward_std": 0.12255740165710449, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.546875, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 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": 42.0, "completions/max_terminated_length": 42.0, "completions/mean_length": 13.125, "completions/mean_terminated_length": 13.125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.15178571428571427, "frac_reward_zero_std": 0.625, "grad_norm": 15.826613426208496, "learning_rate": 2.6932941851065615e-07, "loss": 0.0, "num_tokens": 7542821.0, "reward": 0.5062500238418579, "reward_std": 0.1735912710428238, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "step": 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": 443.0, "completions/max_terminated_length": 443.0, "completions/mean_length": 30.84375, "completions/mean_terminated_length": 30.84375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.15401785714285715, "frac_reward_zero_std": 0.625, "grad_norm": 17.32259750366211, "learning_rate": 2.547882480847461e-07, "loss": -0.0, "num_tokens": 7624195.0, "reward": 0.4125000238418579, "reward_std": 0.1872510462999344, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.3125, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "step": 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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.9375, "completions/mean_terminated_length": 12.9375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.15625, "frac_reward_zero_std": 0.75, "grad_norm": 13.53782844543457, "learning_rate": 2.4051521310939254e-07, "loss": 0.0, "num_tokens": 7711775.0, "reward": 0.6156250238418579, "reward_std": 0.10205793380737305, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.515625, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "step": 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": 17.0, "completions/max_terminated_length": 17.0, "completions/mean_length": 12.75, "completions/mean_terminated_length": 12.75, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.15848214285714285, "frac_reward_zero_std": 0.625, "grad_norm": 12.834123611450195, "learning_rate": 2.2652592093878665e-07, "loss": 0.0, "num_tokens": 7798103.0, "reward": 0.9125000238418579, "reward_std": 0.16675157845020294, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.8125, "rewards/mcq_exact_match_reward/std": 0.39339789748191833, "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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.6875, "completions/mean_terminated_length": 12.6875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.16071428571428573, "frac_reward_zero_std": 1.0, "grad_norm": 0.0, "learning_rate": 2.128356686585282e-07, "loss": 0.0, "num_tokens": 7884851.0, "reward": 0.8500000238418579, "reward_std": 0.0, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.75, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "step": 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": 13.0, "completions/mean_length": 44.625, "completions/mean_terminated_length": 12.825397491455078, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.16294642857142858, "frac_reward_zero_std": 0.875, "grad_norm": 1.9047719240188599, "learning_rate": 1.9945942635848745e-07, "loss": 0.0, "num_tokens": 7992155.0, "reward": 0.45781251788139343, "reward_std": 0.048613592982292175, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.359375, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "step": 73 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 105.0, "completions/max_terminated_length": 105.0, "completions/mean_length": 14.78125, "completions/mean_terminated_length": 14.78125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.16517857142857142, "frac_reward_zero_std": 0.625, "grad_norm": 11.897013664245605, "learning_rate": 1.8641182076323148e-07, "loss": -0.0, "num_tokens": 8109837.0, "reward": 0.62109375, "reward_std": 0.10426764190196991, "rewards/format_reward/mean": 0.7421875, "rewards/format_reward/std": 0.4363647401332855, "rewards/mcq_exact_match_reward/mean": 0.546875, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 74 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.515625, "completions/mean_terminated_length": 12.515625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.16741071428571427, "frac_reward_zero_std": 0.75, "grad_norm": 15.865518569946289, "learning_rate": 1.7370711923791564e-07, "loss": 0.0, "num_tokens": 8193150.0, "reward": 0.7093750238418579, "reward_std": 0.12255740165710449, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.609375, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "step": 75 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.8125, "completions/mean_terminated_length": 12.8125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.16964285714285715, "frac_reward_zero_std": 0.75, "grad_norm": 8.128179550170898, "learning_rate": 1.6135921418712955e-07, "loss": 0.0, "num_tokens": 8289962.0, "reward": 0.5843750238418579, "reward_std": 0.1315089464187622, "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": 353.0, "completions/max_terminated_length": 353.0, "completions/mean_length": 22.734375, "completions/mean_terminated_length": 22.734375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.171875, "frac_reward_zero_std": 0.625, "grad_norm": 6.7019805908203125, "learning_rate": 1.493816078637557e-07, "loss": 0.0, "num_tokens": 8404241.0, "reward": 0.5531250238418579, "reward_std": 0.13258251547813416, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 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": 284.0, "completions/max_terminated_length": 284.0, "completions/mean_length": 17.171875, "completions/mean_terminated_length": 17.171875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.17410714285714285, "frac_reward_zero_std": 0.625, "grad_norm": 17.198396682739258, "learning_rate": 1.3778739760445552e-07, "loss": -0.0, "num_tokens": 8495196.0, "reward": 0.5210937261581421, "reward_std": 0.10426764190196991, "rewards/format_reward/mean": 0.9921875, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.421875, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.734375, "completions/mean_terminated_length": 12.734375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.17633928571428573, "frac_reward_zero_std": 0.75, "grad_norm": 12.526077270507812, "learning_rate": 1.2658926150792322e-07, "loss": 0.0, "num_tokens": 8571915.0, "reward": 0.823437511920929, "reward_std": 0.04861358925700188, "rewards/format_reward/mean": 0.890625, "rewards/format_reward/std": 0.3145764470100403, "rewards/mcq_exact_match_reward/mean": 0.734375, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.765625, "completions/mean_terminated_length": 12.765625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.17857142857142858, "frac_reward_zero_std": 0.75, "grad_norm": 7.538797378540039, "learning_rate": 1.1579944457157059e-07, "loss": 0.0, "num_tokens": 8675380.0, "reward": 0.8187500238418579, "reward_std": 0.0883883461356163, "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": 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": 296.0, "completions/max_terminated_length": 296.0, "completions/mean_length": 17.09375, "completions/mean_terminated_length": 17.09375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.18080357142857142, "frac_reward_zero_std": 0.875, "grad_norm": 9.688939094543457, "learning_rate": 1.0542974530180327e-07, "loss": -0.0, "num_tokens": 8800610.0, "reward": 0.7406250238418579, "reward_std": 0.04419417306780815, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.640625, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "step": 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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.859375, "completions/mean_terminated_length": 12.859375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.18303571428571427, "frac_reward_zero_std": 0.875, "grad_norm": 9.010716438293457, "learning_rate": 9.549150281252632e-08, "loss": 0.0, "num_tokens": 8913425.0, "reward": 0.5062500238418579, "reward_std": 0.0578637570142746, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "step": 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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 17.0, "completions/mean_length": 44.46875, "completions/mean_terminated_length": 12.666666984558105, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.18526785714285715, "frac_reward_zero_std": 0.875, "grad_norm": 1.4834736585617065, "learning_rate": 8.599558442598998e-08, "loss": 0.0, "num_tokens": 9028295.0, "reward": 0.5828125476837158, "reward_std": 0.048613592982292175, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.484375, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.609375, "completions/mean_terminated_length": 12.609375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1875, "frac_reward_zero_std": 0.875, "grad_norm": 14.377280235290527, "learning_rate": 7.695237378953224e-08, "loss": 0.0, "num_tokens": 9165494.0, "reward": 0.5218750238418579, "reward_std": 0.0646936446428299, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.421875, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "step": 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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.734375, "completions/mean_terminated_length": 12.734375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.18973214285714285, "frac_reward_zero_std": 0.75, "grad_norm": 14.24134349822998, "learning_rate": 6.837175952121304e-08, "loss": 0.0, "num_tokens": 9242997.0, "reward": 0.6781250238418579, "reward_std": 0.12255740165710449, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.578125, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "step": 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": 12.6875, "completions/mean_terminated_length": 12.6875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.19196428571428573, "frac_reward_zero_std": 0.5, "grad_norm": 16.786251068115234, "learning_rate": 6.026312439675551e-08, "loss": 0.0, "num_tokens": 9332529.0, "reward": 0.7406250238418579, "reward_std": 0.2109457552433014, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.640625, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "step": 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": 397.0, "completions/max_terminated_length": 397.0, "completions/mean_length": 18.640625, "completions/mean_terminated_length": 18.640625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.19419642857142858, "frac_reward_zero_std": 0.75, "grad_norm": 10.523579597473145, "learning_rate": 5.263533508961826e-08, "loss": -0.0, "num_tokens": 9408450.0, "reward": 0.4281250238418579, "reward_std": 0.11100947856903076, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "step": 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": 14.0, "completions/max_terminated_length": 14.0, "completions/mean_length": 12.65625, "completions/mean_terminated_length": 12.65625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.19642857142857142, "frac_reward_zero_std": 0.875, "grad_norm": 5.4456706047058105, "learning_rate": 4.549673247541874e-08, "loss": 0.0, "num_tokens": 9491852.0, "reward": 0.5984375476837158, "reward_std": 0.0044194171205163, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.5, "rewards/mcq_exact_match_reward/std": 0.5039526224136353, "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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.734375, "completions/mean_terminated_length": 12.734375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.19866071428571427, "frac_reward_zero_std": 0.75, "grad_norm": 12.022665977478027, "learning_rate": 3.8855122511307626e-08, "loss": 0.0, "num_tokens": 9580907.0, "reward": 0.7562500238418579, "reward_std": 0.1293872892856598, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.65625, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "step": 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": 16.0, "completions/max_terminated_length": 16.0, "completions/mean_length": 12.78125, "completions/mean_terminated_length": 12.78125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.20089285714285715, "frac_reward_zero_std": 0.875, "grad_norm": 5.465011119842529, "learning_rate": 3.271776770026963e-08, "loss": -0.0, "num_tokens": 9679413.0, "reward": 0.8187500238418579, "reward_std": 0.0578637570142746, "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": 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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.84375, "completions/mean_terminated_length": 12.84375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.203125, "frac_reward_zero_std": 0.75, "grad_norm": 14.83763313293457, "learning_rate": 2.7091379149682682e-08, "loss": -0.0, "num_tokens": 9780491.0, "reward": 0.16250000894069672, "reward_std": 0.10888782143592834, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.0625, "rewards/mcq_exact_match_reward/std": 0.24397502839565277, "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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.609375, "completions/mean_terminated_length": 12.609375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.20535714285714285, "frac_reward_zero_std": 0.75, "grad_norm": 8.963287353515625, "learning_rate": 2.1982109232821176e-08, "loss": 0.0, "num_tokens": 9874282.0, "reward": 0.8187500238418579, "reward_std": 0.0883883461356163, "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": 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": 130.0, "completions/max_terminated_length": 130.0, "completions/mean_length": 17.890625, "completions/mean_terminated_length": 17.890625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.20758928571428573, "frac_reward_zero_std": 0.875, "grad_norm": 13.938237190246582, "learning_rate": 1.7395544861325718e-08, "loss": 0.0, "num_tokens": 9957723.0, "reward": 0.5531250238418579, "reward_std": 0.0646936446428299, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.671875, "completions/mean_terminated_length": 12.671875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.20982142857142858, "frac_reward_zero_std": 0.5, "grad_norm": 17.77403450012207, "learning_rate": 1.3336701375997127e-08, "loss": 0.0, "num_tokens": 10066814.0, "reward": 0.5218750238418579, "reward_std": 0.2198973149061203, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.421875, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "step": 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": 16.0, "completions/max_terminated_length": 16.0, "completions/mean_length": 12.59375, "completions/mean_terminated_length": 12.59375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.21205357142857142, "frac_reward_zero_std": 0.625, "grad_norm": 20.48771858215332, "learning_rate": 9.810017062595321e-09, "loss": 0.0, "num_tokens": 10162468.0, "reward": 0.44062501192092896, "reward_std": 0.09417472034692764, "rewards/format_reward/mean": 0.96875, "rewards/format_reward/std": 0.17536810040473938, "rewards/mcq_exact_match_reward/mean": 0.34375, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "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": 380.0, "completions/max_terminated_length": 380.0, "completions/mean_length": 27.390625, "completions/mean_terminated_length": 27.390625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.21428571428571427, "frac_reward_zero_std": 0.875, "grad_norm": 3.7267138957977295, "learning_rate": 6.819348298638839e-09, "loss": 0.0, "num_tokens": 10272317.0, "reward": 0.4593750238418579, "reward_std": 0.04419417306780815, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.359375, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "step": 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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.578125, "completions/mean_terminated_length": 12.578125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.21651785714285715, "frac_reward_zero_std": 0.75, "grad_norm": 9.971395492553711, "learning_rate": 4.367965336512403e-09, "loss": 0.0, "num_tokens": 10352002.0, "reward": 0.8187500238418579, "reward_std": 0.1246790662407875, "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": 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": 626.0, "completions/max_terminated_length": 626.0, "completions/mean_length": 22.4375, "completions/mean_terminated_length": 22.4375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.21875, "frac_reward_zero_std": 0.75, "grad_norm": 6.507881164550781, "learning_rate": 2.458548727494292e-09, "loss": 0.0, "num_tokens": 10437238.0, "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": 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": 33.0, "completions/max_terminated_length": 33.0, "completions/mean_length": 14.140625, "completions/mean_terminated_length": 14.140625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.22098214285714285, "frac_reward_zero_std": 0.75, "grad_norm": 6.7576584815979, "learning_rate": 1.0931863906127325e-09, "loss": -0.0, "num_tokens": 10576031.0, "reward": 0.6062500476837158, "reward_std": 0.04998054727911949, "rewards/format_reward/mean": 0.90625, "rewards/format_reward/std": 0.29378482699394226, "rewards/mcq_exact_match_reward/mean": 0.515625, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.71875, "completions/mean_terminated_length": 12.71875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.22321428571428573, "frac_reward_zero_std": 0.875, "grad_norm": 12.548639297485352, "learning_rate": 2.733713295369755e-10, "loss": 0.0, "num_tokens": 10646405.0, "reward": 0.6468750238418579, "reward_std": 0.0646936446428299, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.546875, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 100 } ], "logging_steps": 1, "max_steps": 100, "num_input_tokens_seen": 10646405, "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 }