{ "best_global_step": null, "best_metric": null, "best_model_checkpoint": null, "epoch": 0.2494331065759637, "eval_steps": 500, "global_step": 110, "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3681.0, "completions/mean_length": 1268.2265625, "completions/mean_terminated_length": 980.7562866210938, "completions/min_length": 98.0, "completions/min_terminated_length": 98.0, "entropy": 4.420659868046641, "epoch": 0.0022675736961451248, "frac_reward_zero_std": 0.875, "grad_norm": 0.19921875, "learning_rate": 1e-06, "loss": -0.0093, "num_tokens": 184821.0, "reward": 0.06640625, "reward_std": 0.31472553312778473, "rewards/accuracy_reward_func/mean": 0.046875, "rewards/accuracy_reward_func/std": 0.3003681004047394, "rewards/format_reward_func/mean": 0.01953125, "rewards/format_reward_func/std": 0.09224375709891319, "step": 1, "step_time": 598.9364407490939 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1640625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3782.5, "completions/mean_length": 1565.546875, "completions/mean_terminated_length": 1068.8557739257812, "completions/min_length": 26.5, "completions/min_terminated_length": 26.5, "entropy": 4.373492445796728, "epoch": 0.0045351473922902496, "frac_reward_zero_std": 0.96875, "grad_norm": 0.06201171875, "learning_rate": 9.977324263038547e-07, "loss": 0.0007, "num_tokens": 410327.0, "reward": 0.03125, "reward_std": 0.13729241490364075, "rewards/accuracy_reward_func/mean": 0.015625, "rewards/accuracy_reward_func/std": 0.125, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.06099375709891319, "step": 2, "step_time": 600.5369204720482 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 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": 4096.0, "completions/max_terminated_length": 2910.0, "completions/mean_length": 1133.2265625, "completions/mean_terminated_length": 769.377197265625, "completions/min_length": 5.5, "completions/min_terminated_length": 5.5, "entropy": 2.3305343482643366, "epoch": 0.006802721088435374, "frac_reward_zero_std": 0.9375, "grad_norm": 0.193359375, "learning_rate": 9.954648526077097e-07, "loss": 0.0087, "num_tokens": 578144.0, "reward": 0.03125, "reward_std": 0.19257627427577972, "rewards/accuracy_reward_func/mean": 0.015625, "rewards/accuracy_reward_func/std": 0.125, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08451050892472267, "step": 3, "step_time": 589.7848729742691 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.2265625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3864.0, "completions/mean_length": 1646.265625, "completions/mean_terminated_length": 926.9349060058594, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 3.3248021751642227, "epoch": 0.009070294784580499, "frac_reward_zero_std": 0.90625, "grad_norm": 0.1484375, "learning_rate": 9.931972789115645e-07, "loss": 0.0086, "num_tokens": 812974.0, "reward": 0.0390625, "reward_std": 0.25652606785297394, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.043842025101184845, "step": 4, "step_time": 589.6438818126917 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3125.0, "completions/mean_length": 1278.7578125, "completions/mean_terminated_length": 817.7545166015625, "completions/min_length": 3.5, "completions/min_terminated_length": 3.5, "entropy": 3.488957514986396, "epoch": 0.011337868480725623, "frac_reward_zero_std": 0.90625, "grad_norm": 0.2314453125, "learning_rate": 9.909297052154193e-07, "loss": -0.0221, "num_tokens": 999435.0, "reward": 0.0234375, "reward_std": 0.1596149355173111, "rewards/accuracy_reward_func/mean": 0.015625, "rewards/accuracy_reward_func/std": 0.125, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 5, "step_time": 591.4307740237564 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1328125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3731.0, "completions/mean_length": 1252.328125, "completions/mean_terminated_length": 812.7178344726562, "completions/min_length": 1.5, "completions/min_terminated_length": 1.5, "entropy": 3.201830508187413, "epoch": 0.013605442176870748, "frac_reward_zero_std": 0.8125, "grad_norm": 0.21484375, "learning_rate": 9.886621315192743e-07, "loss": 0.0123, "num_tokens": 1188829.0, "reward": 0.125, "reward_std": 0.4914857894182205, "rewards/accuracy_reward_func/mean": 0.109375, "rewards/accuracy_reward_func/std": 0.45701706409454346, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08451050892472267, "step": 6, "step_time": 629.2572528030723 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1640625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3784.5, "completions/mean_length": 1448.6484375, "completions/mean_terminated_length": 928.8148193359375, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "entropy": 4.067663200199604, "epoch": 0.015873015873015872, "frac_reward_zero_std": 0.9375, "grad_norm": 0.10546875, "learning_rate": 9.86394557823129e-07, "loss": -0.0005, "num_tokens": 1396800.0, "reward": 0.0234375, "reward_std": 0.13152606785297394, "rewards/accuracy_reward_func/mean": 0.015625, "rewards/accuracy_reward_func/std": 0.125, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.043842025101184845, "step": 7, "step_time": 585.323605818674 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3718.5, "completions/mean_length": 1281.5234375, "completions/mean_terminated_length": 915.6332397460938, "completions/min_length": 102.5, "completions/min_terminated_length": 102.5, "entropy": 3.3765318896621466, "epoch": 0.018140589569160998, "frac_reward_zero_std": 0.96875, "grad_norm": 0.1357421875, "learning_rate": 9.84126984126984e-07, "loss": 0.0023, "num_tokens": 1586687.0, "reward": 0.0078125, "reward_std": 0.043842025101184845, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.043842025101184845, "step": 8, "step_time": 603.3711773892865 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.203125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3937.5, "completions/mean_length": 1789.2265625, "completions/mean_terminated_length": 1200.6988525390625, "completions/min_length": 11.0, "completions/min_terminated_length": 11.0, "entropy": 4.331676904112101, "epoch": 0.02040816326530612, "frac_reward_zero_std": 0.84375, "grad_norm": 0.1416015625, "learning_rate": 9.818594104308389e-07, "loss": 0.0198, "num_tokens": 1838880.0, "reward": 0.0625, "reward_std": 0.2596312239766121, "rewards/accuracy_reward_func/mean": 0.046875, "rewards/accuracy_reward_func/std": 0.21304203569889069, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 9, "step_time": 590.4591970965266 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.2109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3845.0, "completions/mean_length": 1610.34375, "completions/mean_terminated_length": 939.8783569335938, "completions/min_length": 20.0, "completions/min_terminated_length": 20.0, "entropy": 4.339755732566118, "epoch": 0.022675736961451247, "frac_reward_zero_std": 0.84375, "grad_norm": 0.171875, "learning_rate": 9.795918367346939e-07, "loss": -0.0001, "num_tokens": 2070844.0, "reward": 0.06640625, "reward_std": 0.26088467985391617, "rewards/accuracy_reward_func/mean": 0.046875, "rewards/accuracy_reward_func/std": 0.21304203569889069, "rewards/format_reward_func/mean": 0.01953125, "rewards/format_reward_func/std": 0.09710253402590752, "step": 10, "step_time": 608.92143144086 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4055.5, "completions/mean_length": 1540.3203125, "completions/mean_terminated_length": 1065.2804565429688, "completions/min_length": 9.5, "completions/min_terminated_length": 9.5, "entropy": 3.709390392526984, "epoch": 0.024943310657596373, "frac_reward_zero_std": 0.90625, "grad_norm": 0.1298828125, "learning_rate": 9.773242630385487e-07, "loss": 0.0305, "num_tokens": 2291637.0, "reward": 0.0390625, "reward_std": 0.17938801646232605, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.17536810040473938, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.043842025101184845, "step": 11, "step_time": 594.1065437020734 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3819.0, "completions/mean_length": 1434.7734375, "completions/mean_terminated_length": 941.9537048339844, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.84947507455945, "epoch": 0.027210884353741496, "frac_reward_zero_std": 0.875, "grad_norm": 0.18359375, "learning_rate": 9.750566893424036e-07, "loss": 0.0622, "num_tokens": 2500052.0, "reward": 0.1015625, "reward_std": 0.4368836283683777, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.41934312880039215, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.043842025101184845, "step": 12, "step_time": 598.8679339522496 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3955.0, "completions/mean_length": 1386.8984375, "completions/mean_terminated_length": 997.9778442382812, "completions/min_length": 4.5, "completions/min_terminated_length": 4.5, "entropy": 3.7934455443173647, "epoch": 0.02947845804988662, "frac_reward_zero_std": 0.90625, "grad_norm": 0.19921875, "learning_rate": 9.727891156462584e-07, "loss": 0.0114, "num_tokens": 2701547.0, "reward": 0.03125, "reward_std": 0.16574779152870178, "rewards/accuracy_reward_func/mean": 0.015625, "rewards/accuracy_reward_func/std": 0.125, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08451050892472267, "step": 13, "step_time": 627.0823687994853 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1328125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3859.5, "completions/mean_length": 1351.640625, "completions/mean_terminated_length": 931.5595703125, "completions/min_length": 46.5, "completions/min_terminated_length": 46.5, "entropy": 3.6855772957205772, "epoch": 0.031746031746031744, "frac_reward_zero_std": 0.84375, "grad_norm": 0.216796875, "learning_rate": 9.705215419501134e-07, "loss": -0.0323, "num_tokens": 2899161.0, "reward": 0.09375, "reward_std": 0.41640444099903107, "rewards/accuracy_reward_func/mean": 0.078125, "rewards/accuracy_reward_func/std": 0.38841013610363007, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08451050892472267, "step": 14, "step_time": 585.9838520288467 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3639.0, "completions/mean_length": 1279.6015625, "completions/mean_terminated_length": 874.271728515625, "completions/min_length": 3.5, "completions/min_terminated_length": 3.5, "entropy": 3.6237228345125914, "epoch": 0.034013605442176874, "frac_reward_zero_std": 0.9375, "grad_norm": 0.146484375, "learning_rate": 9.682539682539682e-07, "loss": 0.036, "num_tokens": 3087398.0, "reward": 0.0234375, "reward_std": 0.16884202510118484, "rewards/accuracy_reward_func/mean": 0.015625, "rewards/accuracy_reward_func/std": 0.125, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.043842025101184845, "step": 15, "step_time": 599.1850006729364 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3629.0, "completions/mean_length": 1518.09375, "completions/mean_terminated_length": 922.2556457519531, "completions/min_length": 4.5, "completions/min_terminated_length": 4.5, "entropy": 4.251630323007703, "epoch": 0.036281179138321996, "frac_reward_zero_std": 0.84375, "grad_norm": 0.1962890625, "learning_rate": 9.65986394557823e-07, "loss": 0.0335, "num_tokens": 3305370.0, "reward": 0.0859375, "reward_std": 0.3918892741203308, "rewards/accuracy_reward_func/mean": 0.078125, "rewards/accuracy_reward_func/std": 0.38841013610363007, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 16, "step_time": 594.1110223559663 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1640625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3773.0, "completions/mean_length": 1507.0234375, "completions/mean_terminated_length": 998.476806640625, "completions/min_length": 42.5, "completions/min_terminated_length": 42.5, "entropy": 3.8557076919823885, "epoch": 0.03854875283446712, "frac_reward_zero_std": 0.8125, "grad_norm": 0.201171875, "learning_rate": 9.63718820861678e-07, "loss": 0.0151, "num_tokens": 3523109.0, "reward": 0.0859375, "reward_std": 0.36247265338897705, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 17, "step_time": 587.1657324088737 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1796875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3525.0, "completions/mean_length": 1425.53125, "completions/mean_terminated_length": 841.9051208496094, "completions/min_length": 62.5, "completions/min_terminated_length": 62.5, "entropy": 3.771521097049117, "epoch": 0.04081632653061224, "frac_reward_zero_std": 0.875, "grad_norm": 0.1708984375, "learning_rate": 9.614512471655328e-07, "loss": -0.0022, "num_tokens": 3727969.0, "reward": 0.07421875, "reward_std": 0.29723551869392395, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.24397502839565277, "rewards/format_reward_func/mean": 0.01171875, "rewards/format_reward_func/std": 0.05326050892472267, "step": 18, "step_time": 581.9889324549586 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3752.0, "completions/mean_length": 1173.984375, "completions/mean_terminated_length": 867.0696716308594, "completions/min_length": 2.5, "completions/min_terminated_length": 2.5, "entropy": 3.8836028147488832, "epoch": 0.04308390022675737, "frac_reward_zero_std": 0.84375, "grad_norm": 0.2177734375, "learning_rate": 9.591836734693876e-07, "loss": 0.0355, "num_tokens": 3903999.0, "reward": 0.08984375, "reward_std": 0.37648390233516693, "rewards/accuracy_reward_func/mean": 0.078125, "rewards/accuracy_reward_func/std": 0.36897502839565277, "rewards/format_reward_func/mean": 0.01171875, "rewards/format_reward_func/std": 0.07509202510118484, "step": 19, "step_time": 659.3856618851423 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1484375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3843.0, "completions/mean_length": 1374.8125, "completions/mean_terminated_length": 902.1538391113281, "completions/min_length": 42.5, "completions/min_terminated_length": 42.5, "entropy": 3.2819528337568045, "epoch": 0.045351473922902494, "frac_reward_zero_std": 0.875, "grad_norm": 0.1669921875, "learning_rate": 9.569160997732426e-07, "loss": 0.0059, "num_tokens": 4103619.0, "reward": 0.0625, "reward_std": 0.3096012622117996, "rewards/accuracy_reward_func/mean": 0.046875, "rewards/accuracy_reward_func/std": 0.3003681004047394, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08451050892472267, "step": 20, "step_time": 586.1947742849588 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.203125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4012.5, "completions/mean_length": 1617.8203125, "completions/mean_terminated_length": 976.201416015625, "completions/min_length": 3.5, "completions/min_terminated_length": 3.5, "entropy": 3.976168217137456, "epoch": 0.047619047619047616, "frac_reward_zero_std": 0.90625, "grad_norm": 0.123046875, "learning_rate": 9.546485260770974e-07, "loss": -0.0038, "num_tokens": 4335104.0, "reward": 0.03515625, "reward_std": 0.2533649355173111, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.00390625, "rewards/format_reward_func/std": 0.03125, "step": 21, "step_time": 613.1841935859993 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1796875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3851.5, "completions/mean_length": 1488.375, "completions/mean_terminated_length": 926.713623046875, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "entropy": 3.7107547111809254, "epoch": 0.049886621315192746, "frac_reward_zero_std": 0.96875, "grad_norm": 0.1298828125, "learning_rate": 9.523809523809522e-07, "loss": 0.0027, "num_tokens": 4555572.0, "reward": 0.0078125, "reward_std": 0.043842025101184845, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.043842025101184845, "step": 22, "step_time": 658.2109148548916 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1015625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4045.5, "completions/mean_length": 1292.171875, "completions/mean_terminated_length": 973.3143920898438, "completions/min_length": 36.0, "completions/min_terminated_length": 36.0, "entropy": 3.6732373889535666, "epoch": 0.05215419501133787, "frac_reward_zero_std": 0.875, "grad_norm": 0.22265625, "learning_rate": 9.501133786848072e-07, "loss": 0.0104, "num_tokens": 4745954.0, "reward": 0.05859375, "reward_std": 0.3089591860771179, "rewards/accuracy_reward_func/mean": 0.046875, "rewards/accuracy_reward_func/std": 0.3003681004047394, "rewards/format_reward_func/mean": 0.01171875, "rewards/format_reward_func/std": 0.07509202510118484, "step": 23, "step_time": 602.3255676180124 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1328125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3506.0, "completions/mean_length": 1260.6484375, "completions/mean_terminated_length": 830.3445129394531, "completions/min_length": 8.5, "completions/min_terminated_length": 8.5, "entropy": 3.2933055758476257, "epoch": 0.05442176870748299, "frac_reward_zero_std": 0.875, "grad_norm": 0.2314453125, "learning_rate": 9.47845804988662e-07, "loss": 0.0031, "num_tokens": 4931917.0, "reward": 0.06640625, "reward_std": 0.35280121862888336, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.00390625, "rewards/format_reward_func/std": 0.03125, "step": 24, "step_time": 598.7310787206516 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3860.5, "completions/mean_length": 1558.296875, "completions/mean_terminated_length": 1142.5206909179688, "completions/min_length": 111.5, "completions/min_terminated_length": 111.5, "entropy": 4.41904532071203, "epoch": 0.05668934240362812, "frac_reward_zero_std": 0.875, "grad_norm": 0.1923828125, "learning_rate": 9.45578231292517e-07, "loss": 0.0251, "num_tokens": 5156903.0, "reward": 0.08203125, "reward_std": 0.3898242563009262, "rewards/accuracy_reward_func/mean": 0.078125, "rewards/accuracy_reward_func/std": 0.38841013610363007, "rewards/format_reward_func/mean": 0.00390625, "rewards/format_reward_func/std": 0.03125, "step": 25, "step_time": 612.8359751384705 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3834.0, "completions/mean_length": 1414.25, "completions/mean_terminated_length": 863.0014343261719, "completions/min_length": 49.5, "completions/min_terminated_length": 49.5, "entropy": 3.6438450161367655, "epoch": 0.05895691609977324, "frac_reward_zero_std": 0.90625, "grad_norm": 0.1162109375, "learning_rate": 9.433106575963718e-07, "loss": -0.003, "num_tokens": 5361427.0, "reward": 0.05078125, "reward_std": 0.29202835261821747, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.01953125, "rewards/format_reward_func/std": 0.09224375709891319, "step": 26, "step_time": 586.6363095985726 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3840.5, "completions/mean_length": 1228.1484375, "completions/mean_terminated_length": 820.9175415039062, "completions/min_length": 110.0, "completions/min_terminated_length": 110.0, "entropy": 2.9272163938730955, "epoch": 0.061224489795918366, "frac_reward_zero_std": 0.875, "grad_norm": 0.1787109375, "learning_rate": 9.410430839002267e-07, "loss": -0.0145, "num_tokens": 5541462.0, "reward": 0.09765625, "reward_std": 0.3758723586797714, "rewards/accuracy_reward_func/mean": 0.078125, "rewards/accuracy_reward_func/std": 0.36897502839565277, "rewards/format_reward_func/mean": 0.01953125, "rewards/format_reward_func/std": 0.09224375709891319, "step": 27, "step_time": 588.392117716372 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3761.0, "completions/mean_length": 1163.484375, "completions/mean_terminated_length": 861.0517272949219, "completions/min_length": 11.5, "completions/min_terminated_length": 11.5, "entropy": 3.6875854432582855, "epoch": 0.06349206349206349, "frac_reward_zero_std": 0.84375, "grad_norm": 0.27734375, "learning_rate": 9.387755102040816e-07, "loss": -0.0093, "num_tokens": 5713684.0, "reward": 0.1171875, "reward_std": 0.4582659602165222, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.41934312880039215, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 28, "step_time": 591.353732842952 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1796875, "completions/max_length": 4096.0, "completions/max_terminated_length": 2511.5, "completions/mean_length": 1238.9765625, "completions/mean_terminated_length": 616.4163208007812, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 2.41994022205472, "epoch": 0.06575963718820861, "frac_reward_zero_std": 0.8125, "grad_norm": 0.3046875, "learning_rate": 9.365079365079365e-07, "loss": -0.0128, "num_tokens": 5899889.0, "reward": 0.1015625, "reward_std": 0.3987704664468765, "rewards/accuracy_reward_func/mean": 0.078125, "rewards/accuracy_reward_func/std": 0.38841013610363007, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10483578220009804, "step": 29, "step_time": 641.2962455619127 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3692.5, "completions/mean_length": 1316.7890625, "completions/mean_terminated_length": 863.9240417480469, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "entropy": 2.6513823810964823, "epoch": 0.06802721088435375, "frac_reward_zero_std": 0.84375, "grad_norm": 0.1982421875, "learning_rate": 9.342403628117913e-07, "loss": 0.0257, "num_tokens": 6092946.0, "reward": 0.0703125, "reward_std": 0.34078919887542725, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.3380420356988907, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.043842025101184845, "step": 30, "step_time": 599.0408421233296 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4026.5, "completions/mean_length": 1242.1640625, "completions/mean_terminated_length": 946.5238647460938, "completions/min_length": 22.5, "completions/min_terminated_length": 22.5, "entropy": 3.833143075928092, "epoch": 0.07029478458049887, "frac_reward_zero_std": 0.96875, "grad_norm": 0.09423828125, "learning_rate": 9.319727891156463e-07, "loss": -0.0087, "num_tokens": 6277247.0, "reward": 0.01171875, "reward_std": 0.05326050519943237, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.01171875, "rewards/format_reward_func/std": 0.05326050892472267, "step": 31, "step_time": 598.7559008421376 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3974.0, "completions/mean_length": 1660.1171875, "completions/mean_terminated_length": 1095.6091613769531, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 3.9972437638789415, "epoch": 0.07256235827664399, "frac_reward_zero_std": 0.90625, "grad_norm": 0.1767578125, "learning_rate": 9.29705215419501e-07, "loss": -0.0009, "num_tokens": 6514618.0, "reward": 0.0625, "reward_std": 0.3126605153083801, "rewards/accuracy_reward_func/mean": 0.046875, "rewards/accuracy_reward_func/std": 0.3003681004047394, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.06099375709891319, "step": 32, "step_time": 588.540596883744 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3492.0, "completions/mean_length": 1256.1171875, "completions/mean_terminated_length": 794.4375, "completions/min_length": 15.0, "completions/min_terminated_length": 15.0, "entropy": 3.574168335646391, "epoch": 0.07482993197278912, "frac_reward_zero_std": 0.84375, "grad_norm": 0.1923828125, "learning_rate": 9.274376417233559e-07, "loss": 0.0144, "num_tokens": 6698337.0, "reward": 0.07421875, "reward_std": 0.3566044270992279, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.01171875, "rewards/format_reward_func/std": 0.05326050892472267, "step": 33, "step_time": 590.0235936194658 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0859375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3585.0, "completions/mean_length": 1064.6640625, "completions/mean_terminated_length": 780.8223876953125, "completions/min_length": 3.5, "completions/min_terminated_length": 3.5, "entropy": 3.5261701587587595, "epoch": 0.07709750566893424, "frac_reward_zero_std": 0.78125, "grad_norm": 0.271484375, "learning_rate": 9.251700680272108e-07, "loss": 0.0511, "num_tokens": 6857330.0, "reward": 0.12109375, "reward_std": 0.4603298455476761, "rewards/accuracy_reward_func/mean": 0.109375, "rewards/accuracy_reward_func/std": 0.45701706409454346, "rewards/format_reward_func/mean": 0.01171875, "rewards/format_reward_func/std": 0.07509202510118484, "step": 34, "step_time": 582.2729406254366 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3875.0, "completions/mean_length": 1183.78125, "completions/mean_terminated_length": 766.4074096679688, "completions/min_length": 22.0, "completions/min_terminated_length": 22.0, "entropy": 3.6374483965337276, "epoch": 0.07936507936507936, "frac_reward_zero_std": 0.75, "grad_norm": 0.2451171875, "learning_rate": 9.229024943310657e-07, "loss": -0.0024, "num_tokens": 7033458.0, "reward": 0.1328125, "reward_std": 0.45352521538734436, "rewards/accuracy_reward_func/mean": 0.109375, "rewards/accuracy_reward_func/std": 0.44585782289505005, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 35, "step_time": 604.7557975919917 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1484375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3507.5, "completions/mean_length": 1338.28125, "completions/mean_terminated_length": 857.5964660644531, "completions/min_length": 20.0, "completions/min_terminated_length": 20.0, "entropy": 3.4108287915587425, "epoch": 0.08163265306122448, "frac_reward_zero_std": 0.875, "grad_norm": 0.2353515625, "learning_rate": 9.206349206349205e-07, "loss": -0.0284, "num_tokens": 7227406.0, "reward": 0.0625, "reward_std": 0.33031322062015533, "rewards/accuracy_reward_func/mean": 0.046875, "rewards/accuracy_reward_func/std": 0.3003681004047394, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08451050892472267, "step": 36, "step_time": 582.625431638211 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4073.5, "completions/mean_length": 1282.84375, "completions/mean_terminated_length": 994.6044311523438, "completions/min_length": 6.5, "completions/min_terminated_length": 6.5, "entropy": 4.335059890523553, "epoch": 0.08390022675736962, "frac_reward_zero_std": 0.875, "grad_norm": 0.21875, "learning_rate": 9.183673469387755e-07, "loss": 0.0204, "num_tokens": 7416650.0, "reward": 0.05859375, "reward_std": 0.21704265475273132, "rewards/accuracy_reward_func/mean": 0.046875, "rewards/accuracy_reward_func/std": 0.21304203569889069, "rewards/format_reward_func/mean": 0.01171875, "rewards/format_reward_func/std": 0.05326050892472267, "step": 37, "step_time": 597.565967454575 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3871.0, "completions/mean_length": 1476.2109375, "completions/mean_terminated_length": 932.4811401367188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.803565353155136, "epoch": 0.08616780045351474, "frac_reward_zero_std": 0.9375, "grad_norm": 0.08837890625, "learning_rate": 9.160997732426303e-07, "loss": -0.0007, "num_tokens": 7629049.0, "reward": 0.0625, "reward_std": 0.24397501349449158, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.24397502839565277, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 38, "step_time": 583.4348162943497 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1640625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3728.0, "completions/mean_length": 1381.0390625, "completions/mean_terminated_length": 849.9523315429688, "completions/min_length": 5.5, "completions/min_terminated_length": 5.5, "entropy": 3.3647786304354668, "epoch": 0.08843537414965986, "frac_reward_zero_std": 0.8125, "grad_norm": 0.1923828125, "learning_rate": 9.138321995464853e-07, "loss": 0.0154, "num_tokens": 7831390.0, "reward": 0.10546875, "reward_std": 0.41434262692928314, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.39548972249031067, "rewards/format_reward_func/mean": 0.01171875, "rewards/format_reward_func/std": 0.07509202510118484, "step": 39, "step_time": 599.8489952124655 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 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": 4096.0, "completions/max_terminated_length": 3683.5, "completions/mean_length": 1284.046875, "completions/mean_terminated_length": 935.6366577148438, "completions/min_length": 20.5, "completions/min_terminated_length": 20.5, "entropy": 3.913011803291738, "epoch": 0.09070294784580499, "frac_reward_zero_std": 0.75, "grad_norm": 0.28515625, "learning_rate": 9.115646258503401e-07, "loss": -0.0161, "num_tokens": 8018060.0, "reward": 0.1484375, "reward_std": 0.4921819716691971, "rewards/accuracy_reward_func/mean": 0.140625, "rewards/accuracy_reward_func/std": 0.48994454741477966, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 40, "step_time": 583.9517429694533 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 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": 4096.0, "completions/max_terminated_length": 3488.0, "completions/mean_length": 1136.1796875, "completions/mean_terminated_length": 772.6929931640625, "completions/min_length": 5.5, "completions/min_terminated_length": 5.5, "entropy": 3.530646240338683, "epoch": 0.09297052154195011, "frac_reward_zero_std": 0.8125, "grad_norm": 0.2373046875, "learning_rate": 9.09297052154195e-07, "loss": 0.0425, "num_tokens": 8187987.0, "reward": 0.1015625, "reward_std": 0.42883121967315674, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.043842025101184845, "step": 41, "step_time": 586.9831876922399 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1015625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3883.0, "completions/mean_length": 1217.671875, "completions/mean_terminated_length": 890.7786254882812, "completions/min_length": 28.0, "completions/min_terminated_length": 28.0, "entropy": 3.6075865300372243, "epoch": 0.09523809523809523, "frac_reward_zero_std": 0.875, "grad_norm": 0.1689453125, "learning_rate": 9.070294784580499e-07, "loss": 0.004, "num_tokens": 8367045.0, "reward": 0.078125, "reward_std": 0.3689750134944916, "rewards/accuracy_reward_func/mean": 0.078125, "rewards/accuracy_reward_func/std": 0.36897502839565277, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 42, "step_time": 593.1046754224226 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3453.5, "completions/mean_length": 1357.359375, "completions/mean_terminated_length": 928.1923522949219, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.721042402088642, "epoch": 0.09750566893424037, "frac_reward_zero_std": 0.8125, "grad_norm": 0.291015625, "learning_rate": 9.047619047619047e-07, "loss": -0.0005, "num_tokens": 8565287.0, "reward": 0.0859375, "reward_std": 0.29998381435871124, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.24397502839565277, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 43, "step_time": 599.4938222169876 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3986.5, "completions/mean_length": 1526.4140625, "completions/mean_terminated_length": 998.0187683105469, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 4.3327610064297915, "epoch": 0.09977324263038549, "frac_reward_zero_std": 0.9375, "grad_norm": 0.1328125, "learning_rate": 9.024943310657595e-07, "loss": 0.0407, "num_tokens": 8784856.0, "reward": 0.0234375, "reward_std": 0.13152606785297394, "rewards/accuracy_reward_func/mean": 0.015625, "rewards/accuracy_reward_func/std": 0.125, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.043842025101184845, "step": 44, "step_time": 601.5610232567415 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 2995.0, "completions/mean_length": 1162.65625, "completions/mean_terminated_length": 682.6545104980469, "completions/min_length": 3.5, "completions/min_terminated_length": 3.5, "entropy": 2.599572140723467, "epoch": 0.10204081632653061, "frac_reward_zero_std": 0.9375, "grad_norm": 0.13671875, "learning_rate": 9.002267573696145e-07, "loss": -0.003, "num_tokens": 8957812.0, "reward": 0.01171875, "reward_std": 0.07509202510118484, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.01171875, "rewards/format_reward_func/std": 0.07509202510118484, "step": 45, "step_time": 599.2623494956642 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1640625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3193.5, "completions/mean_length": 1333.75, "completions/mean_terminated_length": 775.1260681152344, "completions/min_length": 4.5, "completions/min_terminated_length": 4.5, "entropy": 2.9866717774420977, "epoch": 0.10430839002267574, "frac_reward_zero_std": 0.90625, "grad_norm": 0.193359375, "learning_rate": 8.979591836734693e-07, "loss": 0.037, "num_tokens": 9151420.0, "reward": 0.0234375, "reward_std": 0.1596149355173111, "rewards/accuracy_reward_func/mean": 0.015625, "rewards/accuracy_reward_func/std": 0.125, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 46, "step_time": 584.1288063684478 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1328125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3997.5, "completions/mean_length": 1418.0390625, "completions/mean_terminated_length": 1004.8611145019531, "completions/min_length": 48.0, "completions/min_terminated_length": 48.0, "entropy": 3.4238973315805197, "epoch": 0.10657596371882086, "frac_reward_zero_std": 0.96875, "grad_norm": 0.130859375, "learning_rate": 8.956916099773242e-07, "loss": 0.0135, "num_tokens": 9355801.0, "reward": 0.00390625, "reward_std": 0.03125, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.00390625, "rewards/format_reward_func/std": 0.03125, "step": 47, "step_time": 585.5585831161588 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1484375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3749.5, "completions/mean_length": 1428.3359375, "completions/mean_terminated_length": 964.203369140625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.784708108752966, "epoch": 0.10884353741496598, "frac_reward_zero_std": 0.84375, "grad_norm": 0.2041015625, "learning_rate": 8.934240362811791e-07, "loss": 0.0216, "num_tokens": 9563164.0, "reward": 0.08984375, "reward_std": 0.37423862516880035, "rewards/accuracy_reward_func/mean": 0.078125, "rewards/accuracy_reward_func/std": 0.36897502839565277, "rewards/format_reward_func/mean": 0.01171875, "rewards/format_reward_func/std": 0.07509202510118484, "step": 48, "step_time": 588.5648364648223 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1640625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3933.5, "completions/mean_length": 1392.5078125, "completions/mean_terminated_length": 861.5879211425781, "completions/min_length": 56.5, "completions/min_terminated_length": 56.5, "entropy": 3.1068236976861954, "epoch": 0.1111111111111111, "frac_reward_zero_std": 0.90625, "grad_norm": 0.166015625, "learning_rate": 8.91156462585034e-07, "loss": 0.0261, "num_tokens": 9764893.0, "reward": 0.05078125, "reward_std": 0.29285234212875366, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.01953125, "rewards/format_reward_func/std": 0.09710253402590752, "step": 49, "step_time": 612.5280408170074 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1484375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3839.0, "completions/mean_length": 1314.390625, "completions/mean_terminated_length": 831.5285034179688, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 3.5168911181390285, "epoch": 0.11337868480725624, "frac_reward_zero_std": 0.90625, "grad_norm": 0.1533203125, "learning_rate": 8.888888888888888e-07, "loss": -0.0155, "num_tokens": 9956947.0, "reward": 0.04296875, "reward_std": 0.22127514332532883, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.17536810040473938, "rewards/format_reward_func/mean": 0.01171875, "rewards/format_reward_func/std": 0.07509202510118484, "step": 50, "step_time": 587.9333986295387 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3882.0, "completions/mean_length": 1565.3359375, "completions/mean_terminated_length": 976.5892639160156, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 4.300196705386043, "epoch": 0.11564625850340136, "frac_reward_zero_std": 0.90625, "grad_norm": 0.1318359375, "learning_rate": 8.866213151927438e-07, "loss": 0.0146, "num_tokens": 10180654.0, "reward": 0.0234375, "reward_std": 0.1596149355173111, "rewards/accuracy_reward_func/mean": 0.015625, "rewards/accuracy_reward_func/std": 0.125, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 51, "step_time": 596.8772314703092 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 4001.5, "completions/mean_length": 1331.4453125, "completions/mean_terminated_length": 965.4701843261719, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.5988041888922453, "epoch": 0.11791383219954649, "frac_reward_zero_std": 0.875, "grad_norm": 0.2001953125, "learning_rate": 8.843537414965985e-07, "loss": 0.0128, "num_tokens": 10374567.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 52, "step_time": 598.7554965950549 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3918.0, "completions/mean_length": 1333.5234375, "completions/mean_terminated_length": 964.8404235839844, "completions/min_length": 4.5, "completions/min_terminated_length": 4.5, "entropy": 3.7250664979219437, "epoch": 0.12018140589569161, "frac_reward_zero_std": 0.9375, "grad_norm": 0.1201171875, "learning_rate": 8.820861678004535e-07, "loss": 0.0008, "num_tokens": 10567310.0, "reward": 0.0234375, "reward_std": 0.16884202510118484, "rewards/accuracy_reward_func/mean": 0.015625, "rewards/accuracy_reward_func/std": 0.125, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.043842025101184845, "step": 53, "step_time": 580.3850104399025 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1015625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3875.5, "completions/mean_length": 1238.6484375, "completions/mean_terminated_length": 915.7840270996094, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.7390247844159603, "epoch": 0.12244897959183673, "frac_reward_zero_std": 0.84375, "grad_norm": 0.244140625, "learning_rate": 8.798185941043083e-07, "loss": 0.0408, "num_tokens": 10751985.0, "reward": 0.05859375, "reward_std": 0.3089591860771179, "rewards/accuracy_reward_func/mean": 0.046875, "rewards/accuracy_reward_func/std": 0.3003681004047394, "rewards/format_reward_func/mean": 0.01171875, "rewards/format_reward_func/std": 0.07509202510118484, "step": 54, "step_time": 602.8104365291074 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3832.5, "completions/mean_length": 1432.03125, "completions/mean_terminated_length": 1014.6162109375, "completions/min_length": 55.5, "completions/min_terminated_length": 55.5, "entropy": 4.180685954168439, "epoch": 0.12471655328798185, "frac_reward_zero_std": 0.9375, "grad_norm": 0.1357421875, "learning_rate": 8.775510204081632e-07, "loss": 0.0098, "num_tokens": 10958601.0, "reward": 0.03515625, "reward_std": 0.20661810040473938, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.17536810040473938, "rewards/format_reward_func/mean": 0.00390625, "rewards/format_reward_func/std": 0.03125, "step": 55, "step_time": 589.0944160483778 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 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": 4096.0, "completions/max_terminated_length": 3661.0, "completions/mean_length": 1241.0234375, "completions/mean_terminated_length": 890.4122924804688, "completions/min_length": 31.5, "completions/min_terminated_length": 31.5, "entropy": 3.475444095209241, "epoch": 0.12698412698412698, "frac_reward_zero_std": 0.8125, "grad_norm": 0.294921875, "learning_rate": 8.752834467120181e-07, "loss": 0.0468, "num_tokens": 11141100.0, "reward": 0.1171875, "reward_std": 0.4067866802215576, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.39548972249031067, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 56, "step_time": 595.941380049102 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1640625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3977.0, "completions/mean_length": 1423.265625, "completions/mean_terminated_length": 904.7202758789062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.743929397314787, "epoch": 0.1292517006802721, "frac_reward_zero_std": 0.875, "grad_norm": 0.177734375, "learning_rate": 8.73015873015873e-07, "loss": 0.006, "num_tokens": 11348246.0, "reward": 0.078125, "reward_std": 0.37932415306568146, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.3380420356988907, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08451050892472267, "step": 57, "step_time": 621.8577571306378 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.203125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3588.5, "completions/mean_length": 1568.6953125, "completions/mean_terminated_length": 924.08642578125, "completions/min_length": 69.0, "completions/min_terminated_length": 69.0, "entropy": 3.748473282903433, "epoch": 0.13151927437641722, "frac_reward_zero_std": 0.84375, "grad_norm": 0.1962890625, "learning_rate": 8.707482993197278e-07, "loss": -0.0042, "num_tokens": 11571575.0, "reward": 0.07421875, "reward_std": 0.3420426547527313, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.3380420356988907, "rewards/format_reward_func/mean": 0.01171875, "rewards/format_reward_func/std": 0.05326050892472267, "step": 58, "step_time": 584.199871532619 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1640625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3920.5, "completions/mean_length": 1500.640625, "completions/mean_terminated_length": 999.1975402832031, "completions/min_length": 3.5, "completions/min_terminated_length": 3.5, "entropy": 4.530334532260895, "epoch": 0.13378684807256236, "frac_reward_zero_std": 0.90625, "grad_norm": 0.1689453125, "learning_rate": 8.684807256235828e-07, "loss": 0.0044, "num_tokens": 11791385.0, "reward": 0.0234375, "reward_std": 0.1596149355173111, "rewards/accuracy_reward_func/mean": 0.015625, "rewards/accuracy_reward_func/std": 0.125, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 59, "step_time": 609.9015548406169 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1640625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3949.5, "completions/mean_length": 1356.453125, "completions/mean_terminated_length": 818.5648498535156, "completions/min_length": 2.5, "completions/min_terminated_length": 2.5, "entropy": 3.20889993570745, "epoch": 0.1360544217687075, "frac_reward_zero_std": 0.90625, "grad_norm": 0.19921875, "learning_rate": 8.662131519274376e-07, "loss": -0.0039, "num_tokens": 11987719.0, "reward": 0.03515625, "reward_std": 0.17743311822414398, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.17536810040473938, "rewards/format_reward_func/mean": 0.00390625, "rewards/format_reward_func/std": 0.03125, "step": 60, "step_time": 585.9198675043881 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3567.0, "completions/mean_length": 1115.546875, "completions/mean_terminated_length": 859.4793395996094, "completions/min_length": 3.5, "completions/min_terminated_length": 3.5, "entropy": 3.9494722429662943, "epoch": 0.1383219954648526, "frac_reward_zero_std": 0.8125, "grad_norm": 0.251953125, "learning_rate": 8.639455782312925e-07, "loss": 0.0761, "num_tokens": 12153553.0, "reward": 0.09765625, "reward_std": 0.3964306712150574, "rewards/accuracy_reward_func/mean": 0.078125, "rewards/accuracy_reward_func/std": 0.38841013610363007, "rewards/format_reward_func/mean": 0.01953125, "rewards/format_reward_func/std": 0.09710253402590752, "step": 61, "step_time": 588.7423055358231 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3860.0, "completions/mean_length": 1337.1328125, "completions/mean_terminated_length": 885.6817932128906, "completions/min_length": 5.5, "completions/min_terminated_length": 5.5, "entropy": 2.918829394504428, "epoch": 0.14058956916099774, "frac_reward_zero_std": 0.84375, "grad_norm": 0.208984375, "learning_rate": 8.616780045351474e-07, "loss": 0.0125, "num_tokens": 12351434.0, "reward": 0.09375, "reward_std": 0.4053945690393448, "rewards/accuracy_reward_func/mean": 0.078125, "rewards/accuracy_reward_func/std": 0.36897502839565277, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08451050892472267, "step": 62, "step_time": 633.4415109893307 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.2265625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3928.5, "completions/mean_length": 1692.1640625, "completions/mean_terminated_length": 987.1826171875, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 3.8325051330029964, "epoch": 0.14285714285714285, "frac_reward_zero_std": 0.875, "grad_norm": 0.14453125, "learning_rate": 8.594104308390022e-07, "loss": 0.0149, "num_tokens": 12592403.0, "reward": 0.0546875, "reward_std": 0.3068941682577133, "rewards/accuracy_reward_func/mean": 0.046875, "rewards/accuracy_reward_func/std": 0.3003681004047394, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.043842025101184845, "step": 63, "step_time": 590.275172451511 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 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": 4096.0, "completions/max_terminated_length": 3880.5, "completions/mean_length": 1145.0546875, "completions/mean_terminated_length": 781.627197265625, "completions/min_length": 62.5, "completions/min_terminated_length": 62.5, "entropy": 3.939505912363529, "epoch": 0.14512471655328799, "frac_reward_zero_std": 0.8125, "grad_norm": 0.2490234375, "learning_rate": 8.57142857142857e-07, "loss": 0.0198, "num_tokens": 12763162.0, "reward": 0.078125, "reward_std": 0.3489539474248886, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.3380420356988907, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08451050892472267, "step": 64, "step_time": 601.2852159729227 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3328.5, "completions/mean_length": 1344.3671875, "completions/mean_terminated_length": 894.0999755859375, "completions/min_length": 5.5, "completions/min_terminated_length": 5.5, "entropy": 3.6352404467761517, "epoch": 0.1473922902494331, "frac_reward_zero_std": 0.84375, "grad_norm": 0.2333984375, "learning_rate": 8.54875283446712e-07, "loss": 0.0207, "num_tokens": 12958453.0, "reward": 0.06640625, "reward_std": 0.244957834482193, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.24397502839565277, "rewards/format_reward_func/mean": 0.00390625, "rewards/format_reward_func/std": 0.03125, "step": 65, "step_time": 604.1546341096982 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3960.5, "completions/mean_length": 1339.8203125, "completions/mean_terminated_length": 885.8468322753906, "completions/min_length": 43.0, "completions/min_terminated_length": 43.0, "entropy": 3.2994116097688675, "epoch": 0.14965986394557823, "frac_reward_zero_std": 0.90625, "grad_norm": 0.158203125, "learning_rate": 8.526077097505668e-07, "loss": -0.0314, "num_tokens": 13153994.0, "reward": 0.05859375, "reward_std": 0.30986589193344116, "rewards/accuracy_reward_func/mean": 0.046875, "rewards/accuracy_reward_func/std": 0.3003681004047394, "rewards/format_reward_func/mean": 0.01171875, "rewards/format_reward_func/std": 0.05326050892472267, "step": 66, "step_time": 586.3202119162306 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3762.5, "completions/mean_length": 1159.3671875, "completions/mean_terminated_length": 913.3544616699219, "completions/min_length": 22.0, "completions/min_terminated_length": 22.0, "entropy": 4.719866678118706, "epoch": 0.15192743764172337, "frac_reward_zero_std": 0.8125, "grad_norm": 0.322265625, "learning_rate": 8.503401360544217e-07, "loss": 0.0089, "num_tokens": 13327053.0, "reward": 0.07421875, "reward_std": 0.34598222374916077, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.3380420356988907, "rewards/format_reward_func/mean": 0.01171875, "rewards/format_reward_func/std": 0.07509202510118484, "step": 67, "step_time": 597.3318418385461 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3997.0, "completions/mean_length": 1457.734375, "completions/mean_terminated_length": 1028.190673828125, "completions/min_length": 3.5, "completions/min_terminated_length": 3.5, "entropy": 4.309329951182008, "epoch": 0.15419501133786848, "frac_reward_zero_std": 0.90625, "grad_norm": 0.140625, "learning_rate": 8.480725623582766e-07, "loss": 0.0148, "num_tokens": 13537987.0, "reward": 0.08203125, "reward_std": 0.2711479365825653, "rewards/accuracy_reward_func/mean": 0.078125, "rewards/accuracy_reward_func/std": 0.27048972249031067, "rewards/format_reward_func/mean": 0.00390625, "rewards/format_reward_func/std": 0.03125, "step": 68, "step_time": 595.651706719771 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3833.0, "completions/mean_length": 1283.3984375, "completions/mean_terminated_length": 912.62158203125, "completions/min_length": 64.0, "completions/min_terminated_length": 64.0, "entropy": 3.7792484648525715, "epoch": 0.1564625850340136, "frac_reward_zero_std": 0.75, "grad_norm": 0.294921875, "learning_rate": 8.458049886621315e-07, "loss": 0.0182, "num_tokens": 13725062.0, "reward": 0.12109375, "reward_std": 0.46074703335762024, "rewards/accuracy_reward_func/mean": 0.109375, "rewards/accuracy_reward_func/std": 0.45701706409454346, "rewards/format_reward_func/mean": 0.01171875, "rewards/format_reward_func/std": 0.07509202510118484, "step": 69, "step_time": 585.8399355420843 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1328125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3663.0, "completions/mean_length": 1348.53125, "completions/mean_terminated_length": 929.372802734375, "completions/min_length": 155.5, "completions/min_terminated_length": 155.5, "entropy": 3.0831449050456285, "epoch": 0.15873015873015872, "frac_reward_zero_std": 0.875, "grad_norm": 0.18359375, "learning_rate": 8.435374149659864e-07, "loss": 0.0371, "num_tokens": 13921226.0, "reward": 0.1015625, "reward_std": 0.3995128720998764, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.39548972249031067, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 70, "step_time": 592.1932232938707 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3615.0, "completions/mean_length": 1510.4453125, "completions/mean_terminated_length": 907.9366760253906, "completions/min_length": 9.5, "completions/min_terminated_length": 9.5, "entropy": 3.17475145123899, "epoch": 0.16099773242630386, "frac_reward_zero_std": 0.84375, "grad_norm": 0.1669921875, "learning_rate": 8.412698412698413e-07, "loss": 0.0128, "num_tokens": 14138339.0, "reward": 0.0625, "reward_std": 0.3109140843153, "rewards/accuracy_reward_func/mean": 0.046875, "rewards/accuracy_reward_func/std": 0.3003681004047394, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 71, "step_time": 593.9878287529573 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.2109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3995.5, "completions/mean_length": 1693.75, "completions/mean_terminated_length": 1057.754150390625, "completions/min_length": 91.0, "completions/min_terminated_length": 91.0, "entropy": 4.4856348149478436, "epoch": 0.16326530612244897, "frac_reward_zero_std": 0.875, "grad_norm": 0.134765625, "learning_rate": 8.39002267573696e-07, "loss": 0.0219, "num_tokens": 14379223.0, "reward": 0.078125, "reward_std": 0.3471105545759201, "rewards/accuracy_reward_func/mean": 0.046875, "rewards/accuracy_reward_func/std": 0.3003681004047394, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12088293954730034, "step": 72, "step_time": 588.8920545289293 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1953125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3787.0, "completions/mean_length": 1603.9140625, "completions/mean_terminated_length": 1001.0800170898438, "completions/min_length": 4.5, "completions/min_terminated_length": 4.5, "entropy": 4.481541287153959, "epoch": 0.1655328798185941, "frac_reward_zero_std": 0.875, "grad_norm": 0.1279296875, "learning_rate": 8.36734693877551e-07, "loss": -0.0036, "num_tokens": 14610504.0, "reward": 0.0625, "reward_std": 0.33031322062015533, "rewards/accuracy_reward_func/mean": 0.046875, "rewards/accuracy_reward_func/std": 0.3003681004047394, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08451050892472267, "step": 73, "step_time": 599.5370882023126 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3574.5, "completions/mean_length": 1242.8515625, "completions/mean_terminated_length": 779.6492919921875, "completions/min_length": 16.0, "completions/min_terminated_length": 16.0, "entropy": 3.465228695422411, "epoch": 0.16780045351473924, "frac_reward_zero_std": 0.84375, "grad_norm": 0.263671875, "learning_rate": 8.344671201814058e-07, "loss": 0.0705, "num_tokens": 14794581.0, "reward": 0.11328125, "reward_std": 0.4011373072862625, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.39548972249031067, "rewards/format_reward_func/mean": 0.01953125, "rewards/format_reward_func/std": 0.09224375709891319, "step": 74, "step_time": 595.3202944584191 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1484375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3723.0, "completions/mean_length": 1412.9140625, "completions/mean_terminated_length": 945.0740661621094, "completions/min_length": 15.5, "completions/min_terminated_length": 15.5, "entropy": 3.5855388585478067, "epoch": 0.17006802721088435, "frac_reward_zero_std": 0.84375, "grad_norm": 0.2451171875, "learning_rate": 8.321995464852607e-07, "loss": 0.0482, "num_tokens": 14999906.0, "reward": 0.07421875, "reward_std": 0.35682113468647003, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.01171875, "rewards/format_reward_func/std": 0.07509202510118484, "step": 75, "step_time": 587.6844481090084 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1953125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3875.0, "completions/mean_length": 1507.4296875, "completions/mean_terminated_length": 878.4783325195312, "completions/min_length": 87.5, "completions/min_terminated_length": 87.5, "entropy": 4.3120437897741795, "epoch": 0.17233560090702948, "frac_reward_zero_std": 0.875, "grad_norm": 0.142578125, "learning_rate": 8.299319727891156e-07, "loss": 0.0204, "num_tokens": 15215661.0, "reward": 0.06640625, "reward_std": 0.33945615589618683, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.3380420356988907, "rewards/format_reward_func/mean": 0.00390625, "rewards/format_reward_func/std": 0.03125, "step": 76, "step_time": 583.3368950495496 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3274.5, "completions/mean_length": 1353.484375, "completions/mean_terminated_length": 897.9273986816406, "completions/min_length": 20.5, "completions/min_terminated_length": 20.5, "entropy": 3.536152880638838, "epoch": 0.1746031746031746, "frac_reward_zero_std": 0.84375, "grad_norm": 0.1748046875, "learning_rate": 8.276643990929705e-07, "loss": 0.0289, "num_tokens": 15416263.0, "reward": 0.109375, "reward_std": 0.4490707814693451, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08451050892472267, "step": 77, "step_time": 622.0613927459344 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3977.0, "completions/mean_length": 1316.15625, "completions/mean_terminated_length": 918.4803466796875, "completions/min_length": 127.0, "completions/min_terminated_length": 127.0, "entropy": 3.789307551458478, "epoch": 0.17687074829931973, "frac_reward_zero_std": 0.78125, "grad_norm": 0.2470703125, "learning_rate": 8.253968253968253e-07, "loss": -0.005, "num_tokens": 15608259.0, "reward": 0.11328125, "reward_std": 0.4465160369873047, "rewards/accuracy_reward_func/mean": 0.109375, "rewards/accuracy_reward_func/std": 0.44585782289505005, "rewards/format_reward_func/mean": 0.00390625, "rewards/format_reward_func/std": 0.03125, "step": 78, "step_time": 586.2861022176221 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3956.0, "completions/mean_length": 1310.3046875, "completions/mean_terminated_length": 854.2431030273438, "completions/min_length": 83.5, "completions/min_terminated_length": 83.5, "entropy": 3.574035180732608, "epoch": 0.17913832199546487, "frac_reward_zero_std": 0.875, "grad_norm": 0.1640625, "learning_rate": 8.231292517006803e-07, "loss": 0.012, "num_tokens": 15799730.0, "reward": 0.0859375, "reward_std": 0.3755010813474655, "rewards/accuracy_reward_func/mean": 0.078125, "rewards/accuracy_reward_func/std": 0.36897502839565277, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.043842025101184845, "step": 79, "step_time": 588.9511601096019 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3509.0, "completions/mean_length": 1193.9375, "completions/mean_terminated_length": 893.7241516113281, "completions/min_length": 46.0, "completions/min_terminated_length": 46.0, "entropy": 4.014947101473808, "epoch": 0.18140589569160998, "frac_reward_zero_std": 0.875, "grad_norm": 0.2578125, "learning_rate": 8.208616780045351e-07, "loss": -0.0602, "num_tokens": 15975726.0, "reward": 0.0390625, "reward_std": 0.20868311822414398, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.17536810040473938, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 80, "step_time": 585.2185197388753 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4020.0, "completions/mean_length": 1287.953125, "completions/mean_terminated_length": 887.0965881347656, "completions/min_length": 75.5, "completions/min_terminated_length": 75.5, "entropy": 2.9694155864417553, "epoch": 0.1836734693877551, "frac_reward_zero_std": 0.8125, "grad_norm": 0.26953125, "learning_rate": 8.1859410430839e-07, "loss": -0.0177, "num_tokens": 16163236.0, "reward": 0.09765625, "reward_std": 0.42140813171863556, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.41934312880039215, "rewards/format_reward_func/mean": 0.00390625, "rewards/format_reward_func/std": 0.03125, "step": 81, "step_time": 584.5412617325783 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.203125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3618.0, "completions/mean_length": 1509.203125, "completions/mean_terminated_length": 851.4188537597656, "completions/min_length": 62.5, "completions/min_terminated_length": 62.5, "entropy": 2.8872069008648396, "epoch": 0.18594104308390022, "frac_reward_zero_std": 0.9375, "grad_norm": 0.1337890625, "learning_rate": 8.163265306122449e-07, "loss": 0.0251, "num_tokens": 16380682.0, "reward": 0.0390625, "reward_std": 0.1811344400048256, "rewards/accuracy_reward_func/mean": 0.015625, "rewards/accuracy_reward_func/std": 0.125, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10483578220009804, "step": 82, "step_time": 594.9153352538124 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3969.0, "completions/mean_length": 1395.609375, "completions/mean_terminated_length": 895.5370483398438, "completions/min_length": 23.5, "completions/min_terminated_length": 23.5, "entropy": 4.2366653345525265, "epoch": 0.18820861678004536, "frac_reward_zero_std": 0.84375, "grad_norm": 0.1708984375, "learning_rate": 8.140589569160997e-07, "loss": 0.0103, "num_tokens": 16583804.0, "reward": 0.08984375, "reward_std": 0.30299752950668335, "rewards/accuracy_reward_func/mean": 0.078125, "rewards/accuracy_reward_func/std": 0.27048972249031067, "rewards/format_reward_func/mean": 0.01171875, "rewards/format_reward_func/std": 0.07509202510118484, "step": 83, "step_time": 593.6966224908829 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 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": 4096.0, "completions/max_terminated_length": 3577.0, "completions/mean_length": 1251.390625, "completions/mean_terminated_length": 902.0526428222656, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 3.0590660013258457, "epoch": 0.19047619047619047, "frac_reward_zero_std": 0.8125, "grad_norm": 0.244140625, "learning_rate": 8.117913832199546e-07, "loss": 0.0294, "num_tokens": 16766790.0, "reward": 0.0703125, "reward_std": 0.35486623644828796, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 84, "step_time": 580.8785895872861 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.2265625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3973.0, "completions/mean_length": 1558.0078125, "completions/mean_terminated_length": 810.9404907226562, "completions/min_length": 67.0, "completions/min_terminated_length": 67.0, "entropy": 2.7201994080096483, "epoch": 0.1927437641723356, "frac_reward_zero_std": 0.9375, "grad_norm": 0.10205078125, "learning_rate": 8.095238095238095e-07, "loss": 0.0027, "num_tokens": 16991107.0, "reward": 0.01953125, "reward_std": 0.15625, "rewards/accuracy_reward_func/mean": 0.015625, "rewards/accuracy_reward_func/std": 0.125, "rewards/format_reward_func/mean": 0.00390625, "rewards/format_reward_func/std": 0.03125, "step": 85, "step_time": 598.0091357100755 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1484375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3772.0, "completions/mean_length": 1500.359375, "completions/mean_terminated_length": 1039.2779846191406, "completions/min_length": 3.5, "completions/min_terminated_length": 3.5, "entropy": 4.1149397026747465, "epoch": 0.19501133786848074, "frac_reward_zero_std": 0.9375, "grad_norm": 0.1337890625, "learning_rate": 8.072562358276643e-07, "loss": 0.0186, "num_tokens": 17207357.0, "reward": 0.01953125, "reward_std": 0.15625, "rewards/accuracy_reward_func/mean": 0.015625, "rewards/accuracy_reward_func/std": 0.125, "rewards/format_reward_func/mean": 0.00390625, "rewards/format_reward_func/std": 0.03125, "step": 86, "step_time": 587.796465927735 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1484375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3744.0, "completions/mean_length": 1432.7265625, "completions/mean_terminated_length": 970.9487915039062, "completions/min_length": 99.5, "completions/min_terminated_length": 99.5, "entropy": 3.6395829413086176, "epoch": 0.19727891156462585, "frac_reward_zero_std": 0.875, "grad_norm": 0.185546875, "learning_rate": 8.049886621315193e-07, "loss": 0.0666, "num_tokens": 17414518.0, "reward": 0.06640625, "reward_std": 0.3414069563150406, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.3380420356988907, "rewards/format_reward_func/mean": 0.00390625, "rewards/format_reward_func/std": 0.03125, "step": 87, "step_time": 584.6671968176961 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 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": 4096.0, "completions/max_terminated_length": 3460.5, "completions/mean_length": 1004.0546875, "completions/mean_terminated_length": 800.0615234375, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "entropy": 2.9514008220285177, "epoch": 0.19954648526077098, "frac_reward_zero_std": 0.75, "grad_norm": 0.42578125, "learning_rate": 8.027210884353741e-07, "loss": 0.0102, "num_tokens": 17569533.0, "reward": 0.1640625, "reward_std": 0.5127404779195786, "rewards/accuracy_reward_func/mean": 0.140625, "rewards/accuracy_reward_func/std": 0.506826862692833, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10483578220009804, "step": 88, "step_time": 641.7255861982703 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3904.0, "completions/mean_length": 1224.0390625, "completions/mean_terminated_length": 815.5149536132812, "completions/min_length": 3.5, "completions/min_terminated_length": 3.5, "entropy": 3.424195848405361, "epoch": 0.2018140589569161, "frac_reward_zero_std": 0.8125, "grad_norm": 0.263671875, "learning_rate": 8.00453514739229e-07, "loss": 0.0173, "num_tokens": 17750510.0, "reward": 0.15234375, "reward_std": 0.5172130018472672, "rewards/accuracy_reward_func/mean": 0.140625, "rewards/accuracy_reward_func/std": 0.5144647508859634, "rewards/format_reward_func/mean": 0.01171875, "rewards/format_reward_func/std": 0.05326050892472267, "step": 89, "step_time": 610.0827826801687 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0703125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3426.0, "completions/mean_length": 1072.9375, "completions/mean_terminated_length": 843.7853393554688, "completions/min_length": 97.5, "completions/min_terminated_length": 97.5, "entropy": 3.6471029240638018, "epoch": 0.20408163265306123, "frac_reward_zero_std": 0.78125, "grad_norm": 0.3203125, "learning_rate": 7.981859410430839e-07, "loss": 0.0505, "num_tokens": 17910582.0, "reward": 0.14453125, "reward_std": 0.458310067653656, "rewards/accuracy_reward_func/mean": 0.140625, "rewards/accuracy_reward_func/std": 0.4583333432674408, "rewards/format_reward_func/mean": 0.00390625, "rewards/format_reward_func/std": 0.03125, "step": 90, "step_time": 598.1242492264137 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3946.5, "completions/mean_length": 1446.625, "completions/mean_terminated_length": 948.6966552734375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.583383908495307, "epoch": 0.20634920634920634, "frac_reward_zero_std": 0.9375, "grad_norm": 0.10595703125, "learning_rate": 7.959183673469388e-07, "loss": 0.0011, "num_tokens": 18119978.0, "reward": 0.0234375, "reward_std": 0.13152606785297394, "rewards/accuracy_reward_func/mean": 0.015625, "rewards/accuracy_reward_func/std": 0.125, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.043842025101184845, "step": 91, "step_time": 594.175365534611 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1328125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3648.5, "completions/mean_length": 1342.2109375, "completions/mean_terminated_length": 925.1038513183594, "completions/min_length": 98.0, "completions/min_terminated_length": 98.0, "entropy": 3.504945607855916, "epoch": 0.20861678004535147, "frac_reward_zero_std": 0.96875, "grad_norm": 0.09423828125, "learning_rate": 7.936507936507936e-07, "loss": 0.021, "num_tokens": 18315805.0, "reward": 0.015625, "reward_std": 0.125, "rewards/accuracy_reward_func/mean": 0.015625, "rewards/accuracy_reward_func/std": 0.125, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 92, "step_time": 594.0875607496127 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3821.0, "completions/mean_length": 1374.109375, "completions/mean_terminated_length": 924.5326232910156, "completions/min_length": 17.5, "completions/min_terminated_length": 17.5, "entropy": 3.565487129613757, "epoch": 0.2108843537414966, "frac_reward_zero_std": 0.90625, "grad_norm": 0.1416015625, "learning_rate": 7.913832199546485e-07, "loss": -0.0075, "num_tokens": 18514759.0, "reward": 0.05859375, "reward_std": 0.30986589193344116, "rewards/accuracy_reward_func/mean": 0.046875, "rewards/accuracy_reward_func/std": 0.3003681004047394, "rewards/format_reward_func/mean": 0.01171875, "rewards/format_reward_func/std": 0.05326050892472267, "step": 93, "step_time": 582.082564773038 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3819.0, "completions/mean_length": 1283.6640625, "completions/mean_terminated_length": 762.8611145019531, "completions/min_length": 4.5, "completions/min_terminated_length": 4.5, "entropy": 2.893090734258294, "epoch": 0.21315192743764172, "frac_reward_zero_std": 0.8125, "grad_norm": 0.22265625, "learning_rate": 7.891156462585033e-07, "loss": 0.0267, "num_tokens": 18701720.0, "reward": 0.09765625, "reward_std": 0.4274981766939163, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.00390625, "rewards/format_reward_func/std": 0.03125, "step": 94, "step_time": 582.8762040715665 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1953125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3917.0, "completions/mean_length": 1566.3828125, "completions/mean_terminated_length": 956.0291748046875, "completions/min_length": 39.0, "completions/min_terminated_length": 39.0, "entropy": 3.8191556986421347, "epoch": 0.21541950113378686, "frac_reward_zero_std": 0.8125, "grad_norm": 0.2197265625, "learning_rate": 7.868480725623582e-07, "loss": 0.0048, "num_tokens": 18927125.0, "reward": 0.078125, "reward_std": 0.3586694449186325, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08451050892472267, "step": 95, "step_time": 593.5745522053912 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4037.0, "completions/mean_length": 1239.0546875, "completions/mean_terminated_length": 943.9960021972656, "completions/min_length": 72.0, "completions/min_terminated_length": 72.0, "entropy": 4.091521814465523, "epoch": 0.21768707482993196, "frac_reward_zero_std": 0.75, "grad_norm": 0.2353515625, "learning_rate": 7.845804988662131e-07, "loss": 0.0563, "num_tokens": 19110208.0, "reward": 0.1796875, "reward_std": 0.5162215977907181, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5087014436721802, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10483578220009804, "step": 96, "step_time": 605.531951376237 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 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": 4096.0, "completions/max_terminated_length": 3760.0, "completions/mean_length": 1143.9453125, "completions/mean_terminated_length": 780.8703918457031, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.5440098233520985, "epoch": 0.2199546485260771, "frac_reward_zero_std": 0.84375, "grad_norm": 0.240234375, "learning_rate": 7.82312925170068e-07, "loss": 0.0327, "num_tokens": 19282685.0, "reward": 0.0703125, "reward_std": 0.34456808865070343, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.3380420356988907, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.043842025101184845, "step": 97, "step_time": 613.6265974482521 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3890.5, "completions/mean_length": 1450.671875, "completions/mean_terminated_length": 1021.3078918457031, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "entropy": 4.116855856031179, "epoch": 0.2222222222222222, "frac_reward_zero_std": 0.84375, "grad_norm": 0.189453125, "learning_rate": 7.800453514739229e-07, "loss": 0.0454, "num_tokens": 19491999.0, "reward": 0.1015625, "reward_std": 0.38316610455513, "rewards/accuracy_reward_func/mean": 0.078125, "rewards/accuracy_reward_func/std": 0.36897502839565277, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10483578220009804, "step": 98, "step_time": 580.59360064473 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3202.5, "completions/mean_length": 1423.65625, "completions/mean_terminated_length": 867.8502807617188, "completions/min_length": 133.5, "completions/min_terminated_length": 133.5, "entropy": 2.863107653334737, "epoch": 0.22448979591836735, "frac_reward_zero_std": 0.8125, "grad_norm": 0.2275390625, "learning_rate": 7.777777777777778e-07, "loss": 0.0136, "num_tokens": 19698111.0, "reward": 0.1015625, "reward_std": 0.38122110068798065, "rewards/accuracy_reward_func/mean": 0.078125, "rewards/accuracy_reward_func/std": 0.36897502839565277, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 99, "step_time": 594.6842877250165 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1640625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3543.0, "completions/mean_length": 1284.8125, "completions/mean_terminated_length": 734.4361267089844, "completions/min_length": 62.5, "completions/min_terminated_length": 62.5, "entropy": 3.267590468749404, "epoch": 0.22675736961451248, "frac_reward_zero_std": 0.84375, "grad_norm": 0.2431640625, "learning_rate": 7.755102040816326e-07, "loss": 0.0142, "num_tokens": 19884891.0, "reward": 0.0625, "reward_std": 0.31193090975284576, "rewards/accuracy_reward_func/mean": 0.046875, "rewards/accuracy_reward_func/std": 0.3003681004047394, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08451050892472267, "step": 100, "step_time": 585.4415724445134 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3984.5, "completions/mean_length": 1305.8125, "completions/mean_terminated_length": 917.0467224121094, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.693891802802682, "epoch": 0.2290249433106576, "frac_reward_zero_std": 0.84375, "grad_norm": 0.2734375, "learning_rate": 7.732426303854876e-07, "loss": -0.0776, "num_tokens": 20077307.0, "reward": 0.06640625, "reward_std": 0.31276239454746246, "rewards/accuracy_reward_func/mean": 0.046875, "rewards/accuracy_reward_func/std": 0.3003681004047394, "rewards/format_reward_func/mean": 0.01953125, "rewards/format_reward_func/std": 0.09710253402590752, "step": 101, "step_time": 596.8676613187417 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3850.0, "completions/mean_length": 1559.4140625, "completions/mean_terminated_length": 1031.7346801757812, "completions/min_length": 102.5, "completions/min_terminated_length": 102.5, "entropy": 4.4218991585075855, "epoch": 0.23129251700680273, "frac_reward_zero_std": 0.8125, "grad_norm": 0.21875, "learning_rate": 7.709750566893424e-07, "loss": -0.0576, "num_tokens": 20300304.0, "reward": 0.07421875, "reward_std": 0.34598222374916077, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.3380420356988907, "rewards/format_reward_func/mean": 0.01171875, "rewards/format_reward_func/std": 0.07509202510118484, "step": 102, "step_time": 585.8930206624791 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3971.0, "completions/mean_length": 1271.1875, "completions/mean_terminated_length": 868.7754211425781, "completions/min_length": 2.5, "completions/min_terminated_length": 2.5, "entropy": 3.360744481906295, "epoch": 0.23356009070294784, "frac_reward_zero_std": 0.8125, "grad_norm": 0.2197265625, "learning_rate": 7.687074829931972e-07, "loss": -0.0222, "num_tokens": 20487620.0, "reward": 0.10546875, "reward_std": 0.4000239223241806, "rewards/accuracy_reward_func/mean": 0.078125, "rewards/accuracy_reward_func/std": 0.38841013610363007, "rewards/format_reward_func/mean": 0.02734375, "rewards/format_reward_func/std": 0.11425426602363586, "step": 103, "step_time": 608.8104765452445 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0703125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3744.0, "completions/mean_length": 1091.703125, "completions/mean_terminated_length": 865.4914245605469, "completions/min_length": 15.0, "completions/min_terminated_length": 15.0, "entropy": 2.9236565800383687, "epoch": 0.23582766439909297, "frac_reward_zero_std": 0.8125, "grad_norm": 0.25390625, "learning_rate": 7.664399092970521e-07, "loss": 0.0348, "num_tokens": 20651146.0, "reward": 0.12890625, "reward_std": 0.4921843260526657, "rewards/accuracy_reward_func/mean": 0.109375, "rewards/accuracy_reward_func/std": 0.45701706409454346, "rewards/format_reward_func/mean": 0.01953125, "rewards/format_reward_func/std": 0.09224375709891319, "step": 104, "step_time": 589.7991110961884 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1484375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3982.5, "completions/mean_length": 1259.171875, "completions/mean_terminated_length": 764.8927307128906, "completions/min_length": 10.0, "completions/min_terminated_length": 10.0, "entropy": 3.0813568271696568, "epoch": 0.23809523809523808, "frac_reward_zero_std": 0.90625, "grad_norm": 0.1865234375, "learning_rate": 7.64172335600907e-07, "loss": 0.0208, "num_tokens": 20837072.0, "reward": 0.03125, "reward_std": 0.16574779152870178, "rewards/accuracy_reward_func/mean": 0.015625, "rewards/accuracy_reward_func/std": 0.125, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08451050892472267, "step": 105, "step_time": 588.3291091462597 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3898.5, "completions/mean_length": 1365.6796875, "completions/mean_terminated_length": 996.0594787597656, "completions/min_length": 44.5, "completions/min_terminated_length": 44.5, "entropy": 4.366790888831019, "epoch": 0.24036281179138322, "frac_reward_zero_std": 0.90625, "grad_norm": 0.1962890625, "learning_rate": 7.619047619047618e-07, "loss": 0.0413, "num_tokens": 21034843.0, "reward": 0.03515625, "reward_std": 0.1847865730524063, "rewards/accuracy_reward_func/mean": 0.015625, "rewards/accuracy_reward_func/std": 0.125, "rewards/format_reward_func/mean": 0.01953125, "rewards/format_reward_func/std": 0.09710253402590752, "step": 106, "step_time": 580.5758132711053 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3816.5, "completions/mean_length": 1398.8203125, "completions/mean_terminated_length": 840.67041015625, "completions/min_length": 19.0, "completions/min_terminated_length": 19.0, "entropy": 3.526266535744071, "epoch": 0.24263038548752835, "frac_reward_zero_std": 0.78125, "grad_norm": 0.2294921875, "learning_rate": 7.596371882086168e-07, "loss": 0.0453, "num_tokens": 21239968.0, "reward": 0.13671875, "reward_std": 0.4986996203660965, "rewards/accuracy_reward_func/mean": 0.109375, "rewards/accuracy_reward_func/std": 0.45701706409454346, "rewards/format_reward_func/mean": 0.02734375, "rewards/format_reward_func/std": 0.11425426602363586, "step": 107, "step_time": 614.5666125547141 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1484375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3981.5, "completions/mean_length": 1242.21875, "completions/mean_terminated_length": 745.8328247070312, "completions/min_length": 8.5, "completions/min_terminated_length": 8.5, "entropy": 2.7671644650399685, "epoch": 0.24489795918367346, "frac_reward_zero_std": 0.84375, "grad_norm": 0.189453125, "learning_rate": 7.573696145124716e-07, "loss": 0.0189, "num_tokens": 21421980.0, "reward": 0.06640625, "reward_std": 0.2752250134944916, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.24397502839565277, "rewards/format_reward_func/mean": 0.00390625, "rewards/format_reward_func/std": 0.03125, "step": 108, "step_time": 583.8319079326466 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3914.5, "completions/mean_length": 1306.1796875, "completions/mean_terminated_length": 911.6589660644531, "completions/min_length": 26.0, "completions/min_terminated_length": 26.0, "entropy": 3.7471233755350113, "epoch": 0.2471655328798186, "frac_reward_zero_std": 0.84375, "grad_norm": 0.1708984375, "learning_rate": 7.551020408163265e-07, "loss": -0.0116, "num_tokens": 21615723.0, "reward": 0.1171875, "reward_std": 0.3480724096298218, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.29378482699394226, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 109, "step_time": 635.0032001212239 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 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": 4096.0, "completions/max_terminated_length": 3856.5, "completions/mean_length": 1128.1640625, "completions/mean_terminated_length": 759.73974609375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 2.9849752727895975, "epoch": 0.2494331065759637, "frac_reward_zero_std": 0.8125, "grad_norm": 0.2275390625, "learning_rate": 7.528344671201814e-07, "loss": 0.0177, "num_tokens": 21783176.0, "reward": 0.09765625, "reward_std": 0.4203259348869324, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.41934312880039215, "rewards/format_reward_func/mean": 0.00390625, "rewards/format_reward_func/std": 0.03125, "step": 110, "step_time": 591.5175723498687 } ], "logging_steps": 1, "max_steps": 441, "num_input_tokens_seen": 21783176, "num_train_epochs": 1, "save_steps": 500, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": false }, "attributes": {} } }, "total_flos": 0.0, "train_batch_size": 4, "trial_name": null, "trial_params": null }