{ "best_global_step": null, "best_metric": null, "best_model_checkpoint": null, "epoch": 0.11160714285714286, "eval_steps": 500, "global_step": 50, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2011.0, "completions/mean_length": 536.546875, "completions/mean_terminated_length": 512.5556030273438, "completions/min_length": 132.0, "completions/min_terminated_length": 132.0, "epoch": 0.002232142857142857, "frac_reward_zero_std": 0.0, "grad_norm": 0.9917449355125427, "learning_rate": 0.0, "loss": -0.0, "num_tokens": 116139.0, "reward": 0.23906248807907104, "reward_std": 0.3186468482017517, "rewards/format_reward/mean": 0.359375, "rewards/format_reward/std": 0.2741328477859497, "rewards/mcq_exact_match_reward/mean": 0.203125, "rewards/mcq_exact_match_reward/std": 0.40550529956817627, "step": 1 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1485.0, "completions/max_terminated_length": 1485.0, "completions/mean_length": 368.03125, "completions/mean_terminated_length": 368.03125, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "epoch": 0.004464285714285714, "frac_reward_zero_std": 0.0, "grad_norm": 10.428349494934082, "learning_rate": 2e-07, "loss": -0.0, "num_tokens": 225077.0, "reward": 0.2070312350988388, "reward_std": 0.2855132222175598, "rewards/format_reward/mean": 0.3515625, "rewards/format_reward/std": 0.27716949582099915, "rewards/mcq_exact_match_reward/mean": 0.171875, "rewards/mcq_exact_match_reward/std": 0.38025420904159546, "step": 2 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1265.0, "completions/max_terminated_length": 1265.0, "completions/mean_length": 385.75, "completions/mean_terminated_length": 385.75, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.006696428571428571, "frac_reward_zero_std": 0.0, "grad_norm": 11.403729438781738, "learning_rate": 4e-07, "loss": 0.0, "num_tokens": 336837.0, "reward": 0.3789062201976776, "reward_std": 0.38316458463668823, "rewards/format_reward/mean": 0.3515625, "rewards/format_reward/std": 0.26246222853660583, "rewards/mcq_exact_match_reward/mean": 0.34375, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "step": 3 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1341.0, "completions/mean_length": 417.28125, "completions/mean_terminated_length": 364.6773986816406, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.008928571428571428, "frac_reward_zero_std": 0.0, "grad_norm": 22.8345947265625, "learning_rate": 6e-07, "loss": -0.0, "num_tokens": 438239.0, "reward": 0.23281249403953552, "reward_std": 0.3445912003517151, "rewards/format_reward/mean": 0.296875, "rewards/format_reward/std": 0.2916666865348816, "rewards/mcq_exact_match_reward/mean": 0.203125, "rewards/mcq_exact_match_reward/std": 0.40550529956817627, "step": 4 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1608.0, "completions/max_terminated_length": 1608.0, "completions/mean_length": 407.46875, "completions/mean_terminated_length": 407.46875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.011160714285714286, "frac_reward_zero_std": 0.0, "grad_norm": 5.248055458068848, "learning_rate": 8e-07, "loss": -0.0, "num_tokens": 547109.0, "reward": 0.37968748807907104, "reward_std": 0.3280304968357086, "rewards/format_reward/mean": 0.359375, "rewards/format_reward/std": 0.2741328477859497, "rewards/mcq_exact_match_reward/mean": 0.34375, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "step": 5 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1527.0, "completions/max_terminated_length": 1527.0, "completions/mean_length": 405.734375, "completions/mean_terminated_length": 405.734375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.013392857142857142, "frac_reward_zero_std": 0.0, "grad_norm": 11.29982852935791, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 661452.0, "reward": 0.4015624523162842, "reward_std": 0.328488290309906, "rewards/format_reward/mean": 0.421875, "rewards/format_reward/std": 0.27048972249031067, "rewards/mcq_exact_match_reward/mean": 0.359375, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "step": 6 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1392.0, "completions/mean_length": 546.484375, "completions/mean_terminated_length": 498.0483703613281, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.015625, "frac_reward_zero_std": 0.0, "grad_norm": 4.50560188293457, "learning_rate": 9.99726628670463e-07, "loss": 0.0, "num_tokens": 806683.0, "reward": 0.23124998807907104, "reward_std": 0.30694860219955444, "rewards/format_reward/mean": 0.28125, "rewards/format_reward/std": 0.265398770570755, "rewards/mcq_exact_match_reward/mean": 0.203125, "rewards/mcq_exact_match_reward/std": 0.40550529956817627, "step": 7 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1600.0, "completions/mean_length": 626.90625, "completions/mean_terminated_length": 581.0645141601562, "completions/min_length": 19.0, "completions/min_terminated_length": 19.0, "epoch": 0.017857142857142856, "frac_reward_zero_std": 0.125, "grad_norm": 3.363801956176758, "learning_rate": 9.989068136093872e-07, "loss": -0.0, "num_tokens": 940517.0, "reward": 0.29531246423721313, "reward_std": 0.34660738706588745, "rewards/format_reward/mean": 0.453125, "rewards/format_reward/std": 0.21304203569889069, "rewards/mcq_exact_match_reward/mean": 0.25, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "step": 8 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1435.0, "completions/max_terminated_length": 1435.0, "completions/mean_length": 320.125, "completions/mean_terminated_length": 320.125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.020089285714285716, "frac_reward_zero_std": 0.25, "grad_norm": 24.068775177001953, "learning_rate": 9.975414512725056e-07, "loss": 0.0, "num_tokens": 1057261.0, "reward": 0.27421873807907104, "reward_std": 0.3172922134399414, "rewards/format_reward/mean": 0.3984375, "rewards/format_reward/std": 0.2387082874774933, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "step": 9 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1329.0, "completions/max_terminated_length": 1329.0, "completions/mean_length": 380.28125, "completions/mean_terminated_length": 380.28125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.022321428571428572, "frac_reward_zero_std": 0.0, "grad_norm": 10.925803184509277, "learning_rate": 9.956320346634875e-07, "loss": 0.0, "num_tokens": 1197335.0, "reward": 0.3828124701976776, "reward_std": 0.4325318932533264, "rewards/format_reward/mean": 0.390625, "rewards/format_reward/std": 0.2592533528804779, "rewards/mcq_exact_match_reward/mean": 0.34375, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "step": 10 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1757.0, "completions/max_terminated_length": 1757.0, "completions/mean_length": 362.421875, "completions/mean_terminated_length": 362.421875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.024553571428571428, "frac_reward_zero_std": 0.25, "grad_norm": 15.241398811340332, "learning_rate": 9.931806517013612e-07, "loss": -0.0, "num_tokens": 1321602.0, "reward": 0.24296873807907104, "reward_std": 0.2843829393386841, "rewards/format_reward/mean": 0.3984375, "rewards/format_reward/std": 0.20275264978408813, "rewards/mcq_exact_match_reward/mean": 0.203125, "rewards/mcq_exact_match_reward/std": 0.40550529956817627, "step": 11 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1628.0, "completions/mean_length": 350.15625, "completions/mean_terminated_length": 323.20635986328125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.026785714285714284, "frac_reward_zero_std": 0.0, "grad_norm": 17.23383140563965, "learning_rate": 9.901899829374047e-07, "loss": -0.0, "num_tokens": 1427108.0, "reward": 0.3687499761581421, "reward_std": 0.43608352541923523, "rewards/format_reward/mean": 0.40625, "rewards/format_reward/std": 0.33184191584587097, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "step": 12 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1941.0, "completions/mean_length": 420.0625, "completions/mean_terminated_length": 394.2222595214844, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.029017857142857144, "frac_reward_zero_std": 0.0, "grad_norm": 7.505155563354492, "learning_rate": 9.866632986240029e-07, "loss": -0.0, "num_tokens": 1534736.0, "reward": 0.2109375, "reward_std": 0.32654353976249695, "rewards/format_reward/mean": 0.546875, "rewards/format_reward/std": 0.2630521357059479, "rewards/mcq_exact_match_reward/mean": 0.15625, "rewards/mcq_exact_match_reward/std": 0.36596253514289856, "step": 13 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1861.0, "completions/max_terminated_length": 1861.0, "completions/mean_length": 271.203125, "completions/mean_terminated_length": 271.203125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.03125, "frac_reward_zero_std": 0.125, "grad_norm": 18.432064056396484, "learning_rate": 9.826044551386742e-07, "loss": -0.0, "num_tokens": 1658965.0, "reward": 0.49687492847442627, "reward_std": 0.36446183919906616, "rewards/format_reward/mean": 0.4375, "rewards/format_reward/std": 0.1666666716337204, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 14 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1229.0, "completions/mean_length": 150.046875, "completions/mean_terminated_length": 119.92063903808594, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.033482142857142856, "frac_reward_zero_std": 0.625, "grad_norm": 19.670541763305664, "learning_rate": 9.780178907671788e-07, "loss": 0.0, "num_tokens": 1763384.0, "reward": 0.5101562738418579, "reward_std": 0.19758465886116028, "rewards/format_reward/mean": 0.5703125, "rewards/format_reward/std": 0.21578919887542725, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 15 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1672.0, "completions/max_terminated_length": 1672.0, "completions/mean_length": 125.828125, "completions/mean_terminated_length": 125.828125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.03571428571428571, "frac_reward_zero_std": 0.125, "grad_norm": 30.569133758544922, "learning_rate": 9.729086208503173e-07, "loss": 0.0, "num_tokens": 1854541.0, "reward": 0.5414062142372131, "reward_std": 0.3450131118297577, "rewards/format_reward/mean": 0.5703125, "rewards/format_reward/std": 0.19654127955436707, "rewards/mcq_exact_match_reward/mean": 0.484375, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "step": 16 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 18.0, "completions/max_terminated_length": 18.0, "completions/mean_length": 6.28125, "completions/mean_terminated_length": 6.28125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.03794642857142857, "frac_reward_zero_std": 0.5, "grad_norm": 22.943239212036133, "learning_rate": 9.672822322997304e-07, "loss": 0.0, "num_tokens": 1949287.0, "reward": 0.45781248807907104, "reward_std": 0.22371627390384674, "rewards/format_reward/mean": 0.515625, "rewards/format_reward/std": 0.08768405020236969, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "step": 17 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 913.0, "completions/max_terminated_length": 913.0, "completions/mean_length": 103.0625, "completions/mean_terminated_length": 103.0625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.04017857142857143, "frac_reward_zero_std": 0.375, "grad_norm": 25.630355834960938, "learning_rate": 9.611448774886923e-07, "loss": 0.0, "num_tokens": 2043563.0, "reward": 0.37890625, "reward_std": 0.23165243864059448, "rewards/format_reward/mean": 0.5078125, "rewards/format_reward/std": 0.3020063638687134, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "step": 18 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 749.0, "completions/max_terminated_length": 749.0, "completions/mean_length": 17.796875, "completions/mean_terminated_length": 17.796875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.04241071428571429, "frac_reward_zero_std": 0.75, "grad_norm": 22.894906997680664, "learning_rate": 9.545032675245813e-07, "loss": 0.0, "num_tokens": 2136918.0, "reward": 0.67578125, "reward_std": 0.005102896131575108, "rewards/format_reward/mean": 0.5078125, "rewards/format_reward/std": 0.10882443934679031, "rewards/mcq_exact_match_reward/mean": 0.625, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "step": 19 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 922.0, "completions/max_terminated_length": 922.0, "completions/mean_length": 22.28125, "completions/mean_terminated_length": 22.28125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.044642857142857144, "frac_reward_zero_std": 0.625, "grad_norm": 14.886871337890625, "learning_rate": 9.473646649103817e-07, "loss": -0.0, "num_tokens": 2238624.0, "reward": 0.36171871423721313, "reward_std": 0.11244211345911026, "rewards/format_reward/mean": 0.4921875, "rewards/format_reward/std": 0.10882443934679031, "rewards/mcq_exact_match_reward/mean": 0.3125, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "step": 20 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 448.0, "completions/max_terminated_length": 448.0, "completions/mean_length": 36.765625, "completions/mean_terminated_length": 36.765625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.046875, "frac_reward_zero_std": 0.375, "grad_norm": 40.487422943115234, "learning_rate": 9.397368756032444e-07, "loss": 0.0, "num_tokens": 2304801.0, "reward": 0.762499988079071, "reward_std": 0.15898248553276062, "rewards/format_reward/mean": 0.59375, "rewards/format_reward/std": 0.19669894874095917, "rewards/mcq_exact_match_reward/mean": 0.703125, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "step": 21 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 104.0, "completions/max_terminated_length": 104.0, "completions/mean_length": 8.3125, "completions/mean_terminated_length": 8.3125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.049107142857142856, "frac_reward_zero_std": 0.5, "grad_norm": 14.037199020385742, "learning_rate": 9.316282404787869e-07, "loss": 0.0, "num_tokens": 2397005.0, "reward": 0.2906249761581421, "reward_std": 0.20063331723213196, "rewards/format_reward/mean": 0.5625, "rewards/format_reward/std": 0.18898223340511322, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "step": 22 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 390.0, "completions/max_terminated_length": 390.0, "completions/mean_length": 14.9375, "completions/mean_terminated_length": 14.9375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.05133928571428571, "frac_reward_zero_std": 0.375, "grad_norm": 32.929039001464844, "learning_rate": 9.230476262104676e-07, "loss": -0.0, "num_tokens": 2500105.0, "reward": 0.4742187261581421, "reward_std": 0.23467989265918732, "rewards/format_reward/mean": 0.5234375, "rewards/format_reward/std": 0.10652101784944534, "rewards/mcq_exact_match_reward/mean": 0.421875, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "step": 23 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 164.0, "completions/max_terminated_length": 164.0, "completions/mean_length": 9.828125, "completions/mean_terminated_length": 9.828125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.05357142857142857, "frac_reward_zero_std": 0.0, "grad_norm": 34.22531509399414, "learning_rate": 9.1400441557401e-07, "loss": -0.0, "num_tokens": 2579166.0, "reward": 0.43593746423721313, "reward_std": 0.47567078471183777, "rewards/format_reward/mean": 0.609375, "rewards/format_reward/std": 0.2083333432674408, "rewards/mcq_exact_match_reward/mean": 0.375, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "step": 24 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 479.0, "completions/max_terminated_length": 479.0, "completions/mean_length": 42.328125, "completions/mean_terminated_length": 42.328125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.05580357142857143, "frac_reward_zero_std": 0.5, "grad_norm": 15.287548065185547, "learning_rate": 9.045084971874737e-07, "loss": 0.0, "num_tokens": 2671595.0, "reward": 0.5390625, "reward_std": 0.2117736041545868, "rewards/format_reward/mean": 0.546875, "rewards/format_reward/std": 0.23091863095760345, "rewards/mcq_exact_match_reward/mean": 0.484375, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "step": 25 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 684.0, "completions/max_terminated_length": 684.0, "completions/mean_length": 33.109375, "completions/mean_terminated_length": 33.109375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.05803571428571429, "frac_reward_zero_std": 0.625, "grad_norm": 27.692184448242188, "learning_rate": 8.945702546981968e-07, "loss": -0.0, "num_tokens": 2751314.0, "reward": 0.29296875, "reward_std": 0.09059805423021317, "rewards/format_reward/mean": 0.7421875, "rewards/format_reward/std": 0.25185325741767883, "rewards/mcq_exact_match_reward/mean": 0.21875, "rewards/mcq_exact_match_reward/std": 0.4166666865348816, "step": 26 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 47.0, "completions/max_terminated_length": 47.0, "completions/mean_length": 7.390625, "completions/mean_terminated_length": 7.390625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.060267857142857144, "frac_reward_zero_std": 0.25, "grad_norm": 24.060562133789062, "learning_rate": 8.842005554284295e-07, "loss": 0.0, "num_tokens": 2845155.0, "reward": 0.5234375, "reward_std": 0.27853256464004517, "rewards/format_reward/mean": 0.546875, "rewards/format_reward/std": 0.19352105259895325, "rewards/mcq_exact_match_reward/mean": 0.46875, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "step": 27 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 104.0, "completions/max_terminated_length": 104.0, "completions/mean_length": 11.25, "completions/mean_terminated_length": 11.25, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0625, "frac_reward_zero_std": 0.5, "grad_norm": 15.91177749633789, "learning_rate": 8.734107384920769e-07, "loss": -0.0, "num_tokens": 2934323.0, "reward": 0.4273437261581421, "reward_std": 0.19265598058700562, "rewards/format_reward/mean": 0.6796875, "rewards/format_reward/std": 0.2576941251754761, "rewards/mcq_exact_match_reward/mean": 0.359375, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "step": 28 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 796.0, "completions/max_terminated_length": 796.0, "completions/mean_length": 19.109375, "completions/mean_terminated_length": 19.109375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.06473214285714286, "frac_reward_zero_std": 0.25, "grad_norm": 23.2950439453125, "learning_rate": 8.622126023955445e-07, "loss": -0.0, "num_tokens": 3018370.0, "reward": 0.5562499761581421, "reward_std": 0.34034281969070435, "rewards/format_reward/mean": 0.5625, "rewards/format_reward/std": 0.1666666716337204, "rewards/mcq_exact_match_reward/mean": 0.5, "rewards/mcq_exact_match_reward/std": 0.5039526224136353, "step": 29 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 88.0, "completions/max_terminated_length": 88.0, "completions/mean_length": 8.140625, "completions/mean_terminated_length": 8.140625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.06696428571428571, "frac_reward_zero_std": 0.5, "grad_norm": 24.887393951416016, "learning_rate": 8.506183921362442e-07, "loss": 0.0, "num_tokens": 3115531.0, "reward": 0.4429687261581421, "reward_std": 0.19260738790035248, "rewards/format_reward/mean": 0.5234375, "rewards/format_reward/std": 0.22589658200740814, "rewards/mcq_exact_match_reward/mean": 0.390625, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "step": 30 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 19.0, "completions/max_terminated_length": 19.0, "completions/mean_length": 9.109375, "completions/mean_terminated_length": 9.109375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.06919642857142858, "frac_reward_zero_std": 0.625, "grad_norm": 15.75447940826416, "learning_rate": 8.386407858128706e-07, "loss": 0.0, "num_tokens": 3188034.0, "reward": 0.38749998807907104, "reward_std": 0.16675157845020294, "rewards/format_reward/mean": 0.75, "rewards/format_reward/std": 0.2519763112068176, "rewards/mcq_exact_match_reward/mean": 0.3125, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "step": 31 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 339.0, "completions/max_terminated_length": 339.0, "completions/mean_length": 14.4375, "completions/mean_terminated_length": 14.4375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.07142857142857142, "frac_reward_zero_std": 0.25, "grad_norm": 34.575897216796875, "learning_rate": 8.262928807620843e-07, "loss": 0.0, "num_tokens": 3274902.0, "reward": 0.6507812142372131, "reward_std": 0.21801188588142395, "rewards/format_reward/mean": 0.5703125, "rewards/format_reward/std": 0.3068941831588745, "rewards/mcq_exact_match_reward/mean": 0.59375, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "step": 32 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 157.0, "completions/max_terminated_length": 157.0, "completions/mean_length": 16.9375, "completions/mean_terminated_length": 16.9375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.07366071428571429, "frac_reward_zero_std": 0.5, "grad_norm": 30.152292251586914, "learning_rate": 8.135881792367685e-07, "loss": -0.0, "num_tokens": 3353810.0, "reward": 0.24531251192092896, "reward_std": 0.0730312243103981, "rewards/format_reward/mean": 0.734375, "rewards/format_reward/std": 0.2514837086200714, "rewards/mcq_exact_match_reward/mean": 0.171875, "rewards/mcq_exact_match_reward/std": 0.38025420904159546, "step": 33 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 59.0, "completions/max_terminated_length": 59.0, "completions/mean_length": 11.453125, "completions/mean_terminated_length": 11.453125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.07589285714285714, "frac_reward_zero_std": 0.25, "grad_norm": 27.871400833129883, "learning_rate": 8.005405736415125e-07, "loss": 0.0, "num_tokens": 3431815.0, "reward": 0.828125, "reward_std": 0.229772686958313, "rewards/format_reward/mean": 0.78125, "rewards/format_reward/std": 0.265398770570755, "rewards/mcq_exact_match_reward/mean": 0.75, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "step": 34 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 690.0, "completions/max_terminated_length": 690.0, "completions/mean_length": 37.03125, "completions/mean_terminated_length": 37.03125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.078125, "frac_reward_zero_std": 0.375, "grad_norm": 26.60154914855957, "learning_rate": 7.871643313414718e-07, "loss": 0.0, "num_tokens": 3529713.0, "reward": 0.6273437738418579, "reward_std": 0.15764495730400085, "rewards/format_reward/mean": 0.8046875, "rewards/format_reward/std": 0.24587368965148926, "rewards/mcq_exact_match_reward/mean": 0.546875, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 35 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 158.0, "completions/max_terminated_length": 158.0, "completions/mean_length": 25.671875, "completions/mean_terminated_length": 25.671875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08035714285714286, "frac_reward_zero_std": 0.5, "grad_norm": 13.673951148986816, "learning_rate": 7.734740790612136e-07, "loss": -0.0, "num_tokens": 3615652.0, "reward": 0.47968751192092896, "reward_std": 0.23193931579589844, "rewards/format_reward/mean": 0.890625, "rewards/format_reward/std": 0.2083333432674408, "rewards/mcq_exact_match_reward/mean": 0.390625, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "step": 36 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 19.0, "completions/max_terminated_length": 19.0, "completions/mean_length": 12.234375, "completions/mean_terminated_length": 12.234375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08258928571428571, "frac_reward_zero_std": 0.375, "grad_norm": 35.03842544555664, "learning_rate": 7.594847868906076e-07, "loss": 0.0, "num_tokens": 3718803.0, "reward": 0.4921875, "reward_std": 0.21636667847633362, "rewards/format_reward/mean": 0.859375, "rewards/format_reward/std": 0.22658175230026245, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "step": 37 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 334.0, "completions/mean_length": 58.859375, "completions/mean_terminated_length": 27.285715103149414, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.08482142857142858, "frac_reward_zero_std": 0.5, "grad_norm": 16.576311111450195, "learning_rate": 7.452117519152541e-07, "loss": 0.0, "num_tokens": 3804130.0, "reward": 0.518750011920929, "reward_std": 0.1603945940732956, "rewards/format_reward/mean": 0.96875, "rewards/format_reward/std": 0.1510545015335083, "rewards/mcq_exact_match_reward/mean": 0.421875, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "step": 38 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 287.0, "completions/max_terminated_length": 287.0, "completions/mean_length": 27.09375, "completions/mean_terminated_length": 27.09375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08705357142857142, "frac_reward_zero_std": 0.5, "grad_norm": 25.71928596496582, "learning_rate": 7.306705814893439e-07, "loss": 0.0, "num_tokens": 3883480.0, "reward": 0.5218750238418579, "reward_std": 0.15509094297885895, "rewards/format_reward/mean": 0.84375, "rewards/format_reward/std": 0.265398770570755, "rewards/mcq_exact_match_reward/mean": 0.4375, "rewards/mcq_exact_match_reward/std": 0.5, "step": 39 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 348.0, "completions/max_terminated_length": 348.0, "completions/mean_length": 21.84375, "completions/mean_terminated_length": 21.84375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08928571428571429, "frac_reward_zero_std": 0.375, "grad_norm": 33.05258560180664, "learning_rate": 7.158771761692464e-07, "loss": 0.0, "num_tokens": 3983030.0, "reward": 0.7125000357627869, "reward_std": 0.20931637287139893, "rewards/format_reward/mean": 0.875, "rewards/format_reward/std": 0.26726123690605164, "rewards/mcq_exact_match_reward/mean": 0.625, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "step": 40 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 400.0, "completions/max_terminated_length": 400.0, "completions/mean_length": 30.09375, "completions/mean_terminated_length": 30.09375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.09151785714285714, "frac_reward_zero_std": 0.875, "grad_norm": 25.020532608032227, "learning_rate": 7.008477123264847e-07, "loss": -0.0, "num_tokens": 4077068.0, "reward": 0.842968761920929, "reward_std": 0.002209708094596863, "rewards/format_reward/mean": 0.9296875, "rewards/format_reward/std": 0.1751912236213684, "rewards/mcq_exact_match_reward/mean": 0.75, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "step": 41 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 174.0, "completions/max_terminated_length": 174.0, "completions/mean_length": 16.296875, "completions/mean_terminated_length": 16.296875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.09375, "frac_reward_zero_std": 0.5, "grad_norm": 15.59770679473877, "learning_rate": 6.855986244591103e-07, "loss": 0.0, "num_tokens": 4145855.0, "reward": 0.6390625238418579, "reward_std": 0.16434110701084137, "rewards/format_reward/mean": 0.921875, "rewards/format_reward/std": 0.20351573824882507, "rewards/mcq_exact_match_reward/mean": 0.546875, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 42 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 197.0, "completions/max_terminated_length": 197.0, "completions/mean_length": 15.96875, "completions/mean_terminated_length": 15.96875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.09598214285714286, "frac_reward_zero_std": 0.5, "grad_norm": 14.986905097961426, "learning_rate": 6.701465872208216e-07, "loss": -0.0, "num_tokens": 4230677.0, "reward": 0.5843750238418579, "reward_std": 0.19044627249240875, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.484375, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "step": 43 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 424.0, "completions/max_terminated_length": 424.0, "completions/mean_length": 23.390625, "completions/mean_terminated_length": 23.390625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.09821428571428571, "frac_reward_zero_std": 0.625, "grad_norm": 6.284796237945557, "learning_rate": 6.545084971874736e-07, "loss": 0.0, "num_tokens": 4333862.0, "reward": 0.6609375476837158, "reward_std": 0.14709635078907013, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.5625, "rewards/mcq_exact_match_reward/std": 0.5, "step": 44 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 88.0, "completions/max_terminated_length": 88.0, "completions/mean_length": 14.890625, "completions/mean_terminated_length": 14.890625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.10044642857142858, "frac_reward_zero_std": 0.75, "grad_norm": 9.359296798706055, "learning_rate": 6.387014543809223e-07, "loss": 0.0, "num_tokens": 4442279.0, "reward": 0.7562500238418579, "reward_std": 0.10888782143592834, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.65625, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "step": 45 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 297.0, "completions/max_terminated_length": 297.0, "completions/mean_length": 20.421875, "completions/mean_terminated_length": 20.421875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.10267857142857142, "frac_reward_zero_std": 0.625, "grad_norm": 20.838117599487305, "learning_rate": 6.227427435703995e-07, "loss": 0.0, "num_tokens": 4526658.0, "reward": 0.5687500238418579, "reward_std": 0.17570312321186066, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.46875, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "step": 46 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 411.0, "completions/max_terminated_length": 411.0, "completions/mean_length": 35.671875, "completions/mean_terminated_length": 35.671875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.10491071428571429, "frac_reward_zero_std": 0.875, "grad_norm": 3.8892056941986084, "learning_rate": 6.066498153718734e-07, "loss": 0.0, "num_tokens": 4613941.0, "reward": 0.7234375476837158, "reward_std": 0.0044194171205163, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.625, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "step": 47 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 18.0, "completions/max_terminated_length": 18.0, "completions/mean_length": 12.75, "completions/mean_terminated_length": 12.75, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.10714285714285714, "frac_reward_zero_std": 0.75, "grad_norm": 9.113177299499512, "learning_rate": 5.90440267166055e-07, "loss": 0.0, "num_tokens": 4707493.0, "reward": 0.6156250238418579, "reward_std": 0.10205793380737305, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.515625, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "step": 48 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 82.0, "completions/max_terminated_length": 82.0, "completions/mean_length": 13.234375, "completions/mean_terminated_length": 13.234375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.109375, "frac_reward_zero_std": 0.5, "grad_norm": 18.525861740112305, "learning_rate": 5.741318238559209e-07, "loss": 0.0, "num_tokens": 4765364.0, "reward": 0.9265625476837158, "reward_std": 0.24481618404388428, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.828125, "rewards/mcq_exact_match_reward/std": 0.38025420904159546, "step": 49 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 115.0, "completions/max_terminated_length": 115.0, "completions/mean_length": 13.6875, "completions/mean_terminated_length": 13.6875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.11160714285714286, "frac_reward_zero_std": 0.625, "grad_norm": 19.748464584350586, "learning_rate": 5.577423184847931e-07, "loss": 0.0, "num_tokens": 4875504.0, "reward": 0.6468750238418579, "reward_std": 0.189372718334198, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.546875, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 50 } ], "logging_steps": 1, "max_steps": 100, "num_input_tokens_seen": 4875504, "num_train_epochs": 1, "save_steps": 50, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": false }, "attributes": {} } }, "total_flos": 0.0, "train_batch_size": 1, "trial_name": null, "trial_params": null }