{ "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1402.0, "completions/mean_length": 597.796875, "completions/mean_terminated_length": 551.01611328125, "completions/min_length": 130.0, "completions/min_terminated_length": 130.0, "epoch": 0.002232142857142857, "frac_reward_zero_std": 0.0, "grad_norm": 1.3016983270645142, "learning_rate": 0.0, "loss": -0.0, "num_tokens": 134163.0, "reward": 0.22187498211860657, "reward_std": 0.3342322111129761, "rewards/format_reward/mean": 0.34375, "rewards/format_reward/std": 0.29378482699394226, "rewards/mcq_exact_match_reward/mean": 0.1875, "rewards/mcq_exact_match_reward/std": 0.39339789748191833, "step": 1 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1654.0, "completions/mean_length": 545.4375, "completions/mean_terminated_length": 521.5873413085938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.004464285714285714, "frac_reward_zero_std": 0.0, "grad_norm": 15.761076927185059, "learning_rate": 2e-07, "loss": -0.0, "num_tokens": 296327.0, "reward": 0.35390621423721313, "reward_std": 0.26522397994995117, "rewards/format_reward/mean": 0.4140625, "rewards/format_reward/std": 0.19012710452079773, "rewards/mcq_exact_match_reward/mean": 0.3125, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "step": 2 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1734.0, "completions/mean_length": 473.65625, "completions/mean_terminated_length": 448.66668701171875, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "epoch": 0.006696428571428571, "frac_reward_zero_std": 0.0, "grad_norm": 3.760298013687134, "learning_rate": 4e-07, "loss": -0.0, "num_tokens": 411937.0, "reward": 0.30546873807907104, "reward_std": 0.3537239134311676, "rewards/format_reward/mean": 0.3984375, "rewards/format_reward/std": 0.22146137058734894, "rewards/mcq_exact_match_reward/mean": 0.265625, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "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": 1195.0, "completions/mean_length": 515.78125, "completions/mean_terminated_length": 466.3548278808594, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.008928571428571428, "frac_reward_zero_std": 0.0, "grad_norm": 6.041617393493652, "learning_rate": 6e-07, "loss": 0.0, "num_tokens": 525643.0, "reward": 0.2749999761581421, "reward_std": 0.36283189058303833, "rewards/format_reward/mean": 0.40625, "rewards/format_reward/std": 0.25, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "step": 4 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1664.0, "completions/mean_length": 509.484375, "completions/mean_terminated_length": 485.0635070800781, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.011160714285714286, "frac_reward_zero_std": 0.125, "grad_norm": 4.619335651397705, "learning_rate": 8e-07, "loss": -0.0, "num_tokens": 631682.0, "reward": 0.3374999761581421, "reward_std": 0.36903461813926697, "rewards/format_reward/mean": 0.40625, "rewards/format_reward/std": 0.29378482699394226, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "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": 0.046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1556.0, "completions/mean_length": 624.5625, "completions/mean_terminated_length": 554.557373046875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.013392857142857142, "frac_reward_zero_std": 0.0, "grad_norm": 1.5455710887908936, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 776998.0, "reward": 0.15937498211860657, "reward_std": 0.21251952648162842, "rewards/format_reward/mean": 0.34375, "rewards/format_reward/std": 0.25, "rewards/mcq_exact_match_reward/mean": 0.125, "rewards/mcq_exact_match_reward/std": 0.3333333432674408, "step": 6 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 546.515625, "completions/mean_terminated_length": 522.6825561523438, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.015625, "frac_reward_zero_std": 0.0, "grad_norm": 4.492587089538574, "learning_rate": 9.99726628670463e-07, "loss": -0.0, "num_tokens": 917303.0, "reward": 0.39375001192092896, "reward_std": 0.33318066596984863, "rewards/format_reward/mean": 0.34375, "rewards/format_reward/std": 0.265398770570755, "rewards/mcq_exact_match_reward/mean": 0.359375, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "step": 7 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1565.0, "completions/mean_length": 533.21875, "completions/mean_terminated_length": 509.17462158203125, "completions/min_length": 10.0, "completions/min_terminated_length": 10.0, "epoch": 0.017857142857142856, "frac_reward_zero_std": 0.0, "grad_norm": 2.306507110595703, "learning_rate": 9.989068136093872e-07, "loss": -0.0, "num_tokens": 1046317.0, "reward": 0.29374998807907104, "reward_std": 0.3920552134513855, "rewards/format_reward/mean": 0.4375, "rewards/format_reward/std": 0.2597312331199646, "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": 1424.0, "completions/max_terminated_length": 1424.0, "completions/mean_length": 420.34375, "completions/mean_terminated_length": 420.34375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.020089285714285716, "frac_reward_zero_std": 0.125, "grad_norm": 8.412493705749512, "learning_rate": 9.975414512725056e-07, "loss": -0.0, "num_tokens": 1160699.0, "reward": 0.25, "reward_std": 0.26024484634399414, "rewards/format_reward/mean": 0.3125, "rewards/format_reward/std": 0.2597312331199646, "rewards/mcq_exact_match_reward/mean": 0.21875, "rewards/mcq_exact_match_reward/std": 0.4166666865348816, "step": 9 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1864.0, "completions/mean_length": 541.34375, "completions/mean_terminated_length": 517.4285888671875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.022321428571428572, "frac_reward_zero_std": 0.125, "grad_norm": 6.991084098815918, "learning_rate": 9.956320346634875e-07, "loss": -0.0, "num_tokens": 1291769.0, "reward": 0.2632812261581421, "reward_std": 0.3168719708919525, "rewards/format_reward/mean": 0.4453125, "rewards/format_reward/std": 0.29703813791275024, "rewards/mcq_exact_match_reward/mean": 0.21875, "rewards/mcq_exact_match_reward/std": 0.4166666865348816, "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": 1373.0, "completions/max_terminated_length": 1373.0, "completions/mean_length": 499.09375, "completions/mean_terminated_length": 499.09375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.024553571428571428, "frac_reward_zero_std": 0.0, "grad_norm": 2.6405880451202393, "learning_rate": 9.931806517013612e-07, "loss": 0.0, "num_tokens": 1408655.0, "reward": 0.3023437261581421, "reward_std": 0.33734452724456787, "rewards/format_reward/mean": 0.3671875, "rewards/format_reward/std": 0.22257846593856812, "rewards/mcq_exact_match_reward/mean": 0.265625, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "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": 1313.0, "completions/mean_length": 516.921875, "completions/mean_terminated_length": 492.61907958984375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.026785714285714284, "frac_reward_zero_std": 0.0, "grad_norm": 7.266026496887207, "learning_rate": 9.901899829374047e-07, "loss": 0.0, "num_tokens": 1534194.0, "reward": 0.1953125, "reward_std": 0.23524019122123718, "rewards/format_reward/mean": 0.390625, "rewards/format_reward/std": 0.24346621334552765, "rewards/mcq_exact_match_reward/mean": 0.15625, "rewards/mcq_exact_match_reward/std": 0.36596253514289856, "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": 0.0625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1846.0, "completions/mean_length": 791.3125, "completions/mean_terminated_length": 707.5333862304688, "completions/min_length": 86.0, "completions/min_terminated_length": 86.0, "epoch": 0.029017857142857144, "frac_reward_zero_std": 0.0, "grad_norm": 1.0781476497650146, "learning_rate": 9.866632986240029e-07, "loss": -0.0, "num_tokens": 1705878.0, "reward": 0.38359373807907104, "reward_std": 0.3751716613769531, "rewards/format_reward/mean": 0.3984375, "rewards/format_reward/std": 0.22146137058734894, "rewards/mcq_exact_match_reward/mean": 0.34375, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "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": 0.046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1685.0, "completions/mean_length": 611.5625, "completions/mean_terminated_length": 540.9180297851562, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.03125, "frac_reward_zero_std": 0.125, "grad_norm": 0.9777091145515442, "learning_rate": 9.826044551386742e-07, "loss": -0.0, "num_tokens": 1850186.0, "reward": 0.19374999403953552, "reward_std": 0.2785091996192932, "rewards/format_reward/mean": 0.375, "rewards/format_reward/std": 0.29546841979026794, "rewards/mcq_exact_match_reward/mean": 0.15625, "rewards/mcq_exact_match_reward/std": 0.36596253514289856, "step": 14 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1601.0, "completions/mean_length": 556.671875, "completions/mean_terminated_length": 508.56451416015625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.033482142857142856, "frac_reward_zero_std": 0.125, "grad_norm": 16.6350154876709, "learning_rate": 9.780178907671788e-07, "loss": -0.0, "num_tokens": 1988325.0, "reward": 0.4648437201976776, "reward_std": 0.35466837882995605, "rewards/format_reward/mean": 0.4296875, "rewards/format_reward/std": 0.1751912236213684, "rewards/mcq_exact_match_reward/mean": 0.421875, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "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": 1539.0, "completions/max_terminated_length": 1539.0, "completions/mean_length": 446.140625, "completions/mean_terminated_length": 446.140625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.03571428571428571, "frac_reward_zero_std": 0.0, "grad_norm": 37.51299285888672, "learning_rate": 9.729086208503173e-07, "loss": 0.0, "num_tokens": 2101254.0, "reward": 0.2906249761581421, "reward_std": 0.2404833734035492, "rewards/format_reward/mean": 0.40625, "rewards/format_reward/std": 0.265398770570755, "rewards/mcq_exact_match_reward/mean": 0.25, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1922.0, "completions/mean_length": 505.484375, "completions/mean_terminated_length": 455.7257995605469, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.03794642857142857, "frac_reward_zero_std": 0.0, "grad_norm": 11.994452476501465, "learning_rate": 9.672822322997304e-07, "loss": -0.0, "num_tokens": 2219821.0, "reward": 0.2242187261581421, "reward_std": 0.33446377515792847, "rewards/format_reward/mean": 0.3671875, "rewards/format_reward/std": 0.23974503576755524, "rewards/mcq_exact_match_reward/mean": 0.1875, "rewards/mcq_exact_match_reward/std": 0.39339789748191833, "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": 1556.0, "completions/max_terminated_length": 1556.0, "completions/mean_length": 648.90625, "completions/mean_terminated_length": 648.90625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.04017857142857143, "frac_reward_zero_std": 0.125, "grad_norm": 2.126474380493164, "learning_rate": 9.611448774886923e-07, "loss": 0.0, "num_tokens": 2351479.0, "reward": 0.21406248211860657, "reward_std": 0.3205658793449402, "rewards/format_reward/mean": 0.421875, "rewards/format_reward/std": 0.20351573824882507, "rewards/mcq_exact_match_reward/mean": 0.171875, "rewards/mcq_exact_match_reward/std": 0.38025420904159546, "step": 18 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1170.0, "completions/mean_length": 566.203125, "completions/mean_terminated_length": 542.6825561523438, "completions/min_length": 78.0, "completions/min_terminated_length": 78.0, "epoch": 0.04241071428571429, "frac_reward_zero_std": 0.0, "grad_norm": 1.6903095245361328, "learning_rate": 9.545032675245813e-07, "loss": 0.0, "num_tokens": 2497644.0, "reward": 0.26171875, "reward_std": 0.30634066462516785, "rewards/format_reward/mean": 0.4296875, "rewards/format_reward/std": 0.1751912236213684, "rewards/mcq_exact_match_reward/mean": 0.21875, "rewards/mcq_exact_match_reward/std": 0.4166666865348816, "step": 19 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1336.0, "completions/max_terminated_length": 1336.0, "completions/mean_length": 596.484375, "completions/mean_terminated_length": 596.484375, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.044642857142857144, "frac_reward_zero_std": 0.125, "grad_norm": 0.9298074245452881, "learning_rate": 9.473646649103817e-07, "loss": -0.0, "num_tokens": 2626363.0, "reward": 0.4218749701976776, "reward_std": 0.3645807206630707, "rewards/format_reward/mean": 0.46875, "rewards/format_reward/std": 0.17536810040473938, "rewards/mcq_exact_match_reward/mean": 0.375, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "step": 20 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1458.0, "completions/mean_length": 498.96875, "completions/mean_terminated_length": 474.3809814453125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.046875, "frac_reward_zero_std": 0.0, "grad_norm": 15.798802375793457, "learning_rate": 9.397368756032444e-07, "loss": -0.0, "num_tokens": 2753089.0, "reward": 0.4921874701976776, "reward_std": 0.4445301294326782, "rewards/format_reward/mean": 0.390625, "rewards/format_reward/std": 0.2083333432674408, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "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": 1528.0, "completions/max_terminated_length": 1528.0, "completions/mean_length": 509.328125, "completions/mean_terminated_length": 509.328125, "completions/min_length": 129.0, "completions/min_terminated_length": 129.0, "epoch": 0.049107142857142856, "frac_reward_zero_std": 0.125, "grad_norm": 1.0354143381118774, "learning_rate": 9.316282404787869e-07, "loss": 0.0, "num_tokens": 2861742.0, "reward": 0.24375000596046448, "reward_std": 0.24846376478672028, "rewards/format_reward/mean": 0.40625, "rewards/format_reward/std": 0.21593283116817474, "rewards/mcq_exact_match_reward/mean": 0.203125, "rewards/mcq_exact_match_reward/std": 0.40550529956817627, "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": 1469.0, "completions/max_terminated_length": 1469.0, "completions/mean_length": 567.5625, "completions/mean_terminated_length": 567.5625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.05133928571428571, "frac_reward_zero_std": 0.0, "grad_norm": 2.768810987472534, "learning_rate": 9.230476262104676e-07, "loss": -0.0, "num_tokens": 2997922.0, "reward": 0.5914062261581421, "reward_std": 0.39294227957725525, "rewards/format_reward/mean": 0.4453125, "rewards/format_reward/std": 0.15728822350502014, "rewards/mcq_exact_match_reward/mean": 0.546875, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "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": 1711.0, "completions/max_terminated_length": 1711.0, "completions/mean_length": 487.640625, "completions/mean_terminated_length": 487.640625, "completions/min_length": 43.0, "completions/min_terminated_length": 43.0, "epoch": 0.05357142857142857, "frac_reward_zero_std": 0.0, "grad_norm": 1.757157325744629, "learning_rate": 9.1400441557401e-07, "loss": -0.0, "num_tokens": 3113179.0, "reward": 0.44609373807907104, "reward_std": 0.43148568272590637, "rewards/format_reward/mean": 0.3984375, "rewards/format_reward/std": 0.20275264978408813, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1297.0, "completions/mean_length": 595.109375, "completions/mean_terminated_length": 572.0476684570312, "completions/min_length": 65.0, "completions/min_terminated_length": 65.0, "epoch": 0.05580357142857143, "frac_reward_zero_std": 0.0, "grad_norm": 1.4907280206680298, "learning_rate": 9.045084971874737e-07, "loss": -0.0, "num_tokens": 3237106.0, "reward": 0.2109374850988388, "reward_std": 0.36231935024261475, "rewards/format_reward/mean": 0.390625, "rewards/format_reward/std": 0.2741328477859497, "rewards/mcq_exact_match_reward/mean": 0.171875, "rewards/mcq_exact_match_reward/std": 0.38025420904159546, "step": 25 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1619.0, "completions/mean_length": 613.390625, "completions/mean_terminated_length": 590.6190795898438, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.05803571428571429, "frac_reward_zero_std": 0.0, "grad_norm": 2.7617568969726562, "learning_rate": 8.945702546981968e-07, "loss": -0.0, "num_tokens": 3367451.0, "reward": 0.26249998807907104, "reward_std": 0.39707860350608826, "rewards/format_reward/mean": 0.4375, "rewards/format_reward/std": 0.22712838649749756, "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": 1618.0, "completions/max_terminated_length": 1618.0, "completions/mean_length": 440.578125, "completions/mean_terminated_length": 440.578125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.060267857142857144, "frac_reward_zero_std": 0.125, "grad_norm": 7.312154769897461, "learning_rate": 8.842005554284295e-07, "loss": -0.0, "num_tokens": 3484664.0, "reward": 0.4859374761581421, "reward_std": 0.3957879841327667, "rewards/format_reward/mean": 0.484375, "rewards/format_reward/std": 0.21764887869358063, "rewards/mcq_exact_match_reward/mean": 0.4375, "rewards/mcq_exact_match_reward/std": 0.5, "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": 0.078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1915.0, "completions/mean_length": 760.609375, "completions/mean_terminated_length": 651.5084838867188, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0625, "frac_reward_zero_std": 0.125, "grad_norm": 7.172023773193359, "learning_rate": 8.734107384920769e-07, "loss": 0.0, "num_tokens": 3646391.0, "reward": 0.16640624403953552, "reward_std": 0.23528708517551422, "rewards/format_reward/mean": 0.4140625, "rewards/format_reward/std": 0.19012710452079773, "rewards/mcq_exact_match_reward/mean": 0.125, "rewards/mcq_exact_match_reward/std": 0.3333333432674408, "step": 28 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1662.0, "completions/mean_length": 513.109375, "completions/mean_terminated_length": 488.7460632324219, "completions/min_length": 56.0, "completions/min_terminated_length": 56.0, "epoch": 0.06473214285714286, "frac_reward_zero_std": 0.0, "grad_norm": 1.6719025373458862, "learning_rate": 8.622126023955445e-07, "loss": -0.0, "num_tokens": 3753398.0, "reward": 0.5648437142372131, "reward_std": 0.48218396306037903, "rewards/format_reward/mean": 0.4921875, "rewards/format_reward/std": 0.22699186205863953, "rewards/mcq_exact_match_reward/mean": 0.515625, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1655.0, "completions/mean_length": 585.90625, "completions/mean_terminated_length": 562.6984252929688, "completions/min_length": 14.0, "completions/min_terminated_length": 14.0, "epoch": 0.06696428571428571, "frac_reward_zero_std": 0.0, "grad_norm": 2.9140756130218506, "learning_rate": 8.506183921362442e-07, "loss": -0.0, "num_tokens": 3895000.0, "reward": 0.4468749761581421, "reward_std": 0.5032121539115906, "rewards/format_reward/mean": 0.40625, "rewards/format_reward/std": 0.19669894874095917, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "step": 30 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1637.0, "completions/mean_length": 623.65625, "completions/mean_terminated_length": 601.0476684570312, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.06919642857142858, "frac_reward_zero_std": 0.125, "grad_norm": 5.2398505210876465, "learning_rate": 8.386407858128706e-07, "loss": -0.0, "num_tokens": 4029362.0, "reward": 0.43203121423721313, "reward_std": 0.43715861439704895, "rewards/format_reward/mean": 0.4140625, "rewards/format_reward/std": 0.19012710452079773, "rewards/mcq_exact_match_reward/mean": 0.390625, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1342.0, "completions/mean_length": 459.90625, "completions/mean_terminated_length": 434.69842529296875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.07142857142857142, "frac_reward_zero_std": 0.0, "grad_norm": 10.139559745788574, "learning_rate": 8.262928807620843e-07, "loss": -0.0, "num_tokens": 4139252.0, "reward": 0.5257812142372131, "reward_std": 0.48261502385139465, "rewards/format_reward/mean": 0.4140625, "rewards/format_reward/std": 0.244862899184227, "rewards/mcq_exact_match_reward/mean": 0.484375, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "step": 32 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1322.0, "completions/mean_length": 512.953125, "completions/mean_terminated_length": 488.58734130859375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.07366071428571429, "frac_reward_zero_std": 0.0, "grad_norm": 6.937158584594727, "learning_rate": 8.135881792367685e-07, "loss": -0.0, "num_tokens": 4259337.0, "reward": 0.4195312261581421, "reward_std": 0.3889823853969574, "rewards/format_reward/mean": 0.4453125, "rewards/format_reward/std": 0.20152568817138672, "rewards/mcq_exact_match_reward/mean": 0.375, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "step": 33 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1483.0, "completions/mean_length": 463.875, "completions/mean_terminated_length": 412.774169921875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.07589285714285714, "frac_reward_zero_std": 0.25, "grad_norm": 3.2083284854888916, "learning_rate": 8.005405736415125e-07, "loss": 0.0, "num_tokens": 4373121.0, "reward": 0.33984375, "reward_std": 0.3635759949684143, "rewards/format_reward/mean": 0.4296875, "rewards/format_reward/std": 0.23345555365085602, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "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": 1703.0, "completions/max_terminated_length": 1703.0, "completions/mean_length": 505.0625, "completions/mean_terminated_length": 505.0625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.078125, "frac_reward_zero_std": 0.25, "grad_norm": 7.500360488891602, "learning_rate": 7.871643313414718e-07, "loss": -0.0, "num_tokens": 4498525.0, "reward": 0.6585937142372131, "reward_std": 0.27997076511383057, "rewards/format_reward/mean": 0.4921875, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.609375, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1515.0, "completions/mean_length": 397.265625, "completions/mean_terminated_length": 371.0635070800781, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08035714285714286, "frac_reward_zero_std": 0.125, "grad_norm": 27.094533920288086, "learning_rate": 7.734740790612136e-07, "loss": -0.0, "num_tokens": 4620358.0, "reward": 0.5789061784744263, "reward_std": 0.367202490568161, "rewards/format_reward/mean": 0.4765625, "rewards/format_reward/std": 0.22589658200740814, "rewards/mcq_exact_match_reward/mean": 0.53125, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "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": 1575.0, "completions/max_terminated_length": 1575.0, "completions/mean_length": 446.046875, "completions/mean_terminated_length": 446.046875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08258928571428571, "frac_reward_zero_std": 0.125, "grad_norm": 13.872968673706055, "learning_rate": 7.594847868906076e-07, "loss": -0.0, "num_tokens": 4751177.0, "reward": 0.5031249523162842, "reward_std": 0.3504549264907837, "rewards/format_reward/mean": 0.5, "rewards/format_reward/std": 0.2182178944349289, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 37 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1675.0, "completions/max_terminated_length": 1675.0, "completions/mean_length": 373.8125, "completions/mean_terminated_length": 373.8125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08482142857142858, "frac_reward_zero_std": 0.25, "grad_norm": 9.39517593383789, "learning_rate": 7.452117519152541e-07, "loss": 0.0, "num_tokens": 4875565.0, "reward": 0.3593749701976776, "reward_std": 0.25985032320022583, "rewards/format_reward/mean": 0.46875, "rewards/format_reward/std": 0.12198751419782639, "rewards/mcq_exact_match_reward/mean": 0.3125, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "step": 38 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1549.0, "completions/mean_length": 497.59375, "completions/mean_terminated_length": 447.58062744140625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08705357142857142, "frac_reward_zero_std": 0.125, "grad_norm": 13.764936447143555, "learning_rate": 7.306705814893439e-07, "loss": -0.0, "num_tokens": 4989059.0, "reward": 0.34453123807907104, "reward_std": 0.3753530979156494, "rewards/format_reward/mean": 0.4765625, "rewards/format_reward/std": 0.13886408507823944, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "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": 598.0, "completions/max_terminated_length": 598.0, "completions/mean_length": 119.890625, "completions/mean_terminated_length": 119.890625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08928571428571429, "frac_reward_zero_std": 0.5, "grad_norm": 17.380779266357422, "learning_rate": 7.158771761692464e-07, "loss": -0.0, "num_tokens": 5068732.0, "reward": 0.8320311903953552, "reward_std": 0.20367416739463806, "rewards/format_reward/mean": 0.5078125, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.78125, "rewards/mcq_exact_match_reward/std": 0.4166666865348816, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1656.0, "completions/mean_length": 269.6875, "completions/mean_terminated_length": 212.32257080078125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.09151785714285714, "frac_reward_zero_std": 0.125, "grad_norm": 23.273365020751953, "learning_rate": 7.008477123264847e-07, "loss": -0.0, "num_tokens": 5191240.0, "reward": 0.49843746423721313, "reward_std": 0.3545449376106262, "rewards/format_reward/mean": 0.453125, "rewards/format_reward/std": 0.14689241349697113, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "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": 1538.0, "completions/max_terminated_length": 1538.0, "completions/mean_length": 99.0625, "completions/mean_terminated_length": 99.0625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.09375, "frac_reward_zero_std": 0.5, "grad_norm": 14.988553047180176, "learning_rate": 6.855986244591103e-07, "loss": 0.0, "num_tokens": 5287508.0, "reward": 0.4281249940395355, "reward_std": 0.22659148275852203, "rewards/format_reward/mean": 0.53125, "rewards/format_reward/std": 0.1510545015335083, "rewards/mcq_exact_match_reward/mean": 0.375, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 747.0, "completions/max_terminated_length": 747.0, "completions/mean_length": 122.375, "completions/mean_terminated_length": 122.375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.09598214285714286, "frac_reward_zero_std": 0.25, "grad_norm": 19.05467414855957, "learning_rate": 6.701465872208216e-07, "loss": 0.0, "num_tokens": 5390508.0, "reward": 0.5523437261581421, "reward_std": 0.26811331510543823, "rewards/format_reward/mean": 0.5234375, "rewards/format_reward/std": 0.13886408507823944, "rewards/mcq_exact_match_reward/mean": 0.5, "rewards/mcq_exact_match_reward/std": 0.5039526224136353, "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": 1212.0, "completions/max_terminated_length": 1212.0, "completions/mean_length": 106.9375, "completions/mean_terminated_length": 106.9375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.09821428571428571, "frac_reward_zero_std": 0.375, "grad_norm": 19.656740188598633, "learning_rate": 6.545084971874736e-07, "loss": 0.0, "num_tokens": 5487192.0, "reward": 0.6929687261581421, "reward_std": 0.24791689217090607, "rewards/format_reward/mean": 0.5234375, "rewards/format_reward/std": 0.10652101784944534, "rewards/mcq_exact_match_reward/mean": 0.640625, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "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": 1823.0, "completions/max_terminated_length": 1823.0, "completions/mean_length": 85.421875, "completions/mean_terminated_length": 85.421875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.10044642857142858, "frac_reward_zero_std": 0.375, "grad_norm": 42.01066589355469, "learning_rate": 6.387014543809223e-07, "loss": 0.0, "num_tokens": 5582515.0, "reward": 0.5421874523162842, "reward_std": 0.14177924394607544, "rewards/format_reward/mean": 0.578125, "rewards/format_reward/std": 0.20351573824882507, "rewards/mcq_exact_match_reward/mean": 0.484375, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "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": 385.0, "completions/max_terminated_length": 385.0, "completions/mean_length": 25.359375, "completions/mean_terminated_length": 25.359375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.10267857142857142, "frac_reward_zero_std": 0.375, "grad_norm": 34.89189910888672, "learning_rate": 6.227427435703995e-07, "loss": -0.0, "num_tokens": 5686162.0, "reward": 0.4921875, "reward_std": 0.3141977787017822, "rewards/format_reward/mean": 0.546875, "rewards/format_reward/std": 0.1717960685491562, "rewards/mcq_exact_match_reward/mean": 0.4375, "rewards/mcq_exact_match_reward/std": 0.5, "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": 389.0, "completions/max_terminated_length": 389.0, "completions/mean_length": 15.65625, "completions/mean_terminated_length": 15.65625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.10491071428571429, "frac_reward_zero_std": 0.5, "grad_norm": 19.081090927124023, "learning_rate": 6.066498153718734e-07, "loss": -0.0, "num_tokens": 5788092.0, "reward": 0.53515625, "reward_std": 0.1901850551366806, "rewards/format_reward/mean": 0.5078125, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.484375, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "step": 47 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 512.0, "completions/max_terminated_length": 512.0, "completions/mean_length": 17.0, "completions/mean_terminated_length": 17.0, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.10714285714285714, "frac_reward_zero_std": 0.375, "grad_norm": 33.182762145996094, "learning_rate": 5.90440267166055e-07, "loss": 0.0, "num_tokens": 5880436.0, "reward": 0.50390625, "reward_std": 0.2418397068977356, "rewards/format_reward/mean": 0.5078125, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 48 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 495.0, "completions/max_terminated_length": 495.0, "completions/mean_length": 49.828125, "completions/mean_terminated_length": 49.828125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.109375, "frac_reward_zero_std": 0.75, "grad_norm": 18.025070190429688, "learning_rate": 5.741318238559209e-07, "loss": 0.0, "num_tokens": 5985793.0, "reward": 0.28828126192092896, "reward_std": 0.04814022406935692, "rewards/format_reward/mean": 0.5390625, "rewards/format_reward/std": 0.13524486124515533, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "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": 418.0, "completions/max_terminated_length": 418.0, "completions/mean_length": 21.984375, "completions/mean_terminated_length": 21.984375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.11160714285714286, "frac_reward_zero_std": 0.375, "grad_norm": 24.160545349121094, "learning_rate": 5.577423184847931e-07, "loss": 0.0, "num_tokens": 6082544.0, "reward": 0.5703124403953552, "reward_std": 0.21383893489837646, "rewards/format_reward/mean": 0.546875, "rewards/format_reward/std": 0.14689241349697113, "rewards/mcq_exact_match_reward/mean": 0.515625, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "step": 50 } ], "logging_steps": 1, "max_steps": 100, "num_input_tokens_seen": 6082544, "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 }