{ "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": 0.0, "completions/max_length": 1861.0, "completions/max_terminated_length": 1861.0, "completions/mean_length": 598.625, "completions/mean_terminated_length": 598.625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.002232142857142857, "frac_reward_zero_std": 0.0, "grad_norm": 5.51198148727417, "learning_rate": 0.0, "loss": -0.0, "num_tokens": 132096.0, "reward": 0.2382812350988388, "reward_std": 0.3009326457977295, "rewards/format_reward/mean": 0.3515625, "rewards/format_reward/std": 0.24688033759593964, "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": 1602.0, "completions/max_terminated_length": 1602.0, "completions/mean_length": 451.15625, "completions/mean_terminated_length": 451.15625, "completions/min_length": 10.0, "completions/min_terminated_length": 10.0, "epoch": 0.004464285714285714, "frac_reward_zero_std": 0.0, "grad_norm": 2.7671356201171875, "learning_rate": 2e-07, "loss": 0.0, "num_tokens": 255170.0, "reward": 0.12656249105930328, "reward_std": 0.21595951914787292, "rewards/format_reward/mean": 0.328125, "rewards/format_reward/std": 0.27048972249031067, "rewards/mcq_exact_match_reward/mean": 0.09375, "rewards/mcq_exact_match_reward/std": 0.29378482699394226, "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": 1414.0, "completions/max_terminated_length": 1414.0, "completions/mean_length": 500.234375, "completions/mean_terminated_length": 500.234375, "completions/min_length": 50.0, "completions/min_terminated_length": 50.0, "epoch": 0.006696428571428571, "frac_reward_zero_std": 0.0, "grad_norm": 1.8978508710861206, "learning_rate": 4e-07, "loss": 0.0, "num_tokens": 369961.0, "reward": 0.23906248807907104, "reward_std": 0.29380542039871216, "rewards/format_reward/mean": 0.359375, "rewards/format_reward/std": 0.2592533528804779, "rewards/mcq_exact_match_reward/mean": 0.203125, "rewards/mcq_exact_match_reward/std": 0.40550529956817627, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2026.0, "completions/mean_length": 617.234375, "completions/mean_terminated_length": 594.5238647460938, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "epoch": 0.008928571428571428, "frac_reward_zero_std": 0.0, "grad_norm": 5.240273475646973, "learning_rate": 6e-07, "loss": -0.0, "num_tokens": 517816.0, "reward": 0.3828124701976776, "reward_std": 0.3759346902370453, "rewards/format_reward/mean": 0.390625, "rewards/format_reward/std": 0.22658175230026245, "rewards/mcq_exact_match_reward/mean": 0.34375, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "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": 1070.0, "completions/max_terminated_length": 1070.0, "completions/mean_length": 333.421875, "completions/mean_terminated_length": 333.421875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.011160714285714286, "frac_reward_zero_std": 0.0, "grad_norm": 27.351428985595703, "learning_rate": 8e-07, "loss": 0.0, "num_tokens": 617451.0, "reward": 0.2773437201976776, "reward_std": 0.28845131397247314, "rewards/format_reward/mean": 0.4296875, "rewards/format_reward/std": 0.23345555365085602, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "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": 2006.0, "completions/mean_length": 668.578125, "completions/mean_terminated_length": 600.7376708984375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.013392857142857142, "frac_reward_zero_std": 0.125, "grad_norm": 2.6046884059906006, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 768640.0, "reward": 0.43359375, "reward_std": 0.3082636892795563, "rewards/format_reward/mean": 0.4296875, "rewards/format_reward/std": 0.19654127955436707, "rewards/mcq_exact_match_reward/mean": 0.390625, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "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": 1726.0, "completions/mean_length": 595.765625, "completions/mean_terminated_length": 572.7142944335938, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.015625, "frac_reward_zero_std": 0.0, "grad_norm": 4.04109525680542, "learning_rate": 9.99726628670463e-07, "loss": 0.0, "num_tokens": 910537.0, "reward": 0.2476562261581421, "reward_std": 0.2781180143356323, "rewards/format_reward/mean": 0.4453125, "rewards/format_reward/std": 0.2538151443004608, "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": 0.0625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1797.0, "completions/mean_length": 589.15625, "completions/mean_terminated_length": 491.9000244140625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.017857142857142856, "frac_reward_zero_std": 0.0, "grad_norm": 3.8270058631896973, "learning_rate": 9.989068136093872e-07, "loss": -0.0, "num_tokens": 1070555.0, "reward": 0.25312498211860657, "reward_std": 0.27621039748191833, "rewards/format_reward/mean": 0.34375, "rewards/format_reward/std": 0.29378482699394226, "rewards/mcq_exact_match_reward/mean": 0.21875, "rewards/mcq_exact_match_reward/std": 0.4166666865348816, "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": 1734.0, "completions/max_terminated_length": 1734.0, "completions/mean_length": 544.46875, "completions/mean_terminated_length": 544.46875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.020089285714285716, "frac_reward_zero_std": 0.0, "grad_norm": 2.9913084506988525, "learning_rate": 9.975414512725056e-07, "loss": 0.0, "num_tokens": 1181377.0, "reward": 0.2601562440395355, "reward_std": 0.32211291790008545, "rewards/format_reward/mean": 0.4140625, "rewards/format_reward/std": 0.27537402510643005, "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": 1861.0, "completions/mean_length": 516.671875, "completions/mean_terminated_length": 492.3651123046875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.022321428571428572, "frac_reward_zero_std": 0.125, "grad_norm": 6.775735855102539, "learning_rate": 9.956320346634875e-07, "loss": 0.0, "num_tokens": 1297596.0, "reward": 0.3382812440395355, "reward_std": 0.30042457580566406, "rewards/format_reward/mean": 0.4140625, "rewards/format_reward/std": 0.2605654299259186, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1050.0, "completions/mean_length": 292.4375, "completions/mean_terminated_length": 264.5714416503906, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.024553571428571428, "frac_reward_zero_std": 0.0, "grad_norm": 35.66353988647461, "learning_rate": 9.931806517013612e-07, "loss": 0.0, "num_tokens": 1417896.0, "reward": 0.42578125, "reward_std": 0.396813303232193, "rewards/format_reward/mean": 0.3515625, "rewards/format_reward/std": 0.26246222853660583, "rewards/mcq_exact_match_reward/mean": 0.390625, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "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": 0.0, "completions/max_length": 1636.0, "completions/max_terminated_length": 1636.0, "completions/mean_length": 499.890625, "completions/mean_terminated_length": 499.890625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.026785714285714284, "frac_reward_zero_std": 0.0, "grad_norm": 6.420288562774658, "learning_rate": 9.901899829374047e-07, "loss": -0.0, "num_tokens": 1551073.0, "reward": 0.27812501788139343, "reward_std": 0.31929129362106323, "rewards/format_reward/mean": 0.4375, "rewards/format_reward/std": 0.32732683420181274, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "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": 1830.0, "completions/mean_length": 580.359375, "completions/mean_terminated_length": 557.0635375976562, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.029017857142857144, "frac_reward_zero_std": 0.0, "grad_norm": 9.754948616027832, "learning_rate": 9.866632986240029e-07, "loss": -0.0, "num_tokens": 1715992.0, "reward": 0.34999996423721313, "reward_std": 0.38676393032073975, "rewards/format_reward/mean": 0.375, "rewards/format_reward/std": 0.2357022762298584, "rewards/mcq_exact_match_reward/mean": 0.3125, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1502.0, "completions/mean_length": 559.140625, "completions/mean_terminated_length": 535.5079956054688, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.03125, "frac_reward_zero_std": 0.0, "grad_norm": 4.196605205535889, "learning_rate": 9.826044551386742e-07, "loss": 0.0, "num_tokens": 1870537.0, "reward": 0.33046871423721313, "reward_std": 0.3549173176288605, "rewards/format_reward/mean": 0.4921875, "rewards/format_reward/std": 0.2744719088077545, "rewards/mcq_exact_match_reward/mean": 0.28125, "rewards/mcq_exact_match_reward/std": 0.4531635046005249, "step": 14 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1548.0, "completions/mean_length": 591.09375, "completions/mean_terminated_length": 519.4425659179688, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "epoch": 0.033482142857142856, "frac_reward_zero_std": 0.0, "grad_norm": 5.3005290031433105, "learning_rate": 9.780178907671788e-07, "loss": -0.0, "num_tokens": 1995175.0, "reward": 0.36562496423721313, "reward_std": 0.353937029838562, "rewards/format_reward/mean": 0.375, "rewards/format_reward/std": 0.30860671401023865, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "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": 1425.0, "completions/max_terminated_length": 1425.0, "completions/mean_length": 448.15625, "completions/mean_terminated_length": 448.15625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.03571428571428571, "frac_reward_zero_std": 0.125, "grad_norm": 7.560476779937744, "learning_rate": 9.729086208503173e-07, "loss": -0.0, "num_tokens": 2117017.0, "reward": 0.3265624940395355, "reward_std": 0.2529877722263336, "rewards/format_reward/mean": 0.453125, "rewards/format_reward/std": 0.2777281701564789, "rewards/mcq_exact_match_reward/mean": 0.28125, "rewards/mcq_exact_match_reward/std": 0.4531635046005249, "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": 1902.0, "completions/max_terminated_length": 1902.0, "completions/mean_length": 486.375, "completions/mean_terminated_length": 486.375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.03794642857142857, "frac_reward_zero_std": 0.125, "grad_norm": 26.575090408325195, "learning_rate": 9.672822322997304e-07, "loss": -0.0, "num_tokens": 2241897.0, "reward": 0.48359373211860657, "reward_std": 0.3336434066295624, "rewards/format_reward/mean": 0.4609375, "rewards/format_reward/std": 0.2236899733543396, "rewards/mcq_exact_match_reward/mean": 0.4375, "rewards/mcq_exact_match_reward/std": 0.5, "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": 0.046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1314.0, "completions/mean_length": 532.515625, "completions/mean_terminated_length": 457.98358154296875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.04017857142857143, "frac_reward_zero_std": 0.125, "grad_norm": 4.496442794799805, "learning_rate": 9.611448774886923e-07, "loss": -0.0, "num_tokens": 2377914.0, "reward": 0.3531249761581421, "reward_std": 0.292488157749176, "rewards/format_reward/mean": 0.40625, "rewards/format_reward/std": 0.25, "rewards/mcq_exact_match_reward/mean": 0.3125, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "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": 1445.0, "completions/mean_length": 502.09375, "completions/mean_terminated_length": 477.5555725097656, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.04241071428571429, "frac_reward_zero_std": 0.0, "grad_norm": 18.192575454711914, "learning_rate": 9.545032675245813e-07, "loss": -0.0, "num_tokens": 2537832.0, "reward": 0.20468749105930328, "reward_std": 0.2482740581035614, "rewards/format_reward/mean": 0.484375, "rewards/format_reward/std": 0.23517554998397827, "rewards/mcq_exact_match_reward/mean": 0.15625, "rewards/mcq_exact_match_reward/std": 0.36596253514289856, "step": 19 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1414.0, "completions/mean_length": 519.40625, "completions/mean_terminated_length": 495.14288330078125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.044642857142857144, "frac_reward_zero_std": 0.0, "grad_norm": 20.479293823242188, "learning_rate": 9.473646649103817e-07, "loss": -0.0, "num_tokens": 2650002.0, "reward": 0.2382812350988388, "reward_std": 0.29763028025627136, "rewards/format_reward/mean": 0.5078125, "rewards/format_reward/std": 0.32723212242126465, "rewards/mcq_exact_match_reward/mean": 0.1875, "rewards/mcq_exact_match_reward/std": 0.39339789748191833, "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": 0.078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1151.0, "completions/mean_length": 530.015625, "completions/mean_terminated_length": 401.3728942871094, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.046875, "frac_reward_zero_std": 0.0, "grad_norm": 9.621838569641113, "learning_rate": 9.397368756032444e-07, "loss": -0.0, "num_tokens": 2780811.0, "reward": 0.551562488079071, "reward_std": 0.32089635729789734, "rewards/format_reward/mean": 0.671875, "rewards/format_reward/std": 0.33592742681503296, "rewards/mcq_exact_match_reward/mean": 0.484375, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1485.0, "completions/mean_length": 437.828125, "completions/mean_terminated_length": 412.2698669433594, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.049107142857142856, "frac_reward_zero_std": 0.125, "grad_norm": 4.107235908508301, "learning_rate": 9.316282404787869e-07, "loss": 0.0, "num_tokens": 2929728.0, "reward": 0.21250000596046448, "reward_std": 0.23144376277923584, "rewards/format_reward/mean": 0.5625, "rewards/format_reward/std": 0.2597312331199646, "rewards/mcq_exact_match_reward/mean": 0.15625, "rewards/mcq_exact_match_reward/std": 0.36596253514289856, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1956.0, "completions/mean_length": 299.296875, "completions/mean_terminated_length": 271.5397033691406, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.05133928571428571, "frac_reward_zero_std": 0.0, "grad_norm": 20.197181701660156, "learning_rate": 9.230476262104676e-07, "loss": -0.0, "num_tokens": 3044243.0, "reward": 0.6476562023162842, "reward_std": 0.4422643184661865, "rewards/format_reward/mean": 0.5390625, "rewards/format_reward/std": 0.2236899733543396, "rewards/mcq_exact_match_reward/mean": 0.59375, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "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": 682.0, "completions/max_terminated_length": 682.0, "completions/mean_length": 150.328125, "completions/mean_terminated_length": 150.328125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.05357142857142857, "frac_reward_zero_std": 0.0, "grad_norm": 19.86847686767578, "learning_rate": 9.1400441557401e-07, "loss": 0.0, "num_tokens": 3135328.0, "reward": 0.4632812440395355, "reward_std": 0.3330904245376587, "rewards/format_reward/mean": 0.5703125, "rewards/format_reward/std": 0.33174848556518555, "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": 0.046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1512.0, "completions/mean_length": 337.78125, "completions/mean_terminated_length": 253.672119140625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.05580357142857143, "frac_reward_zero_std": 0.375, "grad_norm": 10.876846313476562, "learning_rate": 9.045084971874737e-07, "loss": -0.0, "num_tokens": 3280986.0, "reward": 0.16249999403953552, "reward_std": 0.13564997911453247, "rewards/format_reward/mean": 0.53125, "rewards/format_reward/std": 0.21593283116817474, "rewards/mcq_exact_match_reward/mean": 0.109375, "rewards/mcq_exact_match_reward/std": 0.3145764470100403, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1667.0, "completions/mean_length": 323.828125, "completions/mean_terminated_length": 268.20965576171875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.05803571428571429, "frac_reward_zero_std": 0.5, "grad_norm": 17.0075626373291, "learning_rate": 8.945702546981968e-07, "loss": -0.0, "num_tokens": 3417959.0, "reward": 0.24687498807907104, "reward_std": 0.20000335574150085, "rewards/format_reward/mean": 0.59375, "rewards/format_reward/std": 0.25, "rewards/mcq_exact_match_reward/mean": 0.1875, "rewards/mcq_exact_match_reward/std": 0.39339789748191833, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 967.0, "completions/mean_length": 357.65625, "completions/mean_terminated_length": 303.1290283203125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.060267857142857144, "frac_reward_zero_std": 0.125, "grad_norm": 9.335482597351074, "learning_rate": 8.842005554284295e-07, "loss": -0.0, "num_tokens": 3536609.0, "reward": 0.3062499761581421, "reward_std": 0.35502874851226807, "rewards/format_reward/mean": 0.5625, "rewards/format_reward/std": 0.2597312331199646, "rewards/mcq_exact_match_reward/mean": 0.25, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1430.0, "completions/mean_length": 122.0625, "completions/mean_terminated_length": 91.4920654296875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0625, "frac_reward_zero_std": 0.25, "grad_norm": 21.517492294311523, "learning_rate": 8.734107384920769e-07, "loss": 0.0, "num_tokens": 3662717.0, "reward": 0.32499998807907104, "reward_std": 0.2159128338098526, "rewards/format_reward/mean": 0.59375, "rewards/format_reward/std": 0.233588308095932, "rewards/mcq_exact_match_reward/mean": 0.265625, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "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": 961.0, "completions/mean_length": 194.609375, "completions/mean_terminated_length": 165.19049072265625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.06473214285714286, "frac_reward_zero_std": 0.25, "grad_norm": 18.38661766052246, "learning_rate": 8.622126023955445e-07, "loss": 0.0, "num_tokens": 3803820.0, "reward": 0.2890625, "reward_std": 0.16094207763671875, "rewards/format_reward/mean": 0.546875, "rewards/format_reward/std": 0.24750742316246033, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "step": 29 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1152.0, "completions/max_terminated_length": 1152.0, "completions/mean_length": 94.328125, "completions/mean_terminated_length": 94.328125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.06696428571428571, "frac_reward_zero_std": 0.125, "grad_norm": 32.013633728027344, "learning_rate": 8.506183921362442e-07, "loss": -0.0, "num_tokens": 3906873.0, "reward": 0.20546874403953552, "reward_std": 0.19831590354442596, "rewards/format_reward/mean": 0.6484375, "rewards/format_reward/std": 0.2302463799715042, "rewards/mcq_exact_match_reward/mean": 0.140625, "rewards/mcq_exact_match_reward/std": 0.3503824472427368, "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": 1461.0, "completions/max_terminated_length": 1461.0, "completions/mean_length": 105.5625, "completions/mean_terminated_length": 105.5625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.06919642857142858, "frac_reward_zero_std": 0.25, "grad_norm": 16.25855255126953, "learning_rate": 8.386407858128706e-07, "loss": -0.0, "num_tokens": 4015637.0, "reward": 0.2671875059604645, "reward_std": 0.25722211599349976, "rewards/format_reward/mean": 0.640625, "rewards/format_reward/std": 0.22658175230026245, "rewards/mcq_exact_match_reward/mean": 0.203125, "rewards/mcq_exact_match_reward/std": 0.40550529956817627, "step": 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": 796.0, "completions/mean_length": 69.03125, "completions/mean_terminated_length": 37.619049072265625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.07142857142857142, "frac_reward_zero_std": 0.5, "grad_norm": 14.018732070922852, "learning_rate": 8.262928807620843e-07, "loss": -0.0, "num_tokens": 4125983.0, "reward": 0.4687499701976776, "reward_std": 0.17305275797843933, "rewards/format_reward/mean": 0.625, "rewards/format_reward/std": 0.2357022762298584, "rewards/mcq_exact_match_reward/mean": 0.40625, "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": 848.0, "completions/max_terminated_length": 848.0, "completions/mean_length": 56.140625, "completions/mean_terminated_length": 56.140625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.07366071428571429, "frac_reward_zero_std": 0.0, "grad_norm": 17.901029586791992, "learning_rate": 8.135881792367685e-07, "loss": -0.0, "num_tokens": 4224088.0, "reward": 0.46171873807907104, "reward_std": 0.23906555771827698, "rewards/format_reward/mean": 0.8671875, "rewards/format_reward/std": 0.22257846593856812, "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": 0.0, "completions/max_length": 1427.0, "completions/max_terminated_length": 1427.0, "completions/mean_length": 112.53125, "completions/mean_terminated_length": 112.53125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.07589285714285714, "frac_reward_zero_std": 0.0, "grad_norm": 30.553054809570312, "learning_rate": 8.005405736415125e-07, "loss": 0.0, "num_tokens": 4330018.0, "reward": 0.34687501192092896, "reward_std": 0.31498605012893677, "rewards/format_reward/mean": 0.8125, "rewards/format_reward/std": 0.2597312331199646, "rewards/mcq_exact_match_reward/mean": 0.265625, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 643.0, "completions/mean_length": 64.328125, "completions/mean_terminated_length": 32.841270446777344, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.078125, "frac_reward_zero_std": 0.25, "grad_norm": 26.636701583862305, "learning_rate": 7.871643313414718e-07, "loss": 0.0, "num_tokens": 4426999.0, "reward": 0.46015626192092896, "reward_std": 0.23253700137138367, "rewards/format_reward/mean": 0.8515625, "rewards/format_reward/std": 0.26246222853660583, "rewards/mcq_exact_match_reward/mean": 0.375, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 538.0, "completions/max_terminated_length": 538.0, "completions/mean_length": 38.46875, "completions/mean_terminated_length": 38.46875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08035714285714286, "frac_reward_zero_std": 0.375, "grad_norm": 17.43048095703125, "learning_rate": 7.734740790612136e-07, "loss": 0.0, "num_tokens": 4524477.0, "reward": 0.5718749761581421, "reward_std": 0.281593382358551, "rewards/format_reward/mean": 0.875, "rewards/format_reward/std": 0.2182178944349289, "rewards/mcq_exact_match_reward/mean": 0.484375, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "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": 594.0, "completions/max_terminated_length": 594.0, "completions/mean_length": 62.921875, "completions/mean_terminated_length": 62.921875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08258928571428571, "frac_reward_zero_std": 0.625, "grad_norm": 8.20571231842041, "learning_rate": 7.594847868906076e-07, "loss": 0.0, "num_tokens": 4592592.0, "reward": 0.5609375238418579, "reward_std": 0.15699616074562073, "rewards/format_reward/mean": 0.921875, "rewards/format_reward/std": 0.20351573824882507, "rewards/mcq_exact_match_reward/mean": 0.46875, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "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": 324.0, "completions/max_terminated_length": 324.0, "completions/mean_length": 38.515625, "completions/mean_terminated_length": 38.515625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08482142857142858, "frac_reward_zero_std": 0.25, "grad_norm": 19.207958221435547, "learning_rate": 7.452117519152541e-07, "loss": 0.0, "num_tokens": 4701249.0, "reward": 0.49531251192092896, "reward_std": 0.3524002432823181, "rewards/format_reward/mean": 0.890625, "rewards/format_reward/std": 0.2083333432674408, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "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": 288.0, "completions/max_terminated_length": 288.0, "completions/mean_length": 26.5625, "completions/mean_terminated_length": 26.5625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08705357142857142, "frac_reward_zero_std": 0.375, "grad_norm": 23.38102912902832, "learning_rate": 7.306705814893439e-07, "loss": -0.0, "num_tokens": 4771797.0, "reward": 0.25, "reward_std": 0.18224896490573883, "rewards/format_reward/mean": 0.9375, "rewards/format_reward/std": 0.1666666716337204, "rewards/mcq_exact_match_reward/mean": 0.15625, "rewards/mcq_exact_match_reward/std": 0.36596253514289856, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 440.0, "completions/mean_length": 116.140625, "completions/mean_terminated_length": 53.82257843017578, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08928571428571429, "frac_reward_zero_std": 0.25, "grad_norm": 11.691500663757324, "learning_rate": 7.158771761692464e-07, "loss": 0.0, "num_tokens": 4873502.0, "reward": 0.6382812857627869, "reward_std": 0.3285192549228668, "rewards/format_reward/mean": 0.9140625, "rewards/format_reward/std": 0.244862899184227, "rewards/mcq_exact_match_reward/mean": 0.546875, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "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": 344.0, "completions/max_terminated_length": 344.0, "completions/mean_length": 49.046875, "completions/mean_terminated_length": 49.046875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.09151785714285714, "frac_reward_zero_std": 0.125, "grad_norm": 21.862348556518555, "learning_rate": 7.008477123264847e-07, "loss": 0.0, "num_tokens": 4960465.0, "reward": 0.5023437738418579, "reward_std": 0.2822685241699219, "rewards/format_reward/mean": 0.9609375, "rewards/format_reward/std": 0.18483558297157288, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "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": 758.0, "completions/max_terminated_length": 758.0, "completions/mean_length": 75.3125, "completions/mean_terminated_length": 75.3125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.09375, "frac_reward_zero_std": 0.5, "grad_norm": 29.508949279785156, "learning_rate": 6.855986244591103e-07, "loss": 0.0, "num_tokens": 5056733.0, "reward": 0.4882812798023224, "reward_std": 0.19260738790035248, "rewards/format_reward/mean": 0.9765625, "rewards/format_reward/std": 0.10652101784944534, "rewards/mcq_exact_match_reward/mean": 0.390625, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "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": 832.0, "completions/max_terminated_length": 832.0, "completions/mean_length": 66.359375, "completions/mean_terminated_length": 66.359375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.09598214285714286, "frac_reward_zero_std": 0.375, "grad_norm": 13.661324501037598, "learning_rate": 6.701465872208216e-07, "loss": -0.0, "num_tokens": 5162996.0, "reward": 0.45078128576278687, "reward_std": 0.25875324010849, "rewards/format_reward/mean": 0.9140625, "rewards/format_reward/std": 0.2280818521976471, "rewards/mcq_exact_match_reward/mean": 0.359375, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "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": 521.0, "completions/max_terminated_length": 521.0, "completions/mean_length": 47.65625, "completions/mean_terminated_length": 47.65625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.09821428571428571, "frac_reward_zero_std": 0.625, "grad_norm": 14.362375259399414, "learning_rate": 6.545084971874736e-07, "loss": -0.0, "num_tokens": 5246886.0, "reward": 0.577343761920929, "reward_std": 0.18048441410064697, "rewards/format_reward/mean": 0.9296875, "rewards/format_reward/std": 0.19654127955436707, "rewards/mcq_exact_match_reward/mean": 0.484375, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "step": 44 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1023.0, "completions/mean_length": 131.578125, "completions/mean_terminated_length": 69.75806427001953, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "epoch": 0.10044642857142858, "frac_reward_zero_std": 0.375, "grad_norm": 13.058550834655762, "learning_rate": 6.387014543809223e-07, "loss": 0.0, "num_tokens": 5361291.0, "reward": 0.57421875, "reward_std": 0.17691533267498016, "rewards/format_reward/mean": 0.8984375, "rewards/format_reward/std": 0.25479042530059814, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1929.0, "completions/mean_length": 151.703125, "completions/mean_terminated_length": 90.53225708007812, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.10267857142857142, "frac_reward_zero_std": 0.5, "grad_norm": 5.845555305480957, "learning_rate": 6.227427435703995e-07, "loss": 0.0, "num_tokens": 5465920.0, "reward": 0.6187499761581421, "reward_std": 0.18365247547626495, "rewards/format_reward/mean": 0.875, "rewards/format_reward/std": 0.28171807527542114, "rewards/mcq_exact_match_reward/mean": 0.53125, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 465.0, "completions/mean_length": 99.765625, "completions/mean_terminated_length": 36.91935348510742, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.10491071428571429, "frac_reward_zero_std": 0.25, "grad_norm": 18.143478393554688, "learning_rate": 6.066498153718734e-07, "loss": 0.0, "num_tokens": 5547641.0, "reward": 0.3843750059604645, "reward_std": 0.2764841914176941, "rewards/format_reward/mean": 0.875, "rewards/format_reward/std": 0.26726123690605164, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "step": 47 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 704.0, "completions/mean_length": 143.65625, "completions/mean_terminated_length": 82.2258071899414, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.10714285714285714, "frac_reward_zero_std": 0.25, "grad_norm": 32.34805679321289, "learning_rate": 5.90440267166055e-07, "loss": 0.0, "num_tokens": 5660619.0, "reward": 0.635937511920929, "reward_std": 0.20371733605861664, "rewards/format_reward/mean": 0.890625, "rewards/format_reward/std": 0.24346621334552765, "rewards/mcq_exact_match_reward/mean": 0.546875, "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": 630.0, "completions/max_terminated_length": 630.0, "completions/mean_length": 67.390625, "completions/mean_terminated_length": 67.390625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.109375, "frac_reward_zero_std": 0.25, "grad_norm": 13.70897102355957, "learning_rate": 5.741318238559209e-07, "loss": 0.0, "num_tokens": 5771004.0, "reward": 0.4242187738418579, "reward_std": 0.3430789113044739, "rewards/format_reward/mean": 0.9609375, "rewards/format_reward/std": 0.13524486124515533, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "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": 0.046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 834.0, "completions/mean_length": 197.046875, "completions/mean_terminated_length": 106.01638793945312, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.11160714285714286, "frac_reward_zero_std": 0.0, "grad_norm": 19.1356201171875, "learning_rate": 5.577423184847931e-07, "loss": 0.0, "num_tokens": 5882687.0, "reward": 0.3109375238418579, "reward_std": 0.38860881328582764, "rewards/format_reward/mean": 0.921875, "rewards/format_reward/std": 0.25539806485176086, "rewards/mcq_exact_match_reward/mean": 0.21875, "rewards/mcq_exact_match_reward/std": 0.4166666865348816, "step": 50 } ], "logging_steps": 1, "max_steps": 100, "num_input_tokens_seen": 5882687, "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 }