{ "best_global_step": null, "best_metric": null, "best_model_checkpoint": null, "epoch": 0.25, "eval_steps": 500, "global_step": 100, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1440.0, "completions/mean_length": 411.03125, "completions/mean_terminated_length": 385.0476379394531, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.0025, "frac_reward_zero_std": 0.0, "grad_norm": 13.379964828491211, "learning_rate": 0.0, "loss": -0.0, "num_tokens": 118618.0, "reward": 0.21171873807907104, "reward_std": 0.35573673248291016, "rewards/format_reward/mean": 0.2421875, "rewards/format_reward/std": 0.25185325741767883, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1290.0, "completions/mean_length": 492.96875, "completions/mean_terminated_length": 442.8064270019531, "completions/min_length": 15.0, "completions/min_terminated_length": 15.0, "epoch": 0.005, "frac_reward_zero_std": 0.0, "grad_norm": 2.445894479751587, "learning_rate": 2e-07, "loss": 0.0, "num_tokens": 222928.0, "reward": 0.22109374403953552, "reward_std": 0.3713722825050354, "rewards/format_reward/mean": 0.3359375, "rewards/format_reward/std": 0.30930912494659424, "rewards/mcq_exact_match_reward/mean": 0.1875, "rewards/mcq_exact_match_reward/std": 0.39339789748191833, "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": 1460.0, "completions/max_terminated_length": 1460.0, "completions/mean_length": 469.03125, "completions/mean_terminated_length": 469.03125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0075, "frac_reward_zero_std": 0.125, "grad_norm": 4.150572776794434, "learning_rate": 4e-07, "loss": -0.0, "num_tokens": 351674.0, "reward": 0.3671875, "reward_std": 0.35188227891921997, "rewards/format_reward/mean": 0.390625, "rewards/format_reward/std": 0.2083333432674408, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "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": 1993.0, "completions/mean_length": 664.796875, "completions/mean_terminated_length": 620.1774291992188, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "epoch": 0.01, "frac_reward_zero_std": 0.0, "grad_norm": 1.868766188621521, "learning_rate": 6e-07, "loss": -0.0, "num_tokens": 489285.0, "reward": 0.25468748807907104, "reward_std": 0.3940054774284363, "rewards/format_reward/mean": 0.359375, "rewards/format_reward/std": 0.22658175230026245, "rewards/mcq_exact_match_reward/mean": 0.21875, "rewards/mcq_exact_match_reward/std": 0.4166666865348816, "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": 1038.0, "completions/max_terminated_length": 1038.0, "completions/mean_length": 450.28125, "completions/mean_terminated_length": 450.28125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0125, "frac_reward_zero_std": 0.0, "grad_norm": 5.173692226409912, "learning_rate": 8e-07, "loss": -0.0, "num_tokens": 591591.0, "reward": 0.46015626192092896, "reward_std": 0.3388790488243103, "rewards/format_reward/mean": 0.3828125, "rewards/format_reward/std": 0.3052735924720764, "rewards/mcq_exact_match_reward/mean": 0.421875, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1844.0, "completions/mean_length": 512.328125, "completions/mean_terminated_length": 462.7903137207031, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.015, "frac_reward_zero_std": 0.0, "grad_norm": 8.16497802734375, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 738844.0, "reward": 0.2085937261581421, "reward_std": 0.3302217721939087, "rewards/format_reward/mean": 0.3671875, "rewards/format_reward/std": 0.23974503576755524, "rewards/mcq_exact_match_reward/mean": 0.171875, "rewards/mcq_exact_match_reward/std": 0.38025420904159546, "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": 932.0, "completions/mean_length": 430.546875, "completions/mean_terminated_length": 404.873046875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0175, "frac_reward_zero_std": 0.0, "grad_norm": 10.982437133789062, "learning_rate": 9.99726628670463e-07, "loss": -0.0, "num_tokens": 873671.0, "reward": 0.27265626192092896, "reward_std": 0.2897842228412628, "rewards/format_reward/mean": 0.3828125, "rewards/format_reward/std": 0.2313210517168045, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "step": 7 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1933.0, "completions/mean_length": 733.765625, "completions/mean_terminated_length": 646.1500244140625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.02, "frac_reward_zero_std": 0.125, "grad_norm": 7.61357307434082, "learning_rate": 9.989068136093872e-07, "loss": -0.0, "num_tokens": 1023616.0, "reward": 0.16015625, "reward_std": 0.2687035799026489, "rewards/format_reward/mean": 0.3515625, "rewards/format_reward/std": 0.29113471508026123, "rewards/mcq_exact_match_reward/mean": 0.125, "rewards/mcq_exact_match_reward/std": 0.3333333432674408, "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": 1311.0, "completions/max_terminated_length": 1311.0, "completions/mean_length": 569.65625, "completions/mean_terminated_length": 569.65625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0225, "frac_reward_zero_std": 0.0, "grad_norm": 3.1806397438049316, "learning_rate": 9.975414512725056e-07, "loss": 0.0, "num_tokens": 1154970.0, "reward": 0.30546873807907104, "reward_std": 0.31296759843826294, "rewards/format_reward/mean": 0.3984375, "rewards/format_reward/std": 0.20275264978408813, "rewards/mcq_exact_match_reward/mean": 0.265625, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "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": 1702.0, "completions/max_terminated_length": 1702.0, "completions/mean_length": 520.625, "completions/mean_terminated_length": 520.625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.025, "frac_reward_zero_std": 0.0, "grad_norm": 6.8020806312561035, "learning_rate": 9.956320346634875e-07, "loss": 0.0, "num_tokens": 1266282.0, "reward": 0.37734371423721313, "reward_std": 0.43788474798202515, "rewards/format_reward/mean": 0.3359375, "rewards/format_reward/std": 0.2366211861371994, "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": 1566.0, "completions/max_terminated_length": 1566.0, "completions/mean_length": 417.890625, "completions/mean_terminated_length": 417.890625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0275, "frac_reward_zero_std": 0.0, "grad_norm": 13.776562690734863, "learning_rate": 9.931806517013612e-07, "loss": -0.0, "num_tokens": 1367499.0, "reward": 0.3390624523162842, "reward_std": 0.39278343319892883, "rewards/format_reward/mean": 0.421875, "rewards/format_reward/std": 0.25539806485176086, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "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": 1196.0, "completions/max_terminated_length": 1196.0, "completions/mean_length": 609.40625, "completions/mean_terminated_length": 609.40625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.03, "frac_reward_zero_std": 0.125, "grad_norm": 1.4482215642929077, "learning_rate": 9.901899829374047e-07, "loss": 0.0, "num_tokens": 1529141.0, "reward": 0.13124999403953552, "reward_std": 0.166652113199234, "rewards/format_reward/mean": 0.375, "rewards/format_reward/std": 0.2182178944349289, "rewards/mcq_exact_match_reward/mean": 0.09375, "rewards/mcq_exact_match_reward/std": 0.29378482699394226, "step": 12 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1986.0, "completions/max_terminated_length": 1986.0, "completions/mean_length": 527.765625, "completions/mean_terminated_length": 527.765625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0325, "frac_reward_zero_std": 0.0, "grad_norm": 11.483098983764648, "learning_rate": 9.866632986240029e-07, "loss": -0.0, "num_tokens": 1647174.0, "reward": 0.3203124701976776, "reward_std": 0.3535535931587219, "rewards/format_reward/mean": 0.390625, "rewards/format_reward/std": 0.24346621334552765, "rewards/mcq_exact_match_reward/mean": 0.28125, "rewards/mcq_exact_match_reward/std": 0.4531635046005249, "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": 1367.0, "completions/max_terminated_length": 1367.0, "completions/mean_length": 472.28125, "completions/mean_terminated_length": 472.28125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.035, "frac_reward_zero_std": 0.0, "grad_norm": 3.6256966590881348, "learning_rate": 9.826044551386742e-07, "loss": 0.0, "num_tokens": 1778856.0, "reward": 0.3984375, "reward_std": 0.4515121579170227, "rewards/format_reward/mean": 0.390625, "rewards/format_reward/std": 0.22658175230026245, "rewards/mcq_exact_match_reward/mean": 0.359375, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "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": 0.0, "completions/max_length": 1162.0, "completions/max_terminated_length": 1162.0, "completions/mean_length": 505.078125, "completions/mean_terminated_length": 505.078125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0375, "frac_reward_zero_std": 0.125, "grad_norm": 5.721800327301025, "learning_rate": 9.780178907671788e-07, "loss": -0.0, "num_tokens": 1893813.0, "reward": 0.4039062559604645, "reward_std": 0.364518404006958, "rewards/format_reward/mean": 0.4453125, "rewards/format_reward/std": 0.29703813791275024, "rewards/mcq_exact_match_reward/mean": 0.359375, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "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": 0.046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1924.0, "completions/mean_length": 700.078125, "completions/mean_terminated_length": 633.786865234375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.04, "frac_reward_zero_std": 0.125, "grad_norm": 9.845195770263672, "learning_rate": 9.729086208503173e-07, "loss": 0.0, "num_tokens": 2058154.0, "reward": 0.1874999850988388, "reward_std": 0.24792778491973877, "rewards/format_reward/mean": 0.3125, "rewards/format_reward/std": 0.24397502839565277, "rewards/mcq_exact_match_reward/mean": 0.15625, "rewards/mcq_exact_match_reward/std": 0.36596253514289856, "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": 1582.0, "completions/max_terminated_length": 1582.0, "completions/mean_length": 526.0, "completions/mean_terminated_length": 526.0, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0425, "frac_reward_zero_std": 0.125, "grad_norm": 4.484685897827148, "learning_rate": 9.672822322997304e-07, "loss": -0.0, "num_tokens": 2172050.0, "reward": 0.28359371423721313, "reward_std": 0.3160938024520874, "rewards/format_reward/mean": 0.4921875, "rewards/format_reward/std": 0.18881812691688538, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1344.0, "completions/mean_length": 454.453125, "completions/mean_terminated_length": 429.15875244140625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.045, "frac_reward_zero_std": 0.125, "grad_norm": 12.590941429138184, "learning_rate": 9.611448774886923e-07, "loss": 0.0, "num_tokens": 2278775.0, "reward": 0.22109374403953552, "reward_std": 0.1899556666612625, "rewards/format_reward/mean": 0.4921875, "rewards/format_reward/std": 0.20877929031848907, "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": 0.0, "completions/max_length": 1404.0, "completions/max_terminated_length": 1404.0, "completions/mean_length": 648.90625, "completions/mean_terminated_length": 648.90625, "completions/min_length": 154.0, "completions/min_terminated_length": 154.0, "epoch": 0.0475, "frac_reward_zero_std": 0.125, "grad_norm": 0.6111228466033936, "learning_rate": 9.545032675245813e-07, "loss": -0.0, "num_tokens": 2420609.0, "reward": 0.35624998807907104, "reward_std": 0.29947617650032043, "rewards/format_reward/mean": 0.4375, "rewards/format_reward/std": 0.1666666716337204, "rewards/mcq_exact_match_reward/mean": 0.3125, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1470.0, "completions/mean_length": 445.328125, "completions/mean_terminated_length": 393.6290283203125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.05, "frac_reward_zero_std": 0.0, "grad_norm": 12.61819076538086, "learning_rate": 9.473646649103817e-07, "loss": -0.0, "num_tokens": 2543014.0, "reward": 0.4546874761581421, "reward_std": 0.4531806707382202, "rewards/format_reward/mean": 0.484375, "rewards/format_reward/std": 0.29504841566085815, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1960.0, "completions/mean_length": 592.296875, "completions/mean_terminated_length": 545.3386840820312, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0525, "frac_reward_zero_std": 0.25, "grad_norm": 12.316622734069824, "learning_rate": 9.397368756032444e-07, "loss": 0.0, "num_tokens": 2690841.0, "reward": 0.3203125, "reward_std": 0.29384204745292664, "rewards/format_reward/mean": 0.546875, "rewards/format_reward/std": 0.21304203569889069, "rewards/mcq_exact_match_reward/mean": 0.265625, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "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": 1411.0, "completions/max_terminated_length": 1411.0, "completions/mean_length": 428.9375, "completions/mean_terminated_length": 428.9375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.055, "frac_reward_zero_std": 0.0, "grad_norm": 7.614321231842041, "learning_rate": 9.316282404787869e-07, "loss": -0.0, "num_tokens": 2796701.0, "reward": 0.4468749761581421, "reward_std": 0.4583485722541809, "rewards/format_reward/mean": 0.5625, "rewards/format_reward/std": 0.28867512941360474, "rewards/mcq_exact_match_reward/mean": 0.390625, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "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": 1267.0, "completions/max_terminated_length": 1267.0, "completions/mean_length": 502.234375, "completions/mean_terminated_length": 502.234375, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.0575, "frac_reward_zero_std": 0.0, "grad_norm": 5.405074596405029, "learning_rate": 9.230476262104676e-07, "loss": -0.0, "num_tokens": 2892836.0, "reward": 0.4429687261581421, "reward_std": 0.462321400642395, "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": 23 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1841.0, "completions/mean_length": 505.359375, "completions/mean_terminated_length": 480.873046875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.06, "frac_reward_zero_std": 0.25, "grad_norm": 10.152621269226074, "learning_rate": 9.1400441557401e-07, "loss": 0.0, "num_tokens": 3036515.0, "reward": 0.4546874761581421, "reward_std": 0.25815680623054504, "rewards/format_reward/mean": 0.484375, "rewards/format_reward/std": 0.125, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1206.0, "completions/mean_length": 393.0625, "completions/mean_terminated_length": 339.6773986816406, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0625, "frac_reward_zero_std": 0.125, "grad_norm": 17.456857681274414, "learning_rate": 9.045084971874737e-07, "loss": -0.0, "num_tokens": 3141271.0, "reward": 0.42656248807907104, "reward_std": 0.40213000774383545, "rewards/format_reward/mean": 0.515625, "rewards/format_reward/std": 0.23517554998397827, "rewards/mcq_exact_match_reward/mean": 0.375, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 1467.0, "completions/mean_length": 528.421875, "completions/mean_terminated_length": 479.4031982421875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.065, "frac_reward_zero_std": 0.0, "grad_norm": 14.22276496887207, "learning_rate": 8.945702546981968e-07, "loss": -0.0, "num_tokens": 3270018.0, "reward": 0.4390624761581421, "reward_std": 0.45948851108551025, "rewards/format_reward/mean": 0.484375, "rewards/format_reward/std": 0.1985812783241272, "rewards/mcq_exact_match_reward/mean": 0.390625, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "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": 0.0625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1732.0, "completions/mean_length": 485.03125, "completions/mean_terminated_length": 380.8333435058594, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0675, "frac_reward_zero_std": 0.125, "grad_norm": 10.338312149047852, "learning_rate": 8.842005554284295e-07, "loss": -0.0, "num_tokens": 3393604.0, "reward": 0.18671873211860657, "reward_std": 0.23389193415641785, "rewards/format_reward/mean": 0.4609375, "rewards/format_reward/std": 0.27174752950668335, "rewards/mcq_exact_match_reward/mean": 0.140625, "rewards/mcq_exact_match_reward/std": 0.3503824472427368, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1023.0, "completions/mean_length": 236.375, "completions/mean_terminated_length": 177.93548583984375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.07, "frac_reward_zero_std": 0.5, "grad_norm": 15.465999603271484, "learning_rate": 8.734107384920769e-07, "loss": 0.0, "num_tokens": 3488204.0, "reward": 0.46953123807907104, "reward_std": 0.10991687327623367, "rewards/format_reward/mean": 0.4765625, "rewards/format_reward/std": 0.20758795738220215, "rewards/mcq_exact_match_reward/mean": 0.421875, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "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": 959.0, "completions/max_terminated_length": 959.0, "completions/mean_length": 203.25, "completions/mean_terminated_length": 203.25, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0725, "frac_reward_zero_std": 0.125, "grad_norm": 22.171022415161133, "learning_rate": 8.622126023955445e-07, "loss": 0.0, "num_tokens": 3587436.0, "reward": 0.3312499523162842, "reward_std": 0.3558967113494873, "rewards/format_reward/mean": 0.5, "rewards/format_reward/std": 0.1259881556034088, "rewards/mcq_exact_match_reward/mean": 0.28125, "rewards/mcq_exact_match_reward/std": 0.4531635046005249, "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": 1281.0, "completions/max_terminated_length": 1281.0, "completions/mean_length": 238.390625, "completions/mean_terminated_length": 238.390625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.075, "frac_reward_zero_std": 0.5, "grad_norm": 21.5225887298584, "learning_rate": 8.506183921362442e-07, "loss": -0.0, "num_tokens": 3674637.0, "reward": 0.30312496423721313, "reward_std": 0.1754397749900818, "rewards/format_reward/mean": 0.53125, "rewards/format_reward/std": 0.12198751419782639, "rewards/mcq_exact_match_reward/mean": 0.25, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "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": 1541.0, "completions/max_terminated_length": 1541.0, "completions/mean_length": 176.890625, "completions/mean_terminated_length": 176.890625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0775, "frac_reward_zero_std": 0.25, "grad_norm": 17.143354415893555, "learning_rate": 8.386407858128706e-07, "loss": -0.0, "num_tokens": 3768846.0, "reward": 0.4585937261581421, "reward_std": 0.23260623216629028, "rewards/format_reward/mean": 0.5234375, "rewards/format_reward/std": 0.1649840772151947, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "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": 769.0, "completions/max_terminated_length": 769.0, "completions/mean_length": 102.03125, "completions/mean_terminated_length": 102.03125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08, "frac_reward_zero_std": 0.375, "grad_norm": 24.538957595825195, "learning_rate": 8.262928807620843e-07, "loss": 0.0, "num_tokens": 3868048.0, "reward": 0.5414062142372131, "reward_std": 0.3108735978603363, "rewards/format_reward/mean": 0.5703125, "rewards/format_reward/std": 0.1751912236213684, "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": 0.0, "completions/max_length": 444.0, "completions/max_terminated_length": 444.0, "completions/mean_length": 43.21875, "completions/mean_terminated_length": 43.21875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0825, "frac_reward_zero_std": 0.25, "grad_norm": 18.619295120239258, "learning_rate": 8.135881792367685e-07, "loss": 0.0, "num_tokens": 3960790.0, "reward": 0.42500001192092896, "reward_std": 0.22903475165367126, "rewards/format_reward/mean": 0.5, "rewards/format_reward/std": 0.08908708393573761, "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": 827.0, "completions/max_terminated_length": 827.0, "completions/mean_length": 86.53125, "completions/mean_terminated_length": 86.53125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.085, "frac_reward_zero_std": 0.5, "grad_norm": 11.220178604125977, "learning_rate": 8.005405736415125e-07, "loss": -0.0, "num_tokens": 4056648.0, "reward": 0.6968749761581421, "reward_std": 0.23965352773666382, "rewards/format_reward/mean": 0.5625, "rewards/format_reward/std": 0.18898223340511322, "rewards/mcq_exact_match_reward/mean": 0.640625, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "step": 34 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 751.0, "completions/mean_length": 93.75, "completions/mean_terminated_length": 62.73016357421875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0875, "frac_reward_zero_std": 0.25, "grad_norm": 26.791027069091797, "learning_rate": 7.871643313414718e-07, "loss": 0.0, "num_tokens": 4149944.0, "reward": 0.6312499642372131, "reward_std": 0.25268083810806274, "rewards/format_reward/mean": 0.53125, "rewards/format_reward/std": 0.1510545015335083, "rewards/mcq_exact_match_reward/mean": 0.578125, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "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": 1055.0, "completions/max_terminated_length": 1055.0, "completions/mean_length": 127.0625, "completions/mean_terminated_length": 127.0625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.09, "frac_reward_zero_std": 0.375, "grad_norm": 23.361783981323242, "learning_rate": 7.734740790612136e-07, "loss": -0.0, "num_tokens": 4284420.0, "reward": 0.4249999523162842, "reward_std": 0.2756393849849701, "rewards/format_reward/mean": 0.5, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.375, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 611.0, "completions/max_terminated_length": 611.0, "completions/mean_length": 55.765625, "completions/mean_terminated_length": 55.765625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0925, "frac_reward_zero_std": 0.625, "grad_norm": 9.95936107635498, "learning_rate": 7.594847868906076e-07, "loss": -0.0, "num_tokens": 4381829.0, "reward": 0.37890625, "reward_std": 0.10426763445138931, "rewards/format_reward/mean": 0.5078125, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "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": 454.0, "completions/max_terminated_length": 454.0, "completions/mean_length": 59.5, "completions/mean_terminated_length": 59.5, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.095, "frac_reward_zero_std": 0.125, "grad_norm": 28.146879196166992, "learning_rate": 7.452117519152541e-07, "loss": 0.0, "num_tokens": 4468453.0, "reward": 0.44609373807907104, "reward_std": 0.4099806547164917, "rewards/format_reward/mean": 0.7109375, "rewards/format_reward/std": 0.24888142943382263, "rewards/mcq_exact_match_reward/mean": 0.375, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 723.0, "completions/max_terminated_length": 723.0, "completions/mean_length": 97.953125, "completions/mean_terminated_length": 97.953125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0975, "frac_reward_zero_std": 0.375, "grad_norm": 20.614282608032227, "learning_rate": 7.306705814893439e-07, "loss": 0.0, "num_tokens": 4553410.0, "reward": 0.7882812023162842, "reward_std": 0.29249900579452515, "rewards/format_reward/mean": 0.5390625, "rewards/format_reward/std": 0.16194961965084076, "rewards/mcq_exact_match_reward/mean": 0.734375, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "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": 426.0, "completions/max_terminated_length": 426.0, "completions/mean_length": 65.8125, "completions/mean_terminated_length": 65.8125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1, "frac_reward_zero_std": 0.625, "grad_norm": 11.6072359085083, "learning_rate": 7.158771761692464e-07, "loss": -0.0, "num_tokens": 4656406.0, "reward": 0.4609375, "reward_std": 0.14697228372097015, "rewards/format_reward/mean": 0.546875, "rewards/format_reward/std": 0.19352105259895325, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "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": 880.0, "completions/max_terminated_length": 880.0, "completions/mean_length": 133.046875, "completions/mean_terminated_length": 133.046875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1025, "frac_reward_zero_std": 0.625, "grad_norm": 17.0133113861084, "learning_rate": 7.008477123264847e-07, "loss": -0.0, "num_tokens": 4745449.0, "reward": 0.46328121423721313, "reward_std": 0.13159699738025665, "rewards/format_reward/mean": 0.5703125, "rewards/format_reward/std": 0.1751912236213684, "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": 523.0, "completions/max_terminated_length": 523.0, "completions/mean_length": 89.015625, "completions/mean_terminated_length": 89.015625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.105, "frac_reward_zero_std": 0.625, "grad_norm": 7.227995872497559, "learning_rate": 6.855986244591103e-07, "loss": -0.0, "num_tokens": 4845050.0, "reward": 0.546875, "reward_std": 0.1804211586713791, "rewards/format_reward/mean": 0.625, "rewards/format_reward/std": 0.2182178944349289, "rewards/mcq_exact_match_reward/mean": 0.484375, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "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": 755.0, "completions/max_terminated_length": 755.0, "completions/mean_length": 42.234375, "completions/mean_terminated_length": 42.234375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1075, "frac_reward_zero_std": 0.5, "grad_norm": 28.610111236572266, "learning_rate": 6.701465872208216e-07, "loss": -0.0, "num_tokens": 4950289.0, "reward": 0.5242186784744263, "reward_std": 0.2241736352443695, "rewards/format_reward/mean": 0.5546875, "rewards/format_reward/std": 0.15728822350502014, "rewards/mcq_exact_match_reward/mean": 0.46875, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 430.0, "completions/mean_length": 72.25, "completions/mean_terminated_length": 40.888893127441406, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.11, "frac_reward_zero_std": 0.375, "grad_norm": 32.960147857666016, "learning_rate": 6.545084971874736e-07, "loss": 0.0, "num_tokens": 5046721.0, "reward": 0.4320312440395355, "reward_std": 0.1433981955051422, "rewards/format_reward/mean": 0.7265625, "rewards/format_reward/std": 0.2807259261608124, "rewards/mcq_exact_match_reward/mean": 0.359375, "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": 340.0, "completions/max_terminated_length": 340.0, "completions/mean_length": 12.890625, "completions/mean_terminated_length": 12.890625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1125, "frac_reward_zero_std": 0.5, "grad_norm": 26.44186782836914, "learning_rate": 6.387014543809223e-07, "loss": -0.0, "num_tokens": 5136434.0, "reward": 0.703125, "reward_std": 0.2109457403421402, "rewards/format_reward/mean": 0.625, "rewards/format_reward/std": 0.2182178944349289, "rewards/mcq_exact_match_reward/mean": 0.640625, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "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": 819.0, "completions/max_terminated_length": 819.0, "completions/mean_length": 45.140625, "completions/mean_terminated_length": 45.140625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.115, "frac_reward_zero_std": 0.125, "grad_norm": 29.510046005249023, "learning_rate": 6.227427435703995e-07, "loss": 0.0, "num_tokens": 5259315.0, "reward": 0.6773437261581421, "reward_std": 0.23261141777038574, "rewards/format_reward/mean": 0.5234375, "rewards/format_reward/std": 0.13886408507823944, "rewards/mcq_exact_match_reward/mean": 0.625, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 924.0, "completions/max_terminated_length": 924.0, "completions/mean_length": 65.859375, "completions/mean_terminated_length": 65.859375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1175, "frac_reward_zero_std": 0.375, "grad_norm": 26.25118637084961, "learning_rate": 6.066498153718734e-07, "loss": 0.0, "num_tokens": 5379626.0, "reward": 0.5718749761581421, "reward_std": 0.21378730237483978, "rewards/format_reward/mean": 0.5625, "rewards/format_reward/std": 0.1666666716337204, "rewards/mcq_exact_match_reward/mean": 0.515625, "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": 584.0, "completions/max_terminated_length": 584.0, "completions/mean_length": 43.046875, "completions/mean_terminated_length": 43.046875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.12, "frac_reward_zero_std": 0.5, "grad_norm": 44.91122817993164, "learning_rate": 5.90440267166055e-07, "loss": 0.0, "num_tokens": 5471725.0, "reward": 0.4820312261581421, "reward_std": 0.07460986077785492, "rewards/format_reward/mean": 0.6015625, "rewards/format_reward/std": 0.20275264978408813, "rewards/mcq_exact_match_reward/mean": 0.421875, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "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": 565.0, "completions/max_terminated_length": 565.0, "completions/mean_length": 26.5625, "completions/mean_terminated_length": 26.5625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1225, "frac_reward_zero_std": 0.375, "grad_norm": 52.630611419677734, "learning_rate": 5.741318238559209e-07, "loss": 0.0, "num_tokens": 5560601.0, "reward": 0.7109375, "reward_std": 0.11039550602436066, "rewards/format_reward/mean": 0.703125, "rewards/format_reward/std": 0.24750742316246033, "rewards/mcq_exact_match_reward/mean": 0.640625, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "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": 448.0, "completions/max_terminated_length": 448.0, "completions/mean_length": 47.359375, "completions/mean_terminated_length": 47.359375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.125, "frac_reward_zero_std": 0.625, "grad_norm": 13.111749649047852, "learning_rate": 5.577423184847931e-07, "loss": 0.0, "num_tokens": 5652696.0, "reward": 0.52734375, "reward_std": 0.16526663303375244, "rewards/format_reward/mean": 0.8984375, "rewards/format_reward/std": 0.20275264978408813, "rewards/mcq_exact_match_reward/mean": 0.4375, "rewards/mcq_exact_match_reward/std": 0.5, "step": 50 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 196.0, "completions/max_terminated_length": 196.0, "completions/mean_length": 26.375, "completions/mean_terminated_length": 26.375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1275, "frac_reward_zero_std": 0.625, "grad_norm": 23.74351692199707, "learning_rate": 5.412896727361662e-07, "loss": -0.0, "num_tokens": 5752192.0, "reward": 0.40937498211860657, "reward_std": 0.15992167592048645, "rewards/format_reward/mean": 0.8125, "rewards/format_reward/std": 0.24397502839565277, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "step": 51 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 988.0, "completions/max_terminated_length": 988.0, "completions/mean_length": 105.375, "completions/mean_terminated_length": 105.375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.13, "frac_reward_zero_std": 0.375, "grad_norm": 14.15005874633789, "learning_rate": 5.247918773366111e-07, "loss": 0.0, "num_tokens": 5867984.0, "reward": 0.617968738079071, "reward_std": 0.25049111247062683, "rewards/format_reward/mean": 0.7109375, "rewards/format_reward/std": 0.2789533734321594, "rewards/mcq_exact_match_reward/mean": 0.546875, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 52 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1359.0, "completions/max_terminated_length": 1359.0, "completions/mean_length": 157.828125, "completions/mean_terminated_length": 157.828125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1325, "frac_reward_zero_std": 0.25, "grad_norm": 31.38736343383789, "learning_rate": 5.082669723831793e-07, "loss": 0.0, "num_tokens": 5967237.0, "reward": 0.703125, "reward_std": 0.26853489875793457, "rewards/format_reward/mean": 0.9375, "rewards/format_reward/std": 0.1666666716337204, "rewards/mcq_exact_match_reward/mean": 0.609375, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "step": 53 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1028.0, "completions/max_terminated_length": 1028.0, "completions/mean_length": 93.578125, "completions/mean_terminated_length": 93.578125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.135, "frac_reward_zero_std": 0.625, "grad_norm": 13.873196601867676, "learning_rate": 4.917330276168208e-07, "loss": -0.0, "num_tokens": 6072402.0, "reward": 0.63671875, "reward_std": 0.11330264806747437, "rewards/format_reward/mean": 0.8984375, "rewards/format_reward/std": 0.20275264978408813, "rewards/mcq_exact_match_reward/mean": 0.546875, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 54 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 635.0, "completions/max_terminated_length": 635.0, "completions/mean_length": 65.28125, "completions/mean_terminated_length": 65.28125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1375, "frac_reward_zero_std": 0.5, "grad_norm": 10.523073196411133, "learning_rate": 4.752081226633888e-07, "loss": 0.0, "num_tokens": 6158764.0, "reward": 0.87890625, "reward_std": 0.23447422683238983, "rewards/format_reward/mean": 0.9765625, "rewards/format_reward/std": 0.10652101784944534, "rewards/mcq_exact_match_reward/mean": 0.78125, "rewards/mcq_exact_match_reward/std": 0.4166666865348816, "step": 55 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1239.0, "completions/max_terminated_length": 1239.0, "completions/mean_length": 83.53125, "completions/mean_terminated_length": 83.53125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.14, "frac_reward_zero_std": 0.5, "grad_norm": 10.06303882598877, "learning_rate": 4.5871032726383385e-07, "loss": 0.0, "num_tokens": 6270654.0, "reward": 0.4226562976837158, "reward_std": 0.16703051328659058, "rewards/format_reward/mean": 0.9453125, "rewards/format_reward/std": 0.15728822350502014, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "step": 56 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1656.0, "completions/mean_length": 104.0, "completions/mean_terminated_length": 73.14286041259766, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1425, "frac_reward_zero_std": 0.375, "grad_norm": 16.617080688476562, "learning_rate": 4.4225768151520694e-07, "loss": 0.0, "num_tokens": 6366478.0, "reward": 0.63671875, "reward_std": 0.2360953688621521, "rewards/format_reward/mean": 0.8984375, "rewards/format_reward/std": 0.22146137058734894, "rewards/mcq_exact_match_reward/mean": 0.546875, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 57 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 613.0, "completions/max_terminated_length": 613.0, "completions/mean_length": 39.0625, "completions/mean_terminated_length": 39.0625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.145, "frac_reward_zero_std": 0.625, "grad_norm": 46.76222229003906, "learning_rate": 4.258681761440789e-07, "loss": 0.0, "num_tokens": 6466250.0, "reward": 0.36250001192092896, "reward_std": 0.04963856190443039, "rewards/format_reward/mean": 0.96875, "rewards/format_reward/std": 0.12198751419782639, "rewards/mcq_exact_match_reward/mean": 0.265625, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "step": 58 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 681.0, "completions/max_terminated_length": 681.0, "completions/mean_length": 70.078125, "completions/mean_terminated_length": 70.078125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1475, "frac_reward_zero_std": 0.375, "grad_norm": 14.54373550415039, "learning_rate": 4.095597328339452e-07, "loss": 0.0, "num_tokens": 6543687.0, "reward": 0.70703125, "reward_std": 0.287955105304718, "rewards/format_reward/mean": 0.9765625, "rewards/format_reward/std": 0.13886408507823944, "rewards/mcq_exact_match_reward/mean": 0.609375, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "step": 59 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 574.0, "completions/max_terminated_length": 574.0, "completions/mean_length": 54.96875, "completions/mean_terminated_length": 54.96875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.15, "frac_reward_zero_std": 0.75, "grad_norm": 10.34229564666748, "learning_rate": 3.9335018462812664e-07, "loss": -0.0, "num_tokens": 6622653.0, "reward": 0.5531250238418579, "reward_std": 0.11100947856903076, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 60 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1547.0, "completions/max_terminated_length": 1547.0, "completions/mean_length": 77.0625, "completions/mean_terminated_length": 77.0625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1525, "frac_reward_zero_std": 0.5, "grad_norm": 13.931731224060059, "learning_rate": 3.772572564296004e-07, "loss": 0.0, "num_tokens": 6732169.0, "reward": 0.5625, "reward_std": 0.15013158321380615, "rewards/format_reward/mean": 0.9375, "rewards/format_reward/std": 0.18898223340511322, "rewards/mcq_exact_match_reward/mean": 0.46875, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "step": 61 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 417.0, "completions/max_terminated_length": 417.0, "completions/mean_length": 38.859375, "completions/mean_terminated_length": 38.859375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.155, "frac_reward_zero_std": 0.625, "grad_norm": 19.955392837524414, "learning_rate": 3.612985456190778e-07, "loss": -0.0, "num_tokens": 6821704.0, "reward": 0.2828125059604645, "reward_std": 0.11249187588691711, "rewards/format_reward/mean": 0.953125, "rewards/format_reward/std": 0.14689241349697113, "rewards/mcq_exact_match_reward/mean": 0.1875, "rewards/mcq_exact_match_reward/std": 0.39339789748191833, "step": 62 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1431.0, "completions/max_terminated_length": 1431.0, "completions/mean_length": 134.0625, "completions/mean_terminated_length": 134.0625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1575, "frac_reward_zero_std": 0.5, "grad_norm": 9.703402519226074, "learning_rate": 3.454915028125263e-07, "loss": 0.0, "num_tokens": 6921308.0, "reward": 0.8656250238418579, "reward_std": 0.19044628739356995, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.765625, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "step": 63 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 216.0, "completions/max_terminated_length": 216.0, "completions/mean_length": 25.828125, "completions/mean_terminated_length": 25.828125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.16, "frac_reward_zero_std": 0.75, "grad_norm": 8.14318561553955, "learning_rate": 3.2985341277917846e-07, "loss": 0.0, "num_tokens": 7015673.0, "reward": 0.5218750238418579, "reward_std": 0.10205793380737305, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.421875, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "step": 64 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 352.0, "completions/max_terminated_length": 352.0, "completions/mean_length": 37.640625, "completions/mean_terminated_length": 37.640625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1625, "frac_reward_zero_std": 0.375, "grad_norm": 29.07025718688965, "learning_rate": 3.1440137554088953e-07, "loss": 0.0, "num_tokens": 7104354.0, "reward": 0.55078125, "reward_std": 0.19368097186088562, "rewards/format_reward/mean": 0.9765625, "rewards/format_reward/std": 0.10652101784944534, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 65 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 678.0, "completions/mean_length": 83.828125, "completions/mean_terminated_length": 52.650794982910156, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.165, "frac_reward_zero_std": 0.375, "grad_norm": 15.688562393188477, "learning_rate": 2.9915228767351535e-07, "loss": 0.0, "num_tokens": 7215455.0, "reward": 0.550000011920929, "reward_std": 0.21620942652225494, "rewards/format_reward/mean": 0.96875, "rewards/format_reward/std": 0.17536810040473938, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 66 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 972.0, "completions/max_terminated_length": 972.0, "completions/mean_length": 29.0625, "completions/mean_terminated_length": 29.0625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1675, "frac_reward_zero_std": 0.75, "grad_norm": 12.147396087646484, "learning_rate": 2.841228238307536e-07, "loss": 0.0, "num_tokens": 7313339.0, "reward": 0.6460937857627869, "reward_std": 0.12194531410932541, "rewards/format_reward/mean": 0.9921875, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.546875, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 67 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 216.0, "completions/max_terminated_length": 216.0, "completions/mean_length": 35.25, "completions/mean_terminated_length": 35.25, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.17, "frac_reward_zero_std": 1.0, "grad_norm": 0.0, "learning_rate": 2.6932941851065615e-07, "loss": 0.0, "num_tokens": 7411059.0, "reward": 0.6000000238418579, "reward_std": 0.0, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.5, "rewards/mcq_exact_match_reward/std": 0.5039526224136353, "step": 68 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 518.0, "completions/max_terminated_length": 518.0, "completions/mean_length": 32.984375, "completions/mean_terminated_length": 32.984375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1725, "frac_reward_zero_std": 0.75, "grad_norm": 7.77724027633667, "learning_rate": 2.547882480847461e-07, "loss": 0.0, "num_tokens": 7504818.0, "reward": 0.37968751788139343, "reward_std": 0.06228317320346832, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.28125, "rewards/mcq_exact_match_reward/std": 0.4531635046005249, "step": 69 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1193.0, "completions/max_terminated_length": 1193.0, "completions/mean_length": 52.609375, "completions/mean_terminated_length": 52.609375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.175, "frac_reward_zero_std": 0.5, "grad_norm": 8.265303611755371, "learning_rate": 2.4051521310939254e-07, "loss": 0.0, "num_tokens": 7598889.0, "reward": 0.20781251788139343, "reward_std": 0.17329265177249908, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.109375, "rewards/mcq_exact_match_reward/std": 0.3145764470100403, "step": 70 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 478.0, "completions/max_terminated_length": 478.0, "completions/mean_length": 40.046875, "completions/mean_terminated_length": 40.046875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1775, "frac_reward_zero_std": 0.75, "grad_norm": 11.482749938964844, "learning_rate": 2.2652592093878665e-07, "loss": 0.0, "num_tokens": 7676084.0, "reward": 0.5062500238418579, "reward_std": 0.1246790662407875, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "step": 71 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 145.0, "completions/max_terminated_length": 145.0, "completions/mean_length": 26.1875, "completions/mean_terminated_length": 26.1875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.18, "frac_reward_zero_std": 0.75, "grad_norm": 13.321898460388184, "learning_rate": 2.128356686585282e-07, "loss": 0.0, "num_tokens": 7763944.0, "reward": 0.36250001192092896, "reward_std": 0.04998054727911949, "rewards/format_reward/mean": 0.96875, "rewards/format_reward/std": 0.17536810040473938, "rewards/mcq_exact_match_reward/mean": 0.265625, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "step": 72 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 517.0, "completions/max_terminated_length": 517.0, "completions/mean_length": 41.171875, "completions/mean_terminated_length": 41.171875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1825, "frac_reward_zero_std": 0.375, "grad_norm": 13.54677677154541, "learning_rate": 1.9945942635848745e-07, "loss": -0.0, "num_tokens": 7867979.0, "reward": 0.7085937857627869, "reward_std": 0.2617086172103882, "rewards/format_reward/mean": 0.9921875, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.609375, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "step": 73 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 816.0, "completions/max_terminated_length": 816.0, "completions/mean_length": 38.578125, "completions/mean_terminated_length": 38.578125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.185, "frac_reward_zero_std": 0.625, "grad_norm": 13.316141128540039, "learning_rate": 1.8641182076323148e-07, "loss": 0.0, "num_tokens": 7962568.0, "reward": 0.5062500238418579, "reward_std": 0.16675157845020294, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "step": 74 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 319.0, "completions/max_terminated_length": 319.0, "completions/mean_length": 22.921875, "completions/mean_terminated_length": 22.921875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1875, "frac_reward_zero_std": 0.625, "grad_norm": 16.270925521850586, "learning_rate": 1.7370711923791564e-07, "loss": 0.0, "num_tokens": 8065315.0, "reward": 0.5984375476837158, "reward_std": 0.09280776232481003, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.5, "rewards/mcq_exact_match_reward/std": 0.5039526224136353, "step": 75 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 92.0, "completions/max_terminated_length": 92.0, "completions/mean_length": 15.71875, "completions/mean_terminated_length": 15.71875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.19, "frac_reward_zero_std": 1.0, "grad_norm": 0.0, "learning_rate": 1.6135921418712955e-07, "loss": 0.0, "num_tokens": 8152185.0, "reward": 0.7250000238418579, "reward_std": 0.0, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.625, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "step": 76 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 149.0, "completions/max_terminated_length": 149.0, "completions/mean_length": 20.75, "completions/mean_terminated_length": 20.75, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1925, "frac_reward_zero_std": 0.375, "grad_norm": 14.954136848449707, "learning_rate": 1.493816078637557e-07, "loss": -0.0, "num_tokens": 8246721.0, "reward": 0.6468750238418579, "reward_std": 0.2824692726135254, "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": 77 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 234.0, "completions/max_terminated_length": 234.0, "completions/mean_length": 30.734375, "completions/mean_terminated_length": 30.734375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.195, "frac_reward_zero_std": 0.75, "grad_norm": 10.763412475585938, "learning_rate": 1.3778739760445552e-07, "loss": 0.0, "num_tokens": 8335992.0, "reward": 0.7875000238418579, "reward_std": 0.10888782143592834, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.6875, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "step": 78 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 123.0, "completions/max_terminated_length": 123.0, "completions/mean_length": 14.078125, "completions/mean_terminated_length": 14.078125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1975, "frac_reward_zero_std": 0.75, "grad_norm": 17.061342239379883, "learning_rate": 1.2658926150792322e-07, "loss": 0.0, "num_tokens": 8421421.0, "reward": 0.7875000238418579, "reward_std": 0.10888782143592834, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.6875, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "step": 79 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 240.0, "completions/max_terminated_length": 240.0, "completions/mean_length": 26.671875, "completions/mean_terminated_length": 26.671875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2, "frac_reward_zero_std": 0.75, "grad_norm": 7.0431904792785645, "learning_rate": 1.1579944457157059e-07, "loss": 0.0, "num_tokens": 8520448.0, "reward": 0.7250000238418579, "reward_std": 0.1157275140285492, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.625, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "step": 80 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 340.0, "completions/max_terminated_length": 340.0, "completions/mean_length": 30.234375, "completions/mean_terminated_length": 30.234375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2025, "frac_reward_zero_std": 0.875, "grad_norm": 2.818654775619507, "learning_rate": 1.0542974530180327e-07, "loss": -0.0, "num_tokens": 8612295.0, "reward": 0.7406250238418579, "reward_std": 0.04419417306780815, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.640625, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "step": 81 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 906.0, "completions/max_terminated_length": 906.0, "completions/mean_length": 55.65625, "completions/mean_terminated_length": 55.65625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.205, "frac_reward_zero_std": 0.5, "grad_norm": 15.114845275878906, "learning_rate": 9.549150281252632e-08, "loss": -0.0, "num_tokens": 8707833.0, "reward": 0.48828125, "reward_std": 0.13563647866249084, "rewards/format_reward/mean": 0.9765625, "rewards/format_reward/std": 0.13886408507823944, "rewards/mcq_exact_match_reward/mean": 0.390625, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "step": 82 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 350.0, "completions/max_terminated_length": 350.0, "completions/mean_length": 37.59375, "completions/mean_terminated_length": 37.59375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2075, "frac_reward_zero_std": 0.625, "grad_norm": 11.676969528198242, "learning_rate": 8.599558442598998e-08, "loss": 0.0, "num_tokens": 8811015.0, "reward": 0.6000000238418579, "reward_std": 0.1462520956993103, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.5, "rewards/mcq_exact_match_reward/std": 0.5039526224136353, "step": 83 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 217.0, "completions/max_terminated_length": 217.0, "completions/mean_length": 15.578125, "completions/mean_terminated_length": 15.578125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.21, "frac_reward_zero_std": 0.875, "grad_norm": 2.060880661010742, "learning_rate": 7.695237378953224e-08, "loss": 0.0, "num_tokens": 8899356.0, "reward": 0.5843750238418579, "reward_std": 0.04419417306780815, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.484375, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "step": 84 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 279.0, "completions/max_terminated_length": 279.0, "completions/mean_length": 16.421875, "completions/mean_terminated_length": 16.421875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2125, "frac_reward_zero_std": 0.625, "grad_norm": 16.402795791625977, "learning_rate": 6.837175952121304e-08, "loss": 0.0, "num_tokens": 8987503.0, "reward": 0.5515625476837158, "reward_std": 0.15534965693950653, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 85 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 401.0, "completions/max_terminated_length": 401.0, "completions/mean_length": 30.015625, "completions/mean_terminated_length": 30.015625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.215, "frac_reward_zero_std": 0.625, "grad_norm": 49.13320541381836, "learning_rate": 6.026312439675551e-08, "loss": 0.0, "num_tokens": 9079424.0, "reward": 0.67578125, "reward_std": 0.10529261827468872, "rewards/format_reward/mean": 0.9765625, "rewards/format_reward/std": 0.10652101784944534, "rewards/mcq_exact_match_reward/mean": 0.578125, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "step": 86 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 911.0, "completions/max_terminated_length": 911.0, "completions/mean_length": 30.03125, "completions/mean_terminated_length": 30.03125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.2175, "frac_reward_zero_std": 0.5, "grad_norm": 18.00295066833496, "learning_rate": 5.263533508961826e-08, "loss": -0.0, "num_tokens": 9191058.0, "reward": 0.5046875476837158, "reward_std": 0.15660987794399261, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.08768405020236969, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "step": 87 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 138.0, "completions/max_terminated_length": 138.0, "completions/mean_length": 17.25, "completions/mean_terminated_length": 17.25, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.22, "frac_reward_zero_std": 0.75, "grad_norm": 14.999302864074707, "learning_rate": 4.549673247541874e-08, "loss": 0.0, "num_tokens": 9276314.0, "reward": 0.6468750238418579, "reward_std": 0.11100947856903076, "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": 88 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 414.0, "completions/mean_length": 79.703125, "completions/mean_terminated_length": 48.46031951904297, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2225, "frac_reward_zero_std": 0.5, "grad_norm": 9.31420612335205, "learning_rate": 3.8855122511307626e-08, "loss": 0.0, "num_tokens": 9364495.0, "reward": 0.8765624761581421, "reward_std": 0.21432165801525116, "rewards/format_reward/mean": 0.953125, "rewards/format_reward/std": 0.21304203569889069, "rewards/mcq_exact_match_reward/mean": 0.78125, "rewards/mcq_exact_match_reward/std": 0.4166666865348816, "step": 89 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 148.0, "completions/max_terminated_length": 148.0, "completions/mean_length": 17.421875, "completions/mean_terminated_length": 17.421875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.225, "frac_reward_zero_std": 0.75, "grad_norm": 14.156384468078613, "learning_rate": 3.271776770026963e-08, "loss": 0.0, "num_tokens": 9442594.0, "reward": 0.6781250238418579, "reward_std": 0.10205793380737305, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.578125, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "step": 90 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 430.0, "completions/max_terminated_length": 430.0, "completions/mean_length": 34.875, "completions/mean_terminated_length": 34.875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2275, "frac_reward_zero_std": 1.0, "grad_norm": 0.0, "learning_rate": 2.7091379149682682e-08, "loss": 0.0, "num_tokens": 9518082.0, "reward": 0.8500000238418579, "reward_std": 0.0, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.75, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "step": 91 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 622.0, "completions/max_terminated_length": 622.0, "completions/mean_length": 22.890625, "completions/mean_terminated_length": 22.890625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.23, "frac_reward_zero_std": 0.75, "grad_norm": 11.048893928527832, "learning_rate": 2.1982109232821176e-08, "loss": -0.0, "num_tokens": 9624419.0, "reward": 0.6882812976837158, "reward_std": 0.06007346510887146, "rewards/format_reward/mean": 0.9453125, "rewards/format_reward/std": 0.15728822350502014, "rewards/mcq_exact_match_reward/mean": 0.59375, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "step": 92 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 327.0, "completions/max_terminated_length": 327.0, "completions/mean_length": 24.765625, "completions/mean_terminated_length": 24.765625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2325, "frac_reward_zero_std": 0.625, "grad_norm": 15.352795600891113, "learning_rate": 1.7395544861325718e-08, "loss": 0.0, "num_tokens": 9708148.0, "reward": 0.706250011920929, "reward_std": 0.0530330054461956, "rewards/format_reward/mean": 0.96875, "rewards/format_reward/std": 0.17536810040473938, "rewards/mcq_exact_match_reward/mean": 0.609375, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "step": 93 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 473.0, "completions/mean_length": 62.96875, "completions/mean_terminated_length": 31.46031951904297, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.235, "frac_reward_zero_std": 0.75, "grad_norm": 6.727560997009277, "learning_rate": 1.3336701375997127e-08, "loss": 0.0, "num_tokens": 9812938.0, "reward": 0.8484375476837158, "reward_std": 0.09280776977539062, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.75, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "step": 94 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 208.0, "completions/max_terminated_length": 208.0, "completions/mean_length": 23.109375, "completions/mean_terminated_length": 23.109375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2375, "frac_reward_zero_std": 0.75, "grad_norm": 12.164389610290527, "learning_rate": 9.810017062595321e-09, "loss": 0.0, "num_tokens": 9915161.0, "reward": 0.5687500238418579, "reward_std": 0.0883883461356163, "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": 95 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 318.0, "completions/max_terminated_length": 318.0, "completions/mean_length": 30.46875, "completions/mean_terminated_length": 30.46875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.24, "frac_reward_zero_std": 0.625, "grad_norm": 14.501875877380371, "learning_rate": 6.819348298638839e-09, "loss": 0.0, "num_tokens": 9992359.0, "reward": 0.6625000238418579, "reward_std": 0.1552036553621292, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.5625, "rewards/mcq_exact_match_reward/std": 0.5, "step": 96 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 598.0, "completions/max_terminated_length": 598.0, "completions/mean_length": 43.09375, "completions/mean_terminated_length": 43.09375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2425, "frac_reward_zero_std": 0.5, "grad_norm": 13.192380905151367, "learning_rate": 4.367965336512403e-09, "loss": 0.0, "num_tokens": 10096517.0, "reward": 0.6148437857627869, "reward_std": 0.19081565737724304, "rewards/format_reward/mean": 0.9921875, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.515625, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "step": 97 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 17.0, "completions/max_terminated_length": 17.0, "completions/mean_length": 12.21875, "completions/mean_terminated_length": 12.21875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.245, "frac_reward_zero_std": 0.875, "grad_norm": 8.120609283447266, "learning_rate": 2.458548727494292e-09, "loss": 0.0, "num_tokens": 10218179.0, "reward": 0.4593750238418579, "reward_std": 0.04419417306780815, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.359375, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "step": 98 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 224.0, "completions/max_terminated_length": 224.0, "completions/mean_length": 20.296875, "completions/mean_terminated_length": 20.296875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.2475, "frac_reward_zero_std": 0.5, "grad_norm": 23.42853546142578, "learning_rate": 1.0931863906127325e-09, "loss": 0.0, "num_tokens": 10292118.0, "reward": 0.926562488079071, "reward_std": 0.13547571003437042, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.08768405020236969, "rewards/mcq_exact_match_reward/mean": 0.828125, "rewards/mcq_exact_match_reward/std": 0.38025420904159546, "step": 99 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 405.0, "completions/max_terminated_length": 405.0, "completions/mean_length": 37.109375, "completions/mean_terminated_length": 37.109375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.25, "frac_reward_zero_std": 0.75, "grad_norm": 11.104909896850586, "learning_rate": 2.733713295369755e-10, "loss": 0.0, "num_tokens": 10376717.0, "reward": 0.6000000238418579, "reward_std": 0.0883883461356163, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.5, "rewards/mcq_exact_match_reward/std": 0.5039526224136353, "step": 100 } ], "logging_steps": 1, "max_steps": 100, "num_input_tokens_seen": 10376717, "num_train_epochs": 1, "save_steps": 50, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": true }, "attributes": {} } }, "total_flos": 0.0, "train_batch_size": 1, "trial_name": null, "trial_params": null }