{ "best_global_step": null, "best_metric": null, "best_model_checkpoint": null, "epoch": 0.22321428571428573, "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": 1702.0, "completions/mean_length": 658.671875, "completions/mean_terminated_length": 636.6190795898438, "completions/min_length": 49.0, "completions/min_terminated_length": 49.0, "epoch": 0.002232142857142857, "frac_reward_zero_std": 0.125, "grad_norm": 1.179520606994629, "learning_rate": 0.0, "loss": -0.0, "num_tokens": 158547.0, "reward": 0.3374999761581421, "reward_std": 0.27919313311576843, "rewards/format_reward/mean": 0.40625, "rewards/format_reward/std": 0.21593283116817474, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "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": 1975.0, "completions/max_terminated_length": 1975.0, "completions/mean_length": 456.875, "completions/mean_terminated_length": 456.875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.004464285714285714, "frac_reward_zero_std": 0.0, "grad_norm": 35.00798797607422, "learning_rate": 2e-07, "loss": 0.0, "num_tokens": 283275.0, "reward": 0.2890625, "reward_std": 0.2649531364440918, "rewards/format_reward/mean": 0.390625, "rewards/format_reward/std": 0.2592533528804779, "rewards/mcq_exact_match_reward/mean": 0.25, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "step": 2 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1352.0, "completions/mean_length": 444.15625, "completions/mean_terminated_length": 418.69842529296875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.006696428571428571, "frac_reward_zero_std": 0.0, "grad_norm": 10.055992126464844, "learning_rate": 4e-07, "loss": -0.0, "num_tokens": 397613.0, "reward": 0.39921870827674866, "reward_std": 0.41733628511428833, "rewards/format_reward/mean": 0.3984375, "rewards/format_reward/std": 0.2387082874774933, "rewards/mcq_exact_match_reward/mean": 0.359375, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "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": 1659.0, "completions/mean_length": 489.375, "completions/mean_terminated_length": 464.63494873046875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.008928571428571428, "frac_reward_zero_std": 0.125, "grad_norm": 54.3234748840332, "learning_rate": 6e-07, "loss": 0.0, "num_tokens": 521293.0, "reward": 0.26484376192092896, "reward_std": 0.27863362431526184, "rewards/format_reward/mean": 0.3046875, "rewards/format_reward/std": 0.24587368965148926, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "step": 4 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1320.0, "completions/mean_length": 341.03125, "completions/mean_terminated_length": 313.9365234375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.011160714285714286, "frac_reward_zero_std": 0.0, "grad_norm": 7.101670742034912, "learning_rate": 8e-07, "loss": -0.0, "num_tokens": 629095.0, "reward": 0.47734373807907104, "reward_std": 0.3878313899040222, "rewards/format_reward/mean": 0.3984375, "rewards/format_reward/std": 0.22146137058734894, "rewards/mcq_exact_match_reward/mean": 0.4375, "rewards/mcq_exact_match_reward/std": 0.5, "step": 5 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1347.0, "completions/max_terminated_length": 1347.0, "completions/mean_length": 456.40625, "completions/mean_terminated_length": 456.40625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.013392857142857142, "frac_reward_zero_std": 0.125, "grad_norm": 3.6064534187316895, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 738977.0, "reward": 0.23906248807907104, "reward_std": 0.1744898557662964, "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": 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": 1167.0, "completions/mean_length": 497.59375, "completions/mean_terminated_length": 472.9841613769531, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.015625, "frac_reward_zero_std": 0.0, "grad_norm": 5.452180862426758, "learning_rate": 9.99726628670463e-07, "loss": 0.0, "num_tokens": 869015.0, "reward": 0.3359374701976776, "reward_std": 0.35532253980636597, "rewards/format_reward/mean": 0.390625, "rewards/format_reward/std": 0.22658175230026245, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "step": 7 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1597.0, "completions/mean_length": 590.1875, "completions/mean_terminated_length": 543.1612548828125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.017857142857142856, "frac_reward_zero_std": 0.0, "grad_norm": 3.406330108642578, "learning_rate": 9.989068136093872e-07, "loss": -0.0, "num_tokens": 1002547.0, "reward": 0.19453123211860657, "reward_std": 0.23629224300384521, "rewards/format_reward/mean": 0.3828125, "rewards/format_reward/std": 0.2919853627681732, "rewards/mcq_exact_match_reward/mean": 0.15625, "rewards/mcq_exact_match_reward/std": 0.36596253514289856, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1929.0, "completions/mean_length": 652.171875, "completions/mean_terminated_length": 630.0159301757812, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.020089285714285716, "frac_reward_zero_std": 0.0, "grad_norm": 13.18515682220459, "learning_rate": 9.975414512725056e-07, "loss": 0.0, "num_tokens": 1159886.0, "reward": 0.20624998211860657, "reward_std": 0.3330276906490326, "rewards/format_reward/mean": 0.34375, "rewards/format_reward/std": 0.279951810836792, "rewards/mcq_exact_match_reward/mean": 0.171875, "rewards/mcq_exact_match_reward/std": 0.38025420904159546, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1746.0, "completions/mean_length": 436.953125, "completions/mean_terminated_length": 384.9838562011719, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.022321428571428572, "frac_reward_zero_std": 0.125, "grad_norm": 11.200824737548828, "learning_rate": 9.956320346634875e-07, "loss": 0.0, "num_tokens": 1293211.0, "reward": 0.21015623211860657, "reward_std": 0.2866670489311218, "rewards/format_reward/mean": 0.3828125, "rewards/format_reward/std": 0.21347814798355103, "rewards/mcq_exact_match_reward/mean": 0.171875, "rewards/mcq_exact_match_reward/std": 0.38025420904159546, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1867.0, "completions/mean_length": 574.53125, "completions/mean_terminated_length": 527.0, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.024553571428571428, "frac_reward_zero_std": 0.125, "grad_norm": 4.861082553863525, "learning_rate": 9.931806517013612e-07, "loss": -0.0, "num_tokens": 1417333.0, "reward": 0.48828125, "reward_std": 0.3558398485183716, "rewards/format_reward/mean": 0.3515625, "rewards/format_reward/std": 0.24688033759593964, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "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": 0.046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1187.0, "completions/mean_length": 495.875, "completions/mean_terminated_length": 419.54095458984375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.026785714285714284, "frac_reward_zero_std": 0.0, "grad_norm": 5.903264999389648, "learning_rate": 9.901899829374047e-07, "loss": -0.0, "num_tokens": 1553901.0, "reward": 0.26953125, "reward_std": 0.31017062067985535, "rewards/format_reward/mean": 0.3515625, "rewards/format_reward/std": 0.2302463799715042, "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": 0.0, "completions/max_length": 1661.0, "completions/max_terminated_length": 1661.0, "completions/mean_length": 663.484375, "completions/mean_terminated_length": 663.484375, "completions/min_length": 135.0, "completions/min_terminated_length": 135.0, "epoch": 0.029017857142857144, "frac_reward_zero_std": 0.0, "grad_norm": 0.8600173592567444, "learning_rate": 9.866632986240029e-07, "loss": 0.0, "num_tokens": 1692508.0, "reward": 0.28515625, "reward_std": 0.3843300938606262, "rewards/format_reward/mean": 0.5078125, "rewards/format_reward/std": 0.2744719088077545, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "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": 1094.0, "completions/max_terminated_length": 1094.0, "completions/mean_length": 412.296875, "completions/mean_terminated_length": 412.296875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.03125, "frac_reward_zero_std": 0.0, "grad_norm": 7.163034915924072, "learning_rate": 9.826044551386742e-07, "loss": 0.0, "num_tokens": 1834575.0, "reward": 0.22968748211860657, "reward_std": 0.33986449241638184, "rewards/format_reward/mean": 0.421875, "rewards/format_reward/std": 0.29839184880256653, "rewards/mcq_exact_match_reward/mean": 0.1875, "rewards/mcq_exact_match_reward/std": 0.39339789748191833, "step": 14 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1661.0, "completions/mean_length": 461.03125, "completions/mean_terminated_length": 435.84130859375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.033482142857142856, "frac_reward_zero_std": 0.125, "grad_norm": 45.824703216552734, "learning_rate": 9.780178907671788e-07, "loss": -0.0, "num_tokens": 1976929.0, "reward": 0.2437499761581421, "reward_std": 0.28385066986083984, "rewards/format_reward/mean": 0.40625, "rewards/format_reward/std": 0.21593283116817474, "rewards/mcq_exact_match_reward/mean": 0.203125, "rewards/mcq_exact_match_reward/std": 0.40550529956817627, "step": 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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1875.0, "completions/mean_length": 434.109375, "completions/mean_terminated_length": 408.4920959472656, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.03571428571428571, "frac_reward_zero_std": 0.125, "grad_norm": 19.630359649658203, "learning_rate": 9.729086208503173e-07, "loss": -0.0, "num_tokens": 2096800.0, "reward": 0.3140624761581421, "reward_std": 0.379569947719574, "rewards/format_reward/mean": 0.484375, "rewards/format_reward/std": 0.2514837086200714, "rewards/mcq_exact_match_reward/mean": 0.265625, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "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": 1570.0, "completions/max_terminated_length": 1570.0, "completions/mean_length": 454.203125, "completions/mean_terminated_length": 454.203125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.03794642857142857, "frac_reward_zero_std": 0.0, "grad_norm": 11.983710289001465, "learning_rate": 9.672822322997304e-07, "loss": 0.0, "num_tokens": 2208829.0, "reward": 0.2523437440395355, "reward_std": 0.32918667793273926, "rewards/format_reward/mean": 0.4921875, "rewards/format_reward/std": 0.32723212242126465, "rewards/mcq_exact_match_reward/mean": 0.203125, "rewards/mcq_exact_match_reward/std": 0.40550529956817627, "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": 1754.0, "completions/mean_length": 478.90625, "completions/mean_terminated_length": 454.0000305175781, "completions/min_length": 15.0, "completions/min_terminated_length": 15.0, "epoch": 0.04017857142857143, "frac_reward_zero_std": 0.0, "grad_norm": 8.29274845123291, "learning_rate": 9.611448774886923e-07, "loss": -0.0, "num_tokens": 2327735.0, "reward": 0.36328125, "reward_std": 0.4432058334350586, "rewards/format_reward/mean": 0.5078125, "rewards/format_reward/std": 0.3391421139240265, "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": 0.0, "completions/max_length": 1485.0, "completions/max_terminated_length": 1485.0, "completions/mean_length": 477.625, "completions/mean_terminated_length": 477.625, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.04241071428571429, "frac_reward_zero_std": 0.125, "grad_norm": 2.7630865573883057, "learning_rate": 9.545032675245813e-07, "loss": -0.0, "num_tokens": 2467943.0, "reward": 0.2187499850988388, "reward_std": 0.23242416977882385, "rewards/format_reward/mean": 0.625, "rewards/format_reward/std": 0.28171807527542114, "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": 0.0, "completions/max_length": 1642.0, "completions/max_terminated_length": 1642.0, "completions/mean_length": 444.34375, "completions/mean_terminated_length": 444.34375, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.044642857142857144, "frac_reward_zero_std": 0.0, "grad_norm": 11.05822467803955, "learning_rate": 9.473646649103817e-07, "loss": -0.0, "num_tokens": 2613149.0, "reward": 0.2718749940395355, "reward_std": 0.39301908016204834, "rewards/format_reward/mean": 0.6875, "rewards/format_reward/std": 0.3021090030670166, "rewards/mcq_exact_match_reward/mean": 0.203125, "rewards/mcq_exact_match_reward/std": 0.40550529956817627, "step": 20 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1194.0, "completions/mean_length": 381.609375, "completions/mean_terminated_length": 355.15875244140625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.046875, "frac_reward_zero_std": 0.0, "grad_norm": 9.699033737182617, "learning_rate": 9.397368756032444e-07, "loss": -0.0, "num_tokens": 2746852.0, "reward": 0.578906238079071, "reward_std": 0.42781421542167664, "rewards/format_reward/mean": 0.6328125, "rewards/format_reward/std": 0.34780704975128174, "rewards/mcq_exact_match_reward/mean": 0.515625, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 953.0, "completions/mean_length": 427.015625, "completions/mean_terminated_length": 374.7257995605469, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.049107142857142856, "frac_reward_zero_std": 0.125, "grad_norm": 15.209996223449707, "learning_rate": 9.316282404787869e-07, "loss": 0.0, "num_tokens": 2852493.0, "reward": 0.33203125, "reward_std": 0.26917463541030884, "rewards/format_reward/mean": 0.6640625, "rewards/format_reward/std": 0.29620200395584106, "rewards/mcq_exact_match_reward/mean": 0.265625, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "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": 1582.0, "completions/mean_length": 488.78125, "completions/mean_terminated_length": 464.0317687988281, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.05133928571428571, "frac_reward_zero_std": 0.0, "grad_norm": 9.77275562286377, "learning_rate": 9.230476262104676e-07, "loss": 0.0, "num_tokens": 2998071.0, "reward": 0.45390623807907104, "reward_std": 0.35065871477127075, "rewards/format_reward/mean": 0.6328125, "rewards/format_reward/std": 0.3117053508758545, "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": 0.0, "completions/max_length": 1471.0, "completions/max_terminated_length": 1471.0, "completions/mean_length": 308.28125, "completions/mean_terminated_length": 308.28125, "completions/min_length": 10.0, "completions/min_terminated_length": 10.0, "epoch": 0.05357142857142857, "frac_reward_zero_std": 0.0, "grad_norm": 13.308282852172852, "learning_rate": 9.1400441557401e-07, "loss": -0.0, "num_tokens": 3120153.0, "reward": 0.45624998211860657, "reward_std": 0.4762032628059387, "rewards/format_reward/mean": 0.65625, "rewards/format_reward/std": 0.3766498267650604, "rewards/mcq_exact_match_reward/mean": 0.390625, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "step": 24 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1514.0, "completions/mean_length": 244.515625, "completions/mean_terminated_length": 215.88890075683594, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.05580357142857143, "frac_reward_zero_std": 0.125, "grad_norm": 7.151384353637695, "learning_rate": 9.045084971874737e-07, "loss": 0.0, "num_tokens": 3233706.0, "reward": 0.33906251192092896, "reward_std": 0.29599475860595703, "rewards/format_reward/mean": 0.890625, "rewards/format_reward/std": 0.24346621334552765, "rewards/mcq_exact_match_reward/mean": 0.25, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "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": 0.046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 523.0, "completions/mean_length": 184.375, "completions/mean_terminated_length": 92.72130584716797, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.05803571428571429, "frac_reward_zero_std": 0.0, "grad_norm": 16.7641658782959, "learning_rate": 8.945702546981968e-07, "loss": -0.0, "num_tokens": 3326194.0, "reward": 0.28437501192092896, "reward_std": 0.25573086738586426, "rewards/format_reward/mean": 0.8125, "rewards/format_reward/std": 0.3149704039096832, "rewards/mcq_exact_match_reward/mean": 0.203125, "rewards/mcq_exact_match_reward/std": 0.40550529956817627, "step": 26 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 796.0, "completions/max_terminated_length": 796.0, "completions/mean_length": 120.421875, "completions/mean_terminated_length": 120.421875, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.060267857142857144, "frac_reward_zero_std": 0.0, "grad_norm": 14.986625671386719, "learning_rate": 8.842005554284295e-07, "loss": 0.0, "num_tokens": 3416909.0, "reward": 0.4007812440395355, "reward_std": 0.24702563881874084, "rewards/format_reward/mean": 0.8828125, "rewards/format_reward/std": 0.21347814798355103, "rewards/mcq_exact_match_reward/mean": 0.3125, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "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": 671.0, "completions/mean_length": 161.109375, "completions/mean_terminated_length": 131.1587371826172, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.0625, "frac_reward_zero_std": 0.0, "grad_norm": 19.596328735351562, "learning_rate": 8.734107384920769e-07, "loss": -0.0, "num_tokens": 3530108.0, "reward": 0.31953126192092896, "reward_std": 0.3115481734275818, "rewards/format_reward/mean": 0.8515625, "rewards/format_reward/std": 0.26246222853660583, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "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": 251.0, "completions/mean_length": 53.40625, "completions/mean_terminated_length": 21.74603271484375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.06473214285714286, "frac_reward_zero_std": 0.375, "grad_norm": 18.60841941833496, "learning_rate": 8.622126023955445e-07, "loss": 0.0, "num_tokens": 3634206.0, "reward": 0.375, "reward_std": 0.19709977507591248, "rewards/format_reward/mean": 0.9375, "rewards/format_reward/std": 0.208927720785141, "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": 302.0, "completions/max_terminated_length": 302.0, "completions/mean_length": 18.265625, "completions/mean_terminated_length": 18.265625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.06696428571428571, "frac_reward_zero_std": 0.25, "grad_norm": 25.47766876220703, "learning_rate": 8.506183921362442e-07, "loss": -0.0, "num_tokens": 3731807.0, "reward": 0.4078125059604645, "reward_std": 0.15645788609981537, "rewards/format_reward/mean": 0.953125, "rewards/format_reward/std": 0.19352105259895325, "rewards/mcq_exact_match_reward/mean": 0.3125, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "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": 274.0, "completions/max_terminated_length": 274.0, "completions/mean_length": 26.40625, "completions/mean_terminated_length": 26.40625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.06919642857142858, "frac_reward_zero_std": 0.375, "grad_norm": 9.368828773498535, "learning_rate": 8.386407858128706e-07, "loss": -0.0, "num_tokens": 3811265.0, "reward": 0.5328124761581421, "reward_std": 0.17126736044883728, "rewards/format_reward/mean": 0.953125, "rewards/format_reward/std": 0.14689241349697113, "rewards/mcq_exact_match_reward/mean": 0.4375, "rewards/mcq_exact_match_reward/std": 0.5, "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": 670.0, "completions/max_terminated_length": 670.0, "completions/mean_length": 30.796875, "completions/mean_terminated_length": 30.796875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.07142857142857142, "frac_reward_zero_std": 0.125, "grad_norm": 19.652753829956055, "learning_rate": 8.262928807620843e-07, "loss": 0.0, "num_tokens": 3912804.0, "reward": 0.4515625238418579, "reward_std": 0.303753525018692, "rewards/format_reward/mean": 0.921875, "rewards/format_reward/std": 0.20351573824882507, "rewards/mcq_exact_match_reward/mean": 0.359375, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "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": 24.0, "completions/max_terminated_length": 24.0, "completions/mean_length": 13.140625, "completions/mean_terminated_length": 13.140625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.07366071428571429, "frac_reward_zero_std": 0.375, "grad_norm": 14.285265922546387, "learning_rate": 8.135881792367685e-07, "loss": -0.0, "num_tokens": 4002165.0, "reward": 0.38046878576278687, "reward_std": 0.21527712047100067, "rewards/format_reward/mean": 0.9921875, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.28125, "rewards/mcq_exact_match_reward/std": 0.4531635046005249, "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": 152.0, "completions/max_terminated_length": 152.0, "completions/mean_length": 15.0625, "completions/mean_terminated_length": 15.0625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.07589285714285714, "frac_reward_zero_std": 0.625, "grad_norm": 16.814939498901367, "learning_rate": 8.005405736415125e-07, "loss": -0.0, "num_tokens": 4095201.0, "reward": 0.5210937857627869, "reward_std": 0.16973936557769775, "rewards/format_reward/mean": 0.9921875, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.421875, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "step": 34 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 163.0, "completions/max_terminated_length": 163.0, "completions/mean_length": 15.375, "completions/mean_terminated_length": 15.375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.078125, "frac_reward_zero_std": 0.5, "grad_norm": 22.958662033081055, "learning_rate": 7.871643313414718e-07, "loss": -0.0, "num_tokens": 4189593.0, "reward": 0.7085937261581421, "reward_std": 0.17108294367790222, "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": 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": 598.0, "completions/max_terminated_length": 598.0, "completions/mean_length": 23.421875, "completions/mean_terminated_length": 23.421875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.08035714285714286, "frac_reward_zero_std": 0.5, "grad_norm": 28.723560333251953, "learning_rate": 7.734740790612136e-07, "loss": -0.0, "num_tokens": 4275940.0, "reward": 0.7515624761581421, "reward_std": 0.17957819998264313, "rewards/format_reward/mean": 0.953125, "rewards/format_reward/std": 0.14689241349697113, "rewards/mcq_exact_match_reward/mean": 0.65625, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "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": 236.0, "completions/max_terminated_length": 236.0, "completions/mean_length": 17.046875, "completions/mean_terminated_length": 17.046875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.08258928571428571, "frac_reward_zero_std": 0.375, "grad_norm": 11.786319732666016, "learning_rate": 7.594847868906076e-07, "loss": 0.0, "num_tokens": 4371623.0, "reward": 0.5320312976837158, "reward_std": 0.2051372081041336, "rewards/format_reward/mean": 0.9453125, "rewards/format_reward/std": 0.2203386276960373, "rewards/mcq_exact_match_reward/mean": 0.4375, "rewards/mcq_exact_match_reward/std": 0.5, "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": 834.0, "completions/max_terminated_length": 834.0, "completions/mean_length": 25.78125, "completions/mean_terminated_length": 25.78125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.08482142857142858, "frac_reward_zero_std": 0.625, "grad_norm": 10.964075088500977, "learning_rate": 7.452117519152541e-07, "loss": 0.0, "num_tokens": 4484985.0, "reward": 0.3968750238418579, "reward_std": 0.1530819833278656, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "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": 289.0, "completions/max_terminated_length": 289.0, "completions/mean_length": 17.765625, "completions/mean_terminated_length": 17.765625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.08705357142857142, "frac_reward_zero_std": 0.5, "grad_norm": 16.558963775634766, "learning_rate": 7.306705814893439e-07, "loss": 0.0, "num_tokens": 4574418.0, "reward": 0.6937500238418579, "reward_std": 0.2177756279706955, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.59375, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 509.0, "completions/mean_length": 57.703125, "completions/mean_terminated_length": 26.111112594604492, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.08928571428571429, "frac_reward_zero_std": 0.375, "grad_norm": 14.512017250061035, "learning_rate": 7.158771761692464e-07, "loss": 0.0, "num_tokens": 4689095.0, "reward": 0.2835937738418579, "reward_std": 0.21870046854019165, "rewards/format_reward/mean": 0.9609375, "rewards/format_reward/std": 0.18483558297157288, "rewards/mcq_exact_match_reward/mean": 0.1875, "rewards/mcq_exact_match_reward/std": 0.39339789748191833, "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": 16.0, "completions/max_terminated_length": 16.0, "completions/mean_length": 12.875, "completions/mean_terminated_length": 12.875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.09151785714285714, "frac_reward_zero_std": 0.5, "grad_norm": 20.142478942871094, "learning_rate": 7.008477123264847e-07, "loss": 0.0, "num_tokens": 4805367.0, "reward": 0.4750000238418579, "reward_std": 0.23356688022613525, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.375, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 52.0, "completions/mean_length": 45.671875, "completions/mean_terminated_length": 13.88888931274414, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.09375, "frac_reward_zero_std": 0.5, "grad_norm": 9.706896781921387, "learning_rate": 6.855986244591103e-07, "loss": 0.0, "num_tokens": 4899706.0, "reward": 0.47343751788139343, "reward_std": 0.18119612336158752, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.375, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "step": 42 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.375, "completions/mean_terminated_length": 12.375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.09598214285714286, "frac_reward_zero_std": 0.375, "grad_norm": 18.802722930908203, "learning_rate": 6.701465872208216e-07, "loss": -0.0, "num_tokens": 4991314.0, "reward": 0.6742187738418579, "reward_std": 0.2858760356903076, "rewards/format_reward/mean": 0.9609375, "rewards/format_reward/std": 0.13524486124515533, "rewards/mcq_exact_match_reward/mean": 0.578125, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "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": 1068.0, "completions/max_terminated_length": 1068.0, "completions/mean_length": 83.96875, "completions/mean_terminated_length": 83.96875, "completions/min_length": 10.0, "completions/min_terminated_length": 10.0, "epoch": 0.09821428571428571, "frac_reward_zero_std": 0.625, "grad_norm": 9.638358116149902, "learning_rate": 6.545084971874736e-07, "loss": -0.0, "num_tokens": 5119120.0, "reward": 0.5414062738418579, "reward_std": 0.11321918666362762, "rewards/format_reward/mean": 0.8828125, "rewards/format_reward/std": 0.31800705194473267, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "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": 23.0, "completions/max_terminated_length": 23.0, "completions/mean_length": 13.15625, "completions/mean_terminated_length": 13.15625, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.10044642857142858, "frac_reward_zero_std": 0.625, "grad_norm": 11.176258087158203, "learning_rate": 6.387014543809223e-07, "loss": 0.0, "num_tokens": 5232290.0, "reward": 0.4593750238418579, "reward_std": 0.15992169082164764, "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": 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": 576.0, "completions/max_terminated_length": 576.0, "completions/mean_length": 24.078125, "completions/mean_terminated_length": 24.078125, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.10267857142857142, "frac_reward_zero_std": 0.625, "grad_norm": 10.300027847290039, "learning_rate": 6.227427435703995e-07, "loss": 0.0, "num_tokens": 5317335.0, "reward": 0.3656250238418579, "reward_std": 0.13258251547813416, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.265625, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "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": 446.0, "completions/max_terminated_length": 446.0, "completions/mean_length": 28.484375, "completions/mean_terminated_length": 28.484375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.10491071428571429, "frac_reward_zero_std": 0.625, "grad_norm": 10.058446884155273, "learning_rate": 6.066498153718734e-07, "loss": 0.0, "num_tokens": 5412398.0, "reward": 0.7351562976837158, "reward_std": 0.10825187712907791, "rewards/format_reward/mean": 0.9453125, "rewards/format_reward/std": 0.2203386276960373, "rewards/mcq_exact_match_reward/mean": 0.640625, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 13.0, "completions/mean_terminated_length": 13.0, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.10714285714285714, "frac_reward_zero_std": 0.5, "grad_norm": 17.154449462890625, "learning_rate": 5.90440267166055e-07, "loss": 0.0, "num_tokens": 5507870.0, "reward": 0.5375000238418579, "reward_std": 0.23356688022613525, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.4375, "rewards/mcq_exact_match_reward/std": 0.5, "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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 13.0, "completions/mean_terminated_length": 13.0, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.109375, "frac_reward_zero_std": 0.875, "grad_norm": 6.836947917938232, "learning_rate": 5.741318238559209e-07, "loss": -0.0, "num_tokens": 5599030.0, "reward": 0.4437500238418579, "reward_std": 0.0578637570142746, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.34375, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "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": 96.0, "completions/max_terminated_length": 96.0, "completions/mean_length": 14.1875, "completions/mean_terminated_length": 14.1875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.11160714285714286, "frac_reward_zero_std": 0.625, "grad_norm": 11.428613662719727, "learning_rate": 5.577423184847931e-07, "loss": 0.0, "num_tokens": 5684162.0, "reward": 0.5375000238418579, "reward_std": 0.1962025910615921, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "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": 59.0, "completions/max_terminated_length": 59.0, "completions/mean_length": 14.65625, "completions/mean_terminated_length": 14.65625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.11383928571428571, "frac_reward_zero_std": 0.75, "grad_norm": 9.1956205368042, "learning_rate": 5.412896727361662e-07, "loss": 0.0, "num_tokens": 5784700.0, "reward": 0.9265625476837158, "reward_std": 0.10504289716482162, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.828125, "rewards/mcq_exact_match_reward/std": 0.38025420904159546, "step": 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": 20.0, "completions/max_terminated_length": 20.0, "completions/mean_length": 13.75, "completions/mean_terminated_length": 13.75, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.11607142857142858, "frac_reward_zero_std": 0.75, "grad_norm": 14.34941577911377, "learning_rate": 5.247918773366111e-07, "loss": 0.0, "num_tokens": 5906444.0, "reward": 0.6937500238418579, "reward_std": 0.1246790662407875, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.59375, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "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": 158.0, "completions/max_terminated_length": 158.0, "completions/mean_length": 15.640625, "completions/mean_terminated_length": 15.640625, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.11830357142857142, "frac_reward_zero_std": 0.75, "grad_norm": 12.550302505493164, "learning_rate": 5.082669723831793e-07, "loss": 0.0, "num_tokens": 5984605.0, "reward": 0.37812501192092896, "reward_std": 0.12817874550819397, "rewards/format_reward/mean": 0.96875, "rewards/format_reward/std": 0.17536810040473938, "rewards/mcq_exact_match_reward/mean": 0.28125, "rewards/mcq_exact_match_reward/std": 0.4531635046005249, "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": 267.0, "completions/max_terminated_length": 267.0, "completions/mean_length": 16.953125, "completions/mean_terminated_length": 16.953125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.12053571428571429, "frac_reward_zero_std": 0.625, "grad_norm": 13.103774070739746, "learning_rate": 4.917330276168208e-07, "loss": 0.0, "num_tokens": 6093658.0, "reward": 0.3031250238418579, "reward_std": 0.1804211586713791, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.203125, "rewards/mcq_exact_match_reward/std": 0.40550529956817627, "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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.96875, "completions/mean_terminated_length": 12.96875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.12276785714285714, "frac_reward_zero_std": 0.625, "grad_norm": 12.446603775024414, "learning_rate": 4.752081226633888e-07, "loss": -0.0, "num_tokens": 6205448.0, "reward": 0.4437500238418579, "reward_std": 0.17570313811302185, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.34375, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "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": 142.0, "completions/max_terminated_length": 142.0, "completions/mean_length": 15.03125, "completions/mean_terminated_length": 15.03125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.125, "frac_reward_zero_std": 0.75, "grad_norm": 9.64340877532959, "learning_rate": 4.5871032726383385e-07, "loss": 0.0, "num_tokens": 6282914.0, "reward": 0.8656250238418579, "reward_std": 0.12255740165710449, "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": 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": 0.0, "completions/max_length": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.703125, "completions/mean_terminated_length": 12.703125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.12723214285714285, "frac_reward_zero_std": 0.75, "grad_norm": 26.199546813964844, "learning_rate": 4.4225768151520694e-07, "loss": 0.0, "num_tokens": 6385383.0, "reward": 0.3796875476837158, "reward_std": 0.060756947845220566, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.08768405020236969, "rewards/mcq_exact_match_reward/mean": 0.28125, "rewards/mcq_exact_match_reward/std": 0.4531635046005249, "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": 32.0, "completions/max_terminated_length": 32.0, "completions/mean_length": 13.21875, "completions/mean_terminated_length": 13.21875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.12946428571428573, "frac_reward_zero_std": 0.625, "grad_norm": 13.70984172821045, "learning_rate": 4.258681761440789e-07, "loss": 0.0, "num_tokens": 6477933.0, "reward": 0.6781250238418579, "reward_std": 0.1530819982290268, "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": 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": 19.0, "completions/max_terminated_length": 19.0, "completions/mean_length": 13.015625, "completions/mean_terminated_length": 13.015625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.13169642857142858, "frac_reward_zero_std": 0.875, "grad_norm": 6.949735641479492, "learning_rate": 4.095597328339452e-07, "loss": 0.0, "num_tokens": 6571062.0, "reward": 0.2562500238418579, "reward_std": 0.0578637570142746, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.15625, "rewards/mcq_exact_match_reward/std": 0.36596253514289856, "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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.8125, "completions/mean_terminated_length": 12.8125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.13392857142857142, "frac_reward_zero_std": 0.75, "grad_norm": 7.103476047515869, "learning_rate": 3.9335018462812664e-07, "loss": 0.0, "num_tokens": 6660866.0, "reward": 0.5843750238418579, "reward_std": 0.1315089464187622, "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": 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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.75, "completions/mean_terminated_length": 12.75, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.13616071428571427, "frac_reward_zero_std": 0.75, "grad_norm": 7.502201080322266, "learning_rate": 3.772572564296004e-07, "loss": 0.0, "num_tokens": 6765018.0, "reward": 0.6937500238418579, "reward_std": 0.10888782143592834, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.59375, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.921875, "completions/mean_terminated_length": 12.921875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.13839285714285715, "frac_reward_zero_std": 1.0, "grad_norm": 0.0, "learning_rate": 3.612985456190778e-07, "loss": 0.0, "num_tokens": 6867501.0, "reward": 0.4750000238418579, "reward_std": 0.0, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.375, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.734375, "completions/mean_terminated_length": 12.734375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.140625, "frac_reward_zero_std": 0.75, "grad_norm": 9.751093864440918, "learning_rate": 3.454915028125263e-07, "loss": 0.0, "num_tokens": 6960756.0, "reward": 0.6937500238418579, "reward_std": 0.10888782143592834, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.59375, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "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": 20.0, "completions/max_terminated_length": 20.0, "completions/mean_length": 12.953125, "completions/mean_terminated_length": 12.953125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.14285714285714285, "frac_reward_zero_std": 0.625, "grad_norm": 11.77578353881836, "learning_rate": 3.2985341277917846e-07, "loss": -0.0, "num_tokens": 7061401.0, "reward": 0.4750000238418579, "reward_std": 0.16675157845020294, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.375, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 231.0, "completions/max_terminated_length": 231.0, "completions/mean_length": 16.09375, "completions/mean_terminated_length": 16.09375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.14508928571428573, "frac_reward_zero_std": 0.75, "grad_norm": 8.541019439697266, "learning_rate": 3.1440137554088953e-07, "loss": -0.0, "num_tokens": 7166319.0, "reward": 0.5992187261581421, "reward_std": 0.08812715113162994, "rewards/format_reward/mean": 0.9921875, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.5, "rewards/mcq_exact_match_reward/std": 0.5039526224136353, "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": 0.0, "completions/max_length": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.59375, "completions/mean_terminated_length": 12.59375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.14732142857142858, "frac_reward_zero_std": 0.75, "grad_norm": 13.157929420471191, "learning_rate": 2.9915228767351535e-07, "loss": -0.0, "num_tokens": 7281365.0, "reward": 0.6312500238418579, "reward_std": 0.10888782143592834, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.53125, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "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": 17.0, "completions/max_terminated_length": 17.0, "completions/mean_length": 12.890625, "completions/mean_terminated_length": 12.890625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.14955357142857142, "frac_reward_zero_std": 0.625, "grad_norm": 8.833831787109375, "learning_rate": 2.841228238307536e-07, "loss": 0.0, "num_tokens": 7401510.0, "reward": 0.5671875476837158, "reward_std": 0.11330723762512207, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.46875, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "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": 17.0, "completions/max_terminated_length": 17.0, "completions/mean_length": 12.828125, "completions/mean_terminated_length": 12.828125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.15178571428571427, "frac_reward_zero_std": 0.75, "grad_norm": 13.64869213104248, "learning_rate": 2.6932941851065615e-07, "loss": 0.0, "num_tokens": 7495923.0, "reward": 1.006250023841858, "reward_std": 0.1293872892856598, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.90625, "rewards/mcq_exact_match_reward/std": 0.29378482699394226, "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": 112.0, "completions/max_terminated_length": 112.0, "completions/mean_length": 14.421875, "completions/mean_terminated_length": 14.421875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.15401785714285715, "frac_reward_zero_std": 0.5, "grad_norm": 23.549795150756836, "learning_rate": 2.547882480847461e-07, "loss": 0.0, "num_tokens": 7601574.0, "reward": 0.6468750238418579, "reward_std": 0.2109457552433014, "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": 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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.65625, "completions/mean_terminated_length": 12.65625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.15625, "frac_reward_zero_std": 0.25, "grad_norm": 22.055912017822266, "learning_rate": 2.4051521310939254e-07, "loss": -0.0, "num_tokens": 7706296.0, "reward": 0.5531250238418579, "reward_std": 0.3330284357070923, "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": 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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.90625, "completions/mean_terminated_length": 12.90625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.15848214285714285, "frac_reward_zero_std": 0.5, "grad_norm": 15.57496452331543, "learning_rate": 2.2652592093878665e-07, "loss": 0.0, "num_tokens": 7794866.0, "reward": 0.6156250238418579, "reward_std": 0.2561880350112915, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.515625, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "step": 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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.734375, "completions/mean_terminated_length": 12.734375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.16071428571428573, "frac_reward_zero_std": 0.75, "grad_norm": 11.250571250915527, "learning_rate": 2.128356686585282e-07, "loss": 0.0, "num_tokens": 7895809.0, "reward": 0.7250000238418579, "reward_std": 0.0883883461356163, "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": 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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.671875, "completions/mean_terminated_length": 12.671875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.16294642857142858, "frac_reward_zero_std": 0.625, "grad_norm": 10.136702537536621, "learning_rate": 1.9945942635848745e-07, "loss": 0.0, "num_tokens": 7981116.0, "reward": 0.8500000238418579, "reward_std": 0.1462520956993103, "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": 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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.859375, "completions/mean_terminated_length": 12.859375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.16517857142857142, "frac_reward_zero_std": 0.5, "grad_norm": 16.42728042602539, "learning_rate": 1.8641182076323148e-07, "loss": 0.0, "num_tokens": 8097019.0, "reward": 0.4906250238418579, "reward_std": 0.23144522309303284, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.390625, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "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": 16.0, "completions/max_terminated_length": 16.0, "completions/mean_length": 12.984375, "completions/mean_terminated_length": 12.984375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.16741071428571427, "frac_reward_zero_std": 0.25, "grad_norm": 29.946748733520508, "learning_rate": 1.7370711923791564e-07, "loss": -0.0, "num_tokens": 8194562.0, "reward": 0.6000000238418579, "reward_std": 0.3335031569004059, "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": 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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.9375, "completions/mean_terminated_length": 12.9375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.16964285714285715, "frac_reward_zero_std": 0.75, "grad_norm": 5.190943241119385, "learning_rate": 1.6135921418712955e-07, "loss": 0.0, "num_tokens": 8288814.0, "reward": 0.8656250238418579, "reward_std": 0.10205793380737305, "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": 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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.890625, "completions/mean_terminated_length": 12.890625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.171875, "frac_reward_zero_std": 0.75, "grad_norm": 10.818709373474121, "learning_rate": 1.493816078637557e-07, "loss": 0.0, "num_tokens": 8397335.0, "reward": 0.6000000238418579, "reward_std": 0.1293872892856598, "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": 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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 17.0, "completions/mean_length": 45.03125, "completions/mean_terminated_length": 13.238096237182617, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.17410714285714285, "frac_reward_zero_std": 0.875, "grad_norm": 18.82288932800293, "learning_rate": 1.3778739760445552e-07, "loss": 0.0, "num_tokens": 8480905.0, "reward": 0.53125, "reward_std": 0.07349684089422226, "rewards/format_reward/mean": 0.9375, "rewards/format_reward/std": 0.24397502839565277, "rewards/mcq_exact_match_reward/mean": 0.4375, "rewards/mcq_exact_match_reward/std": 0.5, "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": 27.0, "completions/max_terminated_length": 27.0, "completions/mean_length": 12.859375, "completions/mean_terminated_length": 12.859375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.17633928571428573, "frac_reward_zero_std": 0.625, "grad_norm": 19.21259117126465, "learning_rate": 1.2658926150792322e-07, "loss": 0.0, "num_tokens": 8588392.0, "reward": 0.5843750238418579, "reward_std": 0.1893727034330368, "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": 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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.5, "completions/mean_terminated_length": 12.5, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.17857142857142858, "frac_reward_zero_std": 0.75, "grad_norm": 10.078824043273926, "learning_rate": 1.1579944457157059e-07, "loss": 0.0, "num_tokens": 8715584.0, "reward": 0.4750000238418579, "reward_std": 0.1293872892856598, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.375, "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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.765625, "completions/mean_terminated_length": 12.765625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.18080357142857142, "frac_reward_zero_std": 0.625, "grad_norm": 28.567230224609375, "learning_rate": 1.0542974530180327e-07, "loss": 0.0, "num_tokens": 8816137.0, "reward": 0.6625000238418579, "reward_std": 0.20044592022895813, "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": 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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.84375, "completions/mean_terminated_length": 12.84375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.18303571428571427, "frac_reward_zero_std": 0.75, "grad_norm": 14.302781105041504, "learning_rate": 9.549150281252632e-08, "loss": 0.0, "num_tokens": 8931047.0, "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": 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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.71875, "completions/mean_terminated_length": 12.71875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.18526785714285715, "frac_reward_zero_std": 0.5, "grad_norm": 15.248523712158203, "learning_rate": 8.599558442598998e-08, "loss": 0.0, "num_tokens": 9010469.0, "reward": 0.6625000238418579, "reward_std": 0.2177756428718567, "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": 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": 932.0, "completions/max_terminated_length": 932.0, "completions/mean_length": 81.953125, "completions/mean_terminated_length": 81.953125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1875, "frac_reward_zero_std": 0.5, "grad_norm": 13.376705169677734, "learning_rate": 7.695237378953224e-08, "loss": 0.0, "num_tokens": 9163530.0, "reward": 0.5914062857627869, "reward_std": 0.16998624801635742, "rewards/format_reward/mean": 0.9140625, "rewards/format_reward/std": 0.244862899184227, "rewards/mcq_exact_match_reward/mean": 0.5, "rewards/mcq_exact_match_reward/std": 0.5039526224136353, "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": 17.0, "completions/max_terminated_length": 17.0, "completions/mean_length": 12.640625, "completions/mean_terminated_length": 12.640625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.18973214285714285, "frac_reward_zero_std": 0.75, "grad_norm": 7.489809036254883, "learning_rate": 6.837175952121304e-08, "loss": -0.0, "num_tokens": 9283595.0, "reward": 0.4906250238418579, "reward_std": 0.12255740165710449, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.390625, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.6875, "completions/mean_terminated_length": 12.6875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.19196428571428573, "frac_reward_zero_std": 1.0, "grad_norm": 0.0, "learning_rate": 6.026312439675551e-08, "loss": 0.0, "num_tokens": 9405879.0, "reward": 0.9750000238418579, "reward_std": 0.0, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.875, "rewards/mcq_exact_match_reward/std": 0.3333333432674408, "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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.875, "completions/mean_terminated_length": 12.875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.19419642857142858, "frac_reward_zero_std": 0.75, "grad_norm": 10.523398399353027, "learning_rate": 5.263533508961826e-08, "loss": -0.0, "num_tokens": 9503751.0, "reward": 0.7562500238418579, "reward_std": 0.1246790662407875, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.65625, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "step": 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": 30.0, "completions/max_terminated_length": 30.0, "completions/mean_length": 14.53125, "completions/mean_terminated_length": 14.53125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.19642857142857142, "frac_reward_zero_std": 0.625, "grad_norm": 9.63086223602295, "learning_rate": 4.549673247541874e-08, "loss": 0.0, "num_tokens": 9607329.0, "reward": 0.589062511920929, "reward_std": 0.12014693021774292, "rewards/format_reward/mean": 0.890625, "rewards/format_reward/std": 0.3145764470100403, "rewards/mcq_exact_match_reward/mean": 0.5, "rewards/mcq_exact_match_reward/std": 0.5039526224136353, "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": 0.0, "completions/max_length": 17.0, "completions/max_terminated_length": 17.0, "completions/mean_length": 13.09375, "completions/mean_terminated_length": 13.09375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.19866071428571427, "frac_reward_zero_std": 0.75, "grad_norm": 9.03646183013916, "learning_rate": 3.8855122511307626e-08, "loss": 0.0, "num_tokens": 9682999.0, "reward": 0.9437500238418579, "reward_std": 0.0883883461356163, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.84375, "rewards/mcq_exact_match_reward/std": 0.36596253514289856, "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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.78125, "completions/mean_terminated_length": 12.78125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.20089285714285715, "frac_reward_zero_std": 1.0, "grad_norm": 0.0, "learning_rate": 3.271776770026963e-08, "loss": 0.0, "num_tokens": 9779081.0, "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": 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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.796875, "completions/mean_terminated_length": 12.796875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.203125, "frac_reward_zero_std": 0.5, "grad_norm": 16.86722183227539, "learning_rate": 2.7091379149682682e-08, "loss": -0.0, "num_tokens": 9884644.0, "reward": 0.6156250238418579, "reward_std": 0.2109457552433014, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.515625, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "step": 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": 16.0, "completions/max_terminated_length": 16.0, "completions/mean_length": 12.953125, "completions/mean_terminated_length": 12.953125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.20535714285714285, "frac_reward_zero_std": 0.5, "grad_norm": 23.426349639892578, "learning_rate": 2.1982109232821176e-08, "loss": 0.0, "num_tokens": 9966561.0, "reward": 0.5375000238418579, "reward_std": 0.2314550280570984, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.4375, "rewards/mcq_exact_match_reward/std": 0.5, "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": 17.0, "completions/max_terminated_length": 17.0, "completions/mean_length": 12.75, "completions/mean_terminated_length": 12.75, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.20758928571428573, "frac_reward_zero_std": 0.625, "grad_norm": 13.51418399810791, "learning_rate": 1.7395544861325718e-08, "loss": 0.0, "num_tokens": 10087297.0, "reward": 0.7875000238418579, "reward_std": 0.1462521106004715, "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": 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": 0.0, "completions/max_length": 17.0, "completions/max_terminated_length": 17.0, "completions/mean_length": 12.953125, "completions/mean_terminated_length": 12.953125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.20982142857142858, "frac_reward_zero_std": 0.625, "grad_norm": 10.580665588378906, "learning_rate": 1.3336701375997127e-08, "loss": 0.0, "num_tokens": 10184174.0, "reward": 0.7250000238418579, "reward_std": 0.1462521106004715, "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": 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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.828125, "completions/mean_terminated_length": 12.828125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.21205357142857142, "frac_reward_zero_std": 0.875, "grad_norm": 17.36741065979004, "learning_rate": 9.810017062595321e-09, "loss": 0.0, "num_tokens": 10266723.0, "reward": 0.7093750238418579, "reward_std": 0.04419417306780815, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.609375, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.875, "completions/mean_terminated_length": 12.875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.21428571428571427, "frac_reward_zero_std": 1.0, "grad_norm": 0.0, "learning_rate": 6.819348298638839e-09, "loss": 0.0, "num_tokens": 10370171.0, "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": 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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.703125, "completions/mean_terminated_length": 12.703125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.21651785714285715, "frac_reward_zero_std": 0.625, "grad_norm": 8.915359497070312, "learning_rate": 4.367965336512403e-09, "loss": 0.0, "num_tokens": 10459808.0, "reward": 0.6156250238418579, "reward_std": 0.17358146607875824, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.515625, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "step": 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": 15.0, "completions/max_terminated_length": 15.0, "completions/mean_length": 12.75, "completions/mean_terminated_length": 12.75, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.21875, "frac_reward_zero_std": 0.75, "grad_norm": 9.11575698852539, "learning_rate": 2.458548727494292e-09, "loss": -0.0, "num_tokens": 10554536.0, "reward": 0.9281250238418579, "reward_std": 0.10205793380737305, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.828125, "rewards/mcq_exact_match_reward/std": 0.38025420904159546, "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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.859375, "completions/mean_terminated_length": 12.859375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.22098214285714285, "frac_reward_zero_std": 0.25, "grad_norm": 18.249919891357422, "learning_rate": 1.0931863906127325e-09, "loss": 0.0, "num_tokens": 10656863.0, "reward": 0.6468750238418579, "reward_std": 0.3356248140335083, "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": 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": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 12.6875, "completions/mean_terminated_length": 12.6875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.22321428571428573, "frac_reward_zero_std": 0.625, "grad_norm": 11.268608093261719, "learning_rate": 2.733713295369755e-10, "loss": -0.0, "num_tokens": 10773331.0, "reward": 0.4125000238418579, "reward_std": 0.1462521106004715, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.3125, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "step": 100 } ], "logging_steps": 1, "max_steps": 100, "num_input_tokens_seen": 10773331, "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 }