{ "best_global_step": null, "best_metric": null, "best_model_checkpoint": null, "epoch": 0.11160714285714286, "eval_steps": 500, "global_step": 50, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2011.0, "completions/mean_length": 528.109375, "completions/mean_terminated_length": 503.9841613769531, "completions/min_length": 132.0, "completions/min_terminated_length": 132.0, "epoch": 0.002232142857142857, "frac_reward_zero_std": 0.0, "grad_norm": 0.9355294108390808, "learning_rate": 0.0, "loss": -0.0, "num_tokens": 121119.0, "reward": 0.26953125, "reward_std": 0.37513452768325806, "rewards/format_reward/mean": 0.3515625, "rewards/format_reward/std": 0.27716949582099915, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "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": 1501.0, "completions/max_terminated_length": 1501.0, "completions/mean_length": 427.296875, "completions/mean_terminated_length": 427.296875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.004464285714285714, "frac_reward_zero_std": 0.0, "grad_norm": 12.184840202331543, "learning_rate": 2e-07, "loss": -0.0, "num_tokens": 243994.0, "reward": 0.3226562440395355, "reward_std": 0.42037320137023926, "rewards/format_reward/mean": 0.4140625, "rewards/format_reward/std": 0.2605654299259186, "rewards/mcq_exact_match_reward/mean": 0.28125, "rewards/mcq_exact_match_reward/std": 0.4531635046005249, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1531.0, "completions/mean_length": 544.0, "completions/mean_terminated_length": 495.4838562011719, "completions/min_length": 10.0, "completions/min_terminated_length": 10.0, "epoch": 0.006696428571428571, "frac_reward_zero_std": 0.0, "grad_norm": 4.455195903778076, "learning_rate": 4e-07, "loss": 0.0, "num_tokens": 382882.0, "reward": 0.15468749403953552, "reward_std": 0.24445468187332153, "rewards/format_reward/mean": 0.296875, "rewards/format_reward/std": 0.2630521357059479, "rewards/mcq_exact_match_reward/mean": 0.125, "rewards/mcq_exact_match_reward/std": 0.3333333432674408, "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": 0.046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1767.0, "completions/mean_length": 446.203125, "completions/mean_terminated_length": 367.42620849609375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.008928571428571428, "frac_reward_zero_std": 0.0, "grad_norm": 26.635173797607422, "learning_rate": 6e-07, "loss": -0.0, "num_tokens": 498159.0, "reward": 0.17031247913837433, "reward_std": 0.2350117713212967, "rewards/format_reward/mean": 0.296875, "rewards/format_reward/std": 0.24750742316246033, "rewards/mcq_exact_match_reward/mean": 0.140625, "rewards/mcq_exact_match_reward/std": 0.3503824472427368, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1692.0, "completions/mean_length": 401.390625, "completions/mean_terminated_length": 348.274169921875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.011160714285714286, "frac_reward_zero_std": 0.0, "grad_norm": 22.468753814697266, "learning_rate": 8e-07, "loss": -0.0, "num_tokens": 606640.0, "reward": 0.3656249940395355, "reward_std": 0.4346795678138733, "rewards/format_reward/mean": 0.375, "rewards/format_reward/std": 0.30860671401023865, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "step": 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": 1654.0, "completions/max_terminated_length": 1654.0, "completions/mean_length": 294.734375, "completions/mean_terminated_length": 294.734375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.013392857142857142, "frac_reward_zero_std": 0.0, "grad_norm": 12.87374210357666, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 722783.0, "reward": 0.30937498807907104, "reward_std": 0.37140360474586487, "rewards/format_reward/mean": 0.4375, "rewards/format_reward/std": 0.22712838649749756, "rewards/mcq_exact_match_reward/mean": 0.265625, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "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": 0.0, "completions/max_length": 1824.0, "completions/max_terminated_length": 1824.0, "completions/mean_length": 525.296875, "completions/mean_terminated_length": 525.296875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.015625, "frac_reward_zero_std": 0.0, "grad_norm": 3.5019431114196777, "learning_rate": 9.99726628670463e-07, "loss": 0.0, "num_tokens": 869410.0, "reward": 0.06562500447034836, "reward_std": 0.1088000163435936, "rewards/format_reward/mean": 0.34375, "rewards/format_reward/std": 0.3196600377559662, "rewards/mcq_exact_match_reward/mean": 0.03125, "rewards/mcq_exact_match_reward/std": 0.17536810040473938, "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": 1983.0, "completions/mean_length": 567.390625, "completions/mean_terminated_length": 519.6290283203125, "completions/min_length": 153.0, "completions/min_terminated_length": 153.0, "epoch": 0.017857142857142856, "frac_reward_zero_std": 0.0, "grad_norm": 1.1946548223495483, "learning_rate": 9.989068136093872e-07, "loss": -0.0, "num_tokens": 1001739.0, "reward": 0.36406245827674866, "reward_std": 0.3973655700683594, "rewards/format_reward/mean": 0.359375, "rewards/format_reward/std": 0.24346621334552765, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "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": 1868.0, "completions/mean_length": 313.71875, "completions/mean_terminated_length": 286.19049072265625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.020089285714285716, "frac_reward_zero_std": 0.125, "grad_norm": 14.30042552947998, "learning_rate": 9.975414512725056e-07, "loss": 0.0, "num_tokens": 1119969.0, "reward": 0.3515625, "reward_std": 0.34339872002601624, "rewards/format_reward/mean": 0.390625, "rewards/format_reward/std": 0.2741328477859497, "rewards/mcq_exact_match_reward/mean": 0.3125, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "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": 1410.0, "completions/mean_length": 439.203125, "completions/mean_terminated_length": 387.3064270019531, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.022321428571428572, "frac_reward_zero_std": 0.0, "grad_norm": 8.828161239624023, "learning_rate": 9.956320346634875e-07, "loss": -0.0, "num_tokens": 1259382.0, "reward": 0.43281248211860657, "reward_std": 0.40298882126808167, "rewards/format_reward/mean": 0.421875, "rewards/format_reward/std": 0.27048972249031067, "rewards/mcq_exact_match_reward/mean": 0.390625, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "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": 1549.0, "completions/mean_length": 396.5, "completions/mean_terminated_length": 343.2257995605469, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.024553571428571428, "frac_reward_zero_std": 0.25, "grad_norm": 10.044999122619629, "learning_rate": 9.931806517013612e-07, "loss": -0.0, "num_tokens": 1385830.0, "reward": 0.19921875, "reward_std": 0.276921808719635, "rewards/format_reward/mean": 0.4296875, "rewards/format_reward/std": 0.1751912236213684, "rewards/mcq_exact_match_reward/mean": 0.15625, "rewards/mcq_exact_match_reward/std": 0.36596253514289856, "step": 11 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1999.0, "completions/mean_length": 445.40625, "completions/mean_terminated_length": 419.9682922363281, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.026785714285714284, "frac_reward_zero_std": 0.0, "grad_norm": 6.090798377990723, "learning_rate": 9.901899829374047e-07, "loss": 0.0, "num_tokens": 1481472.0, "reward": 0.515625, "reward_std": 0.46172386407852173, "rewards/format_reward/mean": 0.46875, "rewards/format_reward/std": 0.33184191584587097, "rewards/mcq_exact_match_reward/mean": 0.46875, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1761.0, "completions/mean_length": 581.34375, "completions/mean_terminated_length": 534.0322265625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.029017857142857144, "frac_reward_zero_std": 0.0, "grad_norm": 10.792880058288574, "learning_rate": 9.866632986240029e-07, "loss": 0.0, "num_tokens": 1609150.0, "reward": 0.16874998807907104, "reward_std": 0.27205121517181396, "rewards/format_reward/mean": 0.4375, "rewards/format_reward/std": 0.28867512941360474, "rewards/mcq_exact_match_reward/mean": 0.125, "rewards/mcq_exact_match_reward/std": 0.3333333432674408, "step": 13 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1903.0, "completions/mean_length": 568.484375, "completions/mean_terminated_length": 495.7212829589844, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.03125, "frac_reward_zero_std": 0.125, "grad_norm": 8.320207595825195, "learning_rate": 9.826044551386742e-07, "loss": 0.0, "num_tokens": 1751637.0, "reward": 0.4781249761581421, "reward_std": 0.3682000935077667, "rewards/format_reward/mean": 0.40625, "rewards/format_reward/std": 0.233588308095932, "rewards/mcq_exact_match_reward/mean": 0.4375, "rewards/mcq_exact_match_reward/std": 0.5, "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": 1356.0, "completions/max_terminated_length": 1356.0, "completions/mean_length": 370.96875, "completions/mean_terminated_length": 370.96875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.033482142857142856, "frac_reward_zero_std": 0.0, "grad_norm": 14.99644947052002, "learning_rate": 9.780178907671788e-07, "loss": 0.0, "num_tokens": 1877739.0, "reward": 0.5812499523162842, "reward_std": 0.45640966296195984, "rewards/format_reward/mean": 0.5, "rewards/format_reward/std": 0.26726123690605164, "rewards/mcq_exact_match_reward/mean": 0.53125, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "step": 15 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1714.0, "completions/max_terminated_length": 1714.0, "completions/mean_length": 428.390625, "completions/mean_terminated_length": 428.390625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.03571428571428571, "frac_reward_zero_std": 0.0, "grad_norm": 12.95833683013916, "learning_rate": 9.729086208503173e-07, "loss": 0.0, "num_tokens": 2002500.0, "reward": 0.516406238079071, "reward_std": 0.47470909357070923, "rewards/format_reward/mean": 0.6328125, "rewards/format_reward/std": 0.2987033724784851, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "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": 1412.0, "completions/max_terminated_length": 1412.0, "completions/mean_length": 255.796875, "completions/mean_terminated_length": 255.796875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.03794642857142857, "frac_reward_zero_std": 0.375, "grad_norm": 11.912960052490234, "learning_rate": 9.672822322997304e-07, "loss": -0.0, "num_tokens": 2108527.0, "reward": 0.49296873807907104, "reward_std": 0.28548362851142883, "rewards/format_reward/mean": 0.7109375, "rewards/format_reward/std": 0.2643453776836395, "rewards/mcq_exact_match_reward/mean": 0.421875, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "step": 17 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1803.0, "completions/max_terminated_length": 1803.0, "completions/mean_length": 231.40625, "completions/mean_terminated_length": 231.40625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.04017857142857143, "frac_reward_zero_std": 0.25, "grad_norm": 20.999282836914062, "learning_rate": 9.611448774886923e-07, "loss": -0.0, "num_tokens": 2223681.0, "reward": 0.5687500238418579, "reward_std": 0.35626089572906494, "rewards/format_reward/mean": 0.6875, "rewards/format_reward/std": 0.3021090030670166, "rewards/mcq_exact_match_reward/mean": 0.5, "rewards/mcq_exact_match_reward/std": 0.5039526224136353, "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": 1676.0, "completions/max_terminated_length": 1676.0, "completions/mean_length": 235.875, "completions/mean_terminated_length": 235.875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.04241071428571429, "frac_reward_zero_std": 0.0, "grad_norm": 15.144044876098633, "learning_rate": 9.545032675245813e-07, "loss": -0.0, "num_tokens": 2335865.0, "reward": 0.5882812738418579, "reward_std": 0.3023134469985962, "rewards/format_reward/mean": 0.7265625, "rewards/format_reward/std": 0.3077012598514557, "rewards/mcq_exact_match_reward/mean": 0.515625, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "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": 971.0, "completions/max_terminated_length": 971.0, "completions/mean_length": 119.5, "completions/mean_terminated_length": 119.5, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.044642857142857144, "frac_reward_zero_std": 0.0, "grad_norm": 14.69491195678711, "learning_rate": 9.473646649103817e-07, "loss": 0.0, "num_tokens": 2448625.0, "reward": 0.37812501192092896, "reward_std": 0.402826726436615, "rewards/format_reward/mean": 0.8125, "rewards/format_reward/std": 0.28867512941360474, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "step": 20 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1299.0, "completions/max_terminated_length": 1299.0, "completions/mean_length": 88.09375, "completions/mean_terminated_length": 88.09375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.046875, "frac_reward_zero_std": 0.125, "grad_norm": 37.22273635864258, "learning_rate": 9.397368756032444e-07, "loss": -0.0, "num_tokens": 2531639.0, "reward": 0.62109375, "reward_std": 0.245207741856575, "rewards/format_reward/mean": 0.8984375, "rewards/format_reward/std": 0.22146137058734894, "rewards/mcq_exact_match_reward/mean": 0.53125, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "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": 822.0, "completions/max_terminated_length": 822.0, "completions/mean_length": 32.828125, "completions/mean_terminated_length": 32.828125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.049107142857142856, "frac_reward_zero_std": 0.125, "grad_norm": 29.922048568725586, "learning_rate": 9.316282404787869e-07, "loss": 0.0, "num_tokens": 2643372.0, "reward": 0.5687500238418579, "reward_std": 0.3287316560745239, "rewards/format_reward/mean": 0.84375, "rewards/format_reward/std": 0.233588308095932, "rewards/mcq_exact_match_reward/mean": 0.484375, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "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": 1162.0, "completions/max_terminated_length": 1162.0, "completions/mean_length": 81.90625, "completions/mean_terminated_length": 81.90625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.05133928571428571, "frac_reward_zero_std": 0.125, "grad_norm": 12.92499828338623, "learning_rate": 9.230476262104676e-07, "loss": 0.0, "num_tokens": 2747486.0, "reward": 0.4671875238418579, "reward_std": 0.3623044788837433, "rewards/format_reward/mean": 0.921875, "rewards/format_reward/std": 0.20351573824882507, "rewards/mcq_exact_match_reward/mean": 0.375, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 567.0, "completions/max_terminated_length": 567.0, "completions/mean_length": 41.765625, "completions/mean_terminated_length": 41.765625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.05357142857142857, "frac_reward_zero_std": 0.25, "grad_norm": 24.27642250061035, "learning_rate": 9.1400441557401e-07, "loss": 0.0, "num_tokens": 2828591.0, "reward": 0.39218753576278687, "reward_std": 0.33768123388290405, "rewards/format_reward/mean": 0.953125, "rewards/format_reward/std": 0.14689241349697113, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "step": 24 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 50.0, "completions/max_terminated_length": 50.0, "completions/mean_length": 12.109375, "completions/mean_terminated_length": 12.109375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.05580357142857143, "frac_reward_zero_std": 0.375, "grad_norm": 10.304224967956543, "learning_rate": 9.045084971874737e-07, "loss": -0.0, "num_tokens": 2929198.0, "reward": 0.811718761920929, "reward_std": 0.2634497284889221, "rewards/format_reward/mean": 0.9296875, "rewards/format_reward/std": 0.19654127955436707, "rewards/mcq_exact_match_reward/mean": 0.71875, "rewards/mcq_exact_match_reward/std": 0.4531635046005249, "step": 25 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 926.0, "completions/max_terminated_length": 926.0, "completions/mean_length": 68.359375, "completions/mean_terminated_length": 68.359375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.05803571428571429, "frac_reward_zero_std": 0.25, "grad_norm": 19.061513900756836, "learning_rate": 8.945702546981968e-07, "loss": 0.0, "num_tokens": 3025421.0, "reward": 0.37109375, "reward_std": 0.33500707149505615, "rewards/format_reward/mean": 0.8984375, "rewards/format_reward/std": 0.2387082874774933, "rewards/mcq_exact_match_reward/mean": 0.28125, "rewards/mcq_exact_match_reward/std": 0.4531635046005249, "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": 12.0, "completions/max_terminated_length": 12.0, "completions/mean_length": 12.0, "completions/mean_terminated_length": 12.0, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.060267857142857144, "frac_reward_zero_std": 0.25, "grad_norm": 16.958600997924805, "learning_rate": 8.842005554284295e-07, "loss": 0.0, "num_tokens": 3141461.0, "reward": 0.5062500238418579, "reward_std": 0.3335031569004059, "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": 27 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 40.0, "completions/max_terminated_length": 40.0, "completions/mean_length": 11.71875, "completions/mean_terminated_length": 11.71875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0625, "frac_reward_zero_std": 0.5, "grad_norm": 9.72388744354248, "learning_rate": 8.734107384920769e-07, "loss": -0.0, "num_tokens": 3240051.0, "reward": 0.38984376192092896, "reward_std": 0.21241585910320282, "rewards/format_reward/mean": 0.9296875, "rewards/format_reward/std": 0.19654127955436707, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "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": 90.0, "completions/max_terminated_length": 90.0, "completions/mean_length": 13.328125, "completions/mean_terminated_length": 13.328125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.06473214285714286, "frac_reward_zero_std": 0.25, "grad_norm": 11.685744285583496, "learning_rate": 8.622126023955445e-07, "loss": 0.0, "num_tokens": 3325080.0, "reward": 0.48906251788139343, "reward_std": 0.22539028525352478, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.390625, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "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": 46.0, "completions/max_terminated_length": 46.0, "completions/mean_length": 14.34375, "completions/mean_terminated_length": 14.34375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.06696428571428571, "frac_reward_zero_std": 0.625, "grad_norm": 19.394933700561523, "learning_rate": 8.506183921362442e-07, "loss": -0.0, "num_tokens": 3426318.0, "reward": 0.45468753576278687, "reward_std": 0.1829523742198944, "rewards/format_reward/mean": 0.953125, "rewards/format_reward/std": 0.21304203569889069, "rewards/mcq_exact_match_reward/mean": 0.359375, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "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": 1043.0, "completions/max_terminated_length": 1043.0, "completions/mean_length": 31.171875, "completions/mean_terminated_length": 31.171875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.06919642857142858, "frac_reward_zero_std": 0.5, "grad_norm": 10.200380325317383, "learning_rate": 8.386407858128706e-07, "loss": -0.0, "num_tokens": 3505169.0, "reward": 0.4281250238418579, "reward_std": 0.19939783215522766, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "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": 184.0, "completions/max_terminated_length": 184.0, "completions/mean_length": 18.5625, "completions/mean_terminated_length": 18.5625, "completions/min_length": 11.0, "completions/min_terminated_length": 11.0, "epoch": 0.07142857142857142, "frac_reward_zero_std": 0.75, "grad_norm": 6.094886302947998, "learning_rate": 8.262928807620843e-07, "loss": -0.0, "num_tokens": 3605933.0, "reward": 0.44687503576278687, "reward_std": 0.1315089464187622, "rewards/format_reward/mean": 0.875, "rewards/format_reward/std": 0.3333333432674408, "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": 143.0, "completions/max_terminated_length": 143.0, "completions/mean_length": 22.40625, "completions/mean_terminated_length": 22.40625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.07366071428571429, "frac_reward_zero_std": 0.5, "grad_norm": 25.130882263183594, "learning_rate": 8.135881792367685e-07, "loss": 0.0, "num_tokens": 3683175.0, "reward": 0.33281251788139343, "reward_std": 0.19129052758216858, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "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": 20.0, "completions/max_terminated_length": 20.0, "completions/mean_length": 12.71875, "completions/mean_terminated_length": 12.71875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.07589285714285714, "frac_reward_zero_std": 0.625, "grad_norm": 10.324932098388672, "learning_rate": 8.005405736415125e-07, "loss": 0.0, "num_tokens": 3761261.0, "reward": 0.8812500238418579, "reward_std": 0.16675157845020294, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.78125, "rewards/mcq_exact_match_reward/std": 0.4166666865348816, "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": 157.0, "completions/max_terminated_length": 157.0, "completions/mean_length": 14.390625, "completions/mean_terminated_length": 14.390625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.078125, "frac_reward_zero_std": 0.75, "grad_norm": 35.36943817138672, "learning_rate": 7.871643313414718e-07, "loss": 0.0, "num_tokens": 3860406.0, "reward": 0.643750011920929, "reward_std": 0.06803441047668457, "rewards/format_reward/mean": 0.96875, "rewards/format_reward/std": 0.12198751419782639, "rewards/mcq_exact_match_reward/mean": 0.546875, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 35 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 118.0, "completions/max_terminated_length": 118.0, "completions/mean_length": 16.1875, "completions/mean_terminated_length": 16.1875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.08035714285714286, "frac_reward_zero_std": 0.75, "grad_norm": 9.857135772705078, "learning_rate": 7.734740790612136e-07, "loss": 0.0, "num_tokens": 3950842.0, "reward": 0.6625000238418579, "reward_std": 0.1157275140285492, "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": 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": 565.0, "completions/max_terminated_length": 565.0, "completions/mean_length": 21.984375, "completions/mean_terminated_length": 21.984375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.08258928571428571, "frac_reward_zero_std": 0.625, "grad_norm": 17.679731369018555, "learning_rate": 7.594847868906076e-07, "loss": 0.0, "num_tokens": 4056305.0, "reward": 0.5843750238418579, "reward_std": 0.17358146607875824, "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": 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": 664.0, "completions/max_terminated_length": 664.0, "completions/mean_length": 27.4375, "completions/mean_terminated_length": 27.4375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.08482142857142858, "frac_reward_zero_std": 0.625, "grad_norm": 7.502120494842529, "learning_rate": 7.452117519152541e-07, "loss": 0.0, "num_tokens": 4151749.0, "reward": 0.3812500238418579, "reward_std": 0.1462521106004715, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.28125, "rewards/mcq_exact_match_reward/std": 0.4531635046005249, "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": 19.0, "completions/max_terminated_length": 19.0, "completions/mean_length": 12.109375, "completions/mean_terminated_length": 12.109375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.08705357142857142, "frac_reward_zero_std": 0.75, "grad_norm": 7.0521392822265625, "learning_rate": 7.306705814893439e-07, "loss": 0.0, "num_tokens": 4230140.0, "reward": 0.5828125476837158, "reward_std": 0.04861358925700188, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.484375, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "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": 875.0, "completions/max_terminated_length": 875.0, "completions/mean_length": 32.421875, "completions/mean_terminated_length": 32.421875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.08928571428571429, "frac_reward_zero_std": 0.75, "grad_norm": 11.019020080566406, "learning_rate": 7.158771761692464e-07, "loss": 0.0, "num_tokens": 4325367.0, "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": 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": 66.0, "completions/max_terminated_length": 66.0, "completions/mean_length": 13.0, "completions/mean_terminated_length": 13.0, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.09151785714285714, "frac_reward_zero_std": 1.0, "grad_norm": 0.0, "learning_rate": 7.008477123264847e-07, "loss": 0.0, "num_tokens": 4422711.0, "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": 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": 44.0, "completions/max_terminated_length": 44.0, "completions/mean_length": 13.265625, "completions/mean_terminated_length": 13.265625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.09375, "frac_reward_zero_std": 0.5, "grad_norm": 17.952268600463867, "learning_rate": 6.855986244591103e-07, "loss": 0.0, "num_tokens": 4507048.0, "reward": 0.5296875238418579, "reward_std": 0.12682846188545227, "rewards/format_reward/mean": 0.921875, "rewards/format_reward/std": 0.27048972249031067, "rewards/mcq_exact_match_reward/mean": 0.4375, "rewards/mcq_exact_match_reward/std": 0.5, "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": 296.0, "completions/max_terminated_length": 296.0, "completions/mean_length": 18.1875, "completions/mean_terminated_length": 18.1875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.09598214285714286, "frac_reward_zero_std": 0.75, "grad_norm": 5.754403591156006, "learning_rate": 6.701465872208216e-07, "loss": 0.0, "num_tokens": 4605428.0, "reward": 0.7093750238418579, "reward_std": 0.10205793380737305, "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": 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": 1060.0, "completions/max_terminated_length": 1060.0, "completions/mean_length": 41.875, "completions/mean_terminated_length": 41.875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.09821428571428571, "frac_reward_zero_std": 0.375, "grad_norm": 18.143627166748047, "learning_rate": 6.545084971874736e-07, "loss": 0.0, "num_tokens": 4702924.0, "reward": 0.5671875476837158, "reward_std": 0.22903475165367126, "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": 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": 19.0, "completions/max_terminated_length": 19.0, "completions/mean_length": 14.015625, "completions/mean_terminated_length": 14.015625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.10044642857142858, "frac_reward_zero_std": 0.625, "grad_norm": 12.582730293273926, "learning_rate": 6.387014543809223e-07, "loss": 0.0, "num_tokens": 4826389.0, "reward": 0.6312500238418579, "reward_std": 0.1462521106004715, "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": 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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 115.0, "completions/mean_length": 48.5625, "completions/mean_terminated_length": 16.825397491455078, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.10267857142857142, "frac_reward_zero_std": 0.375, "grad_norm": 29.780662536621094, "learning_rate": 6.227427435703995e-07, "loss": -0.0, "num_tokens": 4927657.0, "reward": 0.45781251788139343, "reward_std": 0.26638925075531006, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.359375, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "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": 656.0, "completions/max_terminated_length": 656.0, "completions/mean_length": 29.828125, "completions/mean_terminated_length": 29.828125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.10491071428571429, "frac_reward_zero_std": 0.75, "grad_norm": 8.41478443145752, "learning_rate": 6.066498153718734e-07, "loss": 0.0, "num_tokens": 5018206.0, "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": 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": 478.0, "completions/max_terminated_length": 478.0, "completions/mean_length": 20.40625, "completions/mean_terminated_length": 20.40625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.10714285714285714, "frac_reward_zero_std": 0.75, "grad_norm": 10.784804344177246, "learning_rate": 5.90440267166055e-07, "loss": -0.0, "num_tokens": 5120456.0, "reward": 0.5375000238418579, "reward_std": 0.10888782143592834, "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": 18.0, "completions/max_terminated_length": 18.0, "completions/mean_length": 12.625, "completions/mean_terminated_length": 12.625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.109375, "frac_reward_zero_std": 0.5, "grad_norm": 21.239360809326172, "learning_rate": 5.741318238559209e-07, "loss": 0.0, "num_tokens": 5192800.0, "reward": 0.870312511920929, "reward_std": 0.24963238835334778, "rewards/format_reward/mean": 0.890625, "rewards/format_reward/std": 0.3145764470100403, "rewards/mcq_exact_match_reward/mean": 0.78125, "rewards/mcq_exact_match_reward/std": 0.4166666865348816, "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": 113.0, "completions/max_terminated_length": 113.0, "completions/mean_length": 15.75, "completions/mean_terminated_length": 15.75, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.11160714285714286, "frac_reward_zero_std": 0.375, "grad_norm": 18.710081100463867, "learning_rate": 5.577423184847931e-07, "loss": 0.0, "num_tokens": 5301304.0, "reward": 0.4750000238418579, "reward_std": 0.2709311842918396, "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": 50 } ], "logging_steps": 1, "max_steps": 100, "num_input_tokens_seen": 5301304, "num_train_epochs": 1, "save_steps": 50, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": false }, "attributes": {} } }, "total_flos": 0.0, "train_batch_size": 1, "trial_name": null, "trial_params": null }