{ "best_global_step": null, "best_metric": null, "best_model_checkpoint": null, "epoch": 0.25, "eval_steps": 500, "global_step": 100, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1784.0, "completions/mean_length": 578.84375, "completions/mean_terminated_length": 555.5238647460938, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "epoch": 0.0025, "frac_reward_zero_std": 0.0, "grad_norm": 6.383375644683838, "learning_rate": 0.0, "loss": 0.0, "num_tokens": 128782.0, "reward": 0.11171874403953552, "reward_std": 0.20118111371994019, "rewards/format_reward/mean": 0.3359375, "rewards/format_reward/std": 0.2680720090866089, "rewards/mcq_exact_match_reward/mean": 0.078125, "rewards/mcq_exact_match_reward/std": 0.27048972249031067, "step": 1 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1313.0, "completions/mean_length": 452.078125, "completions/mean_terminated_length": 400.5967712402344, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.005, "frac_reward_zero_std": 0.0, "grad_norm": 11.18359375, "learning_rate": 2e-07, "loss": -0.0, "num_tokens": 245859.0, "reward": 0.19062498211860657, "reward_std": 0.27377599477767944, "rewards/format_reward/mean": 0.34375, "rewards/format_reward/std": 0.29378482699394226, "rewards/mcq_exact_match_reward/mean": 0.15625, "rewards/mcq_exact_match_reward/std": 0.36596253514289856, "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": 1247.0, "completions/mean_length": 440.0625, "completions/mean_terminated_length": 414.5397033691406, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0075, "frac_reward_zero_std": 0.0, "grad_norm": 15.244134902954102, "learning_rate": 4e-07, "loss": -0.0, "num_tokens": 372479.0, "reward": 0.2570312023162842, "reward_std": 0.25679144263267517, "rewards/format_reward/mean": 0.3828125, "rewards/format_reward/std": 0.2634054720401764, "rewards/mcq_exact_match_reward/mean": 0.21875, "rewards/mcq_exact_match_reward/std": 0.4166666865348816, "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": 0.0, "completions/max_length": 1025.0, "completions/max_terminated_length": 1025.0, "completions/mean_length": 460.390625, "completions/mean_terminated_length": 460.390625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.01, "frac_reward_zero_std": 0.0, "grad_norm": 7.774961471557617, "learning_rate": 6e-07, "loss": 0.0, "num_tokens": 483936.0, "reward": 0.14531248807907104, "reward_std": 0.1944168359041214, "rewards/format_reward/mean": 0.359375, "rewards/format_reward/std": 0.24346621334552765, "rewards/mcq_exact_match_reward/mean": 0.109375, "rewards/mcq_exact_match_reward/std": 0.3145764470100403, "step": 4 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1556.0, "completions/max_terminated_length": 1556.0, "completions/mean_length": 506.5625, "completions/mean_terminated_length": 506.5625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.0125, "frac_reward_zero_std": 0.125, "grad_norm": 3.9458694458007812, "learning_rate": 8e-07, "loss": -0.0, "num_tokens": 586852.0, "reward": 0.3507812023162842, "reward_std": 0.4010940194129944, "rewards/format_reward/mean": 0.3828125, "rewards/format_reward/std": 0.2313210517168045, "rewards/mcq_exact_match_reward/mean": 0.3125, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "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": 1834.0, "completions/max_terminated_length": 1834.0, "completions/mean_length": 627.625, "completions/mean_terminated_length": 627.625, "completions/min_length": 128.0, "completions/min_terminated_length": 128.0, "epoch": 0.015, "frac_reward_zero_std": 0.0, "grad_norm": 1.2029720544815063, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 720428.0, "reward": 0.12890625, "reward_std": 0.18070563673973083, "rewards/format_reward/mean": 0.3515625, "rewards/format_reward/std": 0.2302463799715042, "rewards/mcq_exact_match_reward/mean": 0.09375, "rewards/mcq_exact_match_reward/std": 0.29378482699394226, "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": 1736.0, "completions/mean_length": 511.796875, "completions/mean_terminated_length": 487.4127197265625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.0175, "frac_reward_zero_std": 0.25, "grad_norm": 6.808908462524414, "learning_rate": 9.99726628670463e-07, "loss": -0.0, "num_tokens": 815263.0, "reward": 0.10078124701976776, "reward_std": 0.17372390627861023, "rewards/format_reward/mean": 0.2265625, "rewards/format_reward/std": 0.29452258348464966, "rewards/mcq_exact_match_reward/mean": 0.078125, "rewards/mcq_exact_match_reward/std": 0.27048972249031067, "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": 0.0, "completions/max_length": 1077.0, "completions/max_terminated_length": 1077.0, "completions/mean_length": 314.015625, "completions/mean_terminated_length": 314.015625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.02, "frac_reward_zero_std": 0.125, "grad_norm": 28.814695358276367, "learning_rate": 9.989068136093872e-07, "loss": -0.0, "num_tokens": 905440.0, "reward": 0.24140623211860657, "reward_std": 0.3407360315322876, "rewards/format_reward/mean": 0.3828125, "rewards/format_reward/std": 0.3052735924720764, "rewards/mcq_exact_match_reward/mean": 0.203125, "rewards/mcq_exact_match_reward/std": 0.40550529956817627, "step": 8 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1785.0, "completions/max_terminated_length": 1785.0, "completions/mean_length": 599.3125, "completions/mean_terminated_length": 599.3125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0225, "frac_reward_zero_std": 0.0, "grad_norm": 9.018033981323242, "learning_rate": 9.975414512725056e-07, "loss": 0.0, "num_tokens": 1029812.0, "reward": 0.2867187261581421, "reward_std": 0.369856595993042, "rewards/format_reward/mean": 0.3671875, "rewards/format_reward/std": 0.23974503576755524, "rewards/mcq_exact_match_reward/mean": 0.25, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "step": 9 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1523.0, "completions/mean_length": 485.984375, "completions/mean_terminated_length": 461.19049072265625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.025, "frac_reward_zero_std": 0.125, "grad_norm": 11.238686561584473, "learning_rate": 9.956320346634875e-07, "loss": 0.0, "num_tokens": 1117883.0, "reward": 0.33984375, "reward_std": 0.30625492334365845, "rewards/format_reward/mean": 0.4296875, "rewards/format_reward/std": 0.23345555365085602, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "step": 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": 0.046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1825.0, "completions/mean_length": 507.4375, "completions/mean_terminated_length": 431.672119140625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0275, "frac_reward_zero_std": 0.0, "grad_norm": 10.189136505126953, "learning_rate": 9.931806517013612e-07, "loss": -0.0, "num_tokens": 1228839.0, "reward": 0.2640624940395355, "reward_std": 0.33808624744415283, "rewards/format_reward/mean": 0.453125, "rewards/format_reward/std": 0.2777281701564789, "rewards/mcq_exact_match_reward/mean": 0.21875, "rewards/mcq_exact_match_reward/std": 0.4166666865348816, "step": 11 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1529.0, "completions/max_terminated_length": 1529.0, "completions/mean_length": 334.203125, "completions/mean_terminated_length": 334.203125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.03, "frac_reward_zero_std": 0.125, "grad_norm": 12.628010749816895, "learning_rate": 9.901899829374047e-07, "loss": 0.0, "num_tokens": 1329108.0, "reward": 0.38359373807907104, "reward_std": 0.2515554428100586, "rewards/format_reward/mean": 0.5546875, "rewards/format_reward/std": 0.2833639681339264, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "step": 12 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 992.0, "completions/mean_length": 378.015625, "completions/mean_terminated_length": 351.5079650878906, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0325, "frac_reward_zero_std": 0.0, "grad_norm": 8.161300659179688, "learning_rate": 9.866632986240029e-07, "loss": 0.0, "num_tokens": 1426573.0, "reward": 0.2265625, "reward_std": 0.38044559955596924, "rewards/format_reward/mean": 0.390625, "rewards/format_reward/std": 0.28824523091316223, "rewards/mcq_exact_match_reward/mean": 0.1875, "rewards/mcq_exact_match_reward/std": 0.39339789748191833, "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": 1640.0, "completions/max_terminated_length": 1640.0, "completions/mean_length": 416.640625, "completions/mean_terminated_length": 416.640625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.035, "frac_reward_zero_std": 0.25, "grad_norm": 15.543842315673828, "learning_rate": 9.826044551386742e-07, "loss": -0.0, "num_tokens": 1530902.0, "reward": 0.37187498807907104, "reward_std": 0.3173729181289673, "rewards/format_reward/mean": 0.4375, "rewards/format_reward/std": 0.24397502839565277, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "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": 1488.0, "completions/max_terminated_length": 1488.0, "completions/mean_length": 276.5625, "completions/mean_terminated_length": 276.5625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0375, "frac_reward_zero_std": 0.0, "grad_norm": 11.145133018493652, "learning_rate": 9.780178907671788e-07, "loss": 0.0, "num_tokens": 1621042.0, "reward": 0.3046875, "reward_std": 0.32315585017204285, "rewards/format_reward/mean": 0.546875, "rewards/format_reward/std": 0.21304203569889069, "rewards/mcq_exact_match_reward/mean": 0.25, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "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": 1014.0, "completions/max_terminated_length": 1014.0, "completions/mean_length": 327.515625, "completions/mean_terminated_length": 327.515625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.04, "frac_reward_zero_std": 0.125, "grad_norm": 11.127957344055176, "learning_rate": 9.729086208503173e-07, "loss": 0.0, "num_tokens": 1712915.0, "reward": 0.16796875, "reward_std": 0.16902750730514526, "rewards/format_reward/mean": 0.5859375, "rewards/format_reward/std": 0.28942593932151794, "rewards/mcq_exact_match_reward/mean": 0.109375, "rewards/mcq_exact_match_reward/std": 0.3145764470100403, "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": 1885.0, "completions/max_terminated_length": 1885.0, "completions/mean_length": 371.5, "completions/mean_terminated_length": 371.5, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0425, "frac_reward_zero_std": 0.25, "grad_norm": 5.605539321899414, "learning_rate": 9.672822322997304e-07, "loss": 0.0, "num_tokens": 1820587.0, "reward": 0.5023437142372131, "reward_std": 0.306302011013031, "rewards/format_reward/mean": 0.4921875, "rewards/format_reward/std": 0.18881812691688538, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "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": 1342.0, "completions/max_terminated_length": 1342.0, "completions/mean_length": 270.53125, "completions/mean_terminated_length": 270.53125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.045, "frac_reward_zero_std": 0.25, "grad_norm": 14.470562934875488, "learning_rate": 9.611448774886923e-07, "loss": -0.0, "num_tokens": 1911469.0, "reward": 0.45781248807907104, "reward_std": 0.3108960688114166, "rewards/format_reward/mean": 0.515625, "rewards/format_reward/std": 0.1985812783241272, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 735.0, "completions/mean_length": 287.84375, "completions/mean_terminated_length": 231.06451416015625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0475, "frac_reward_zero_std": 0.125, "grad_norm": 10.86657428741455, "learning_rate": 9.545032675245813e-07, "loss": 0.0, "num_tokens": 1998491.0, "reward": 0.36796873807907104, "reward_std": 0.35315409302711487, "rewards/format_reward/mean": 0.5546875, "rewards/format_reward/std": 0.2833639681339264, "rewards/mcq_exact_match_reward/mean": 0.3125, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "step": 19 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 998.0, "completions/max_terminated_length": 998.0, "completions/mean_length": 204.859375, "completions/mean_terminated_length": 204.859375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.05, "frac_reward_zero_std": 0.125, "grad_norm": 10.854552268981934, "learning_rate": 9.473646649103817e-07, "loss": 0.0, "num_tokens": 2087002.0, "reward": 0.39921876788139343, "reward_std": 0.35974013805389404, "rewards/format_reward/mean": 0.7109375, "rewards/format_reward/std": 0.2789533734321594, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "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": 905.0, "completions/max_terminated_length": 905.0, "completions/mean_length": 122.0, "completions/mean_terminated_length": 122.0, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0525, "frac_reward_zero_std": 0.0, "grad_norm": 22.049701690673828, "learning_rate": 9.397368756032444e-07, "loss": 0.0, "num_tokens": 2216714.0, "reward": 0.35859373211860657, "reward_std": 0.40040650963783264, "rewards/format_reward/mean": 0.6171875, "rewards/format_reward/std": 0.2313210517168045, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "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": 832.0, "completions/max_terminated_length": 832.0, "completions/mean_length": 121.1875, "completions/mean_terminated_length": 121.1875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.055, "frac_reward_zero_std": 0.25, "grad_norm": 22.305938720703125, "learning_rate": 9.316282404787869e-07, "loss": 0.0, "num_tokens": 2307542.0, "reward": 0.27656251192092896, "reward_std": 0.2901719808578491, "rewards/format_reward/mean": 0.578125, "rewards/format_reward/std": 0.23935678601264954, "rewards/mcq_exact_match_reward/mean": 0.21875, "rewards/mcq_exact_match_reward/std": 0.4166666865348816, "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": 1094.0, "completions/max_terminated_length": 1094.0, "completions/mean_length": 91.65625, "completions/mean_terminated_length": 91.65625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0575, "frac_reward_zero_std": 0.375, "grad_norm": 10.896363258361816, "learning_rate": 9.230476262104676e-07, "loss": 0.0, "num_tokens": 2393488.0, "reward": 0.23671874403953552, "reward_std": 0.1915045529603958, "rewards/format_reward/mean": 0.6484375, "rewards/format_reward/std": 0.2302463799715042, "rewards/mcq_exact_match_reward/mean": 0.171875, "rewards/mcq_exact_match_reward/std": 0.38025420904159546, "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": 748.0, "completions/max_terminated_length": 748.0, "completions/mean_length": 64.203125, "completions/mean_terminated_length": 64.203125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.06, "frac_reward_zero_std": 0.375, "grad_norm": 23.73764419555664, "learning_rate": 9.1400441557401e-07, "loss": 0.0, "num_tokens": 2482213.0, "reward": 0.38984376192092896, "reward_std": 0.16612984240055084, "rewards/format_reward/mean": 0.6171875, "rewards/format_reward/std": 0.21347814798355103, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "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": 737.0, "completions/max_terminated_length": 737.0, "completions/mean_length": 79.875, "completions/mean_terminated_length": 79.875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0625, "frac_reward_zero_std": 0.375, "grad_norm": 33.28952407836914, "learning_rate": 9.045084971874737e-07, "loss": -0.0, "num_tokens": 2542725.0, "reward": 0.48828125, "reward_std": 0.19378912448883057, "rewards/format_reward/mean": 0.6640625, "rewards/format_reward/std": 0.2366211861371994, "rewards/mcq_exact_match_reward/mean": 0.421875, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "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": 1185.0, "completions/max_terminated_length": 1185.0, "completions/mean_length": 71.765625, "completions/mean_terminated_length": 71.765625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.065, "frac_reward_zero_std": 0.5, "grad_norm": 20.798845291137695, "learning_rate": 8.945702546981968e-07, "loss": -0.0, "num_tokens": 2649342.0, "reward": 0.46328121423721313, "reward_std": 0.24288779497146606, "rewards/format_reward/mean": 0.5703125, "rewards/format_reward/std": 0.1751912236213684, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "step": 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": 582.0, "completions/max_terminated_length": 582.0, "completions/mean_length": 82.296875, "completions/mean_terminated_length": 82.296875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0675, "frac_reward_zero_std": 0.375, "grad_norm": 22.627351760864258, "learning_rate": 8.842005554284295e-07, "loss": 0.0, "num_tokens": 2745809.0, "reward": 0.5914062261581421, "reward_std": 0.2724077105522156, "rewards/format_reward/mean": 0.6015625, "rewards/format_reward/std": 0.22146137058734894, "rewards/mcq_exact_match_reward/mean": 0.53125, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "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": 434.0, "completions/max_terminated_length": 434.0, "completions/mean_length": 36.09375, "completions/mean_terminated_length": 36.09375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.07, "frac_reward_zero_std": 0.625, "grad_norm": 17.067331314086914, "learning_rate": 8.734107384920769e-07, "loss": 0.0, "num_tokens": 2841695.0, "reward": 0.35859376192092896, "reward_std": 0.13537800312042236, "rewards/format_reward/mean": 0.7734375, "rewards/format_reward/std": 0.2662152051925659, "rewards/mcq_exact_match_reward/mean": 0.28125, "rewards/mcq_exact_match_reward/std": 0.4531635046005249, "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": 672.0, "completions/max_terminated_length": 672.0, "completions/mean_length": 75.9375, "completions/mean_terminated_length": 75.9375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0725, "frac_reward_zero_std": 0.375, "grad_norm": 21.372982025146484, "learning_rate": 8.622126023955445e-07, "loss": -0.0, "num_tokens": 2918907.0, "reward": 0.28593748807907104, "reward_std": 0.15852639079093933, "rewards/format_reward/mean": 0.828125, "rewards/format_reward/std": 0.23935678601264954, "rewards/mcq_exact_match_reward/mean": 0.203125, "rewards/mcq_exact_match_reward/std": 0.40550529956817627, "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": 245.0, "completions/max_terminated_length": 245.0, "completions/mean_length": 26.046875, "completions/mean_terminated_length": 26.046875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.075, "frac_reward_zero_std": 0.0, "grad_norm": 29.563072204589844, "learning_rate": 8.506183921362442e-07, "loss": 0.0, "num_tokens": 2992974.0, "reward": 0.46796876192092896, "reward_std": 0.40812966227531433, "rewards/format_reward/mean": 0.7734375, "rewards/format_reward/std": 0.2807259261608124, "rewards/mcq_exact_match_reward/mean": 0.390625, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "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": 87.0, "completions/max_terminated_length": 87.0, "completions/mean_length": 10.90625, "completions/mean_terminated_length": 10.90625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0775, "frac_reward_zero_std": 0.5, "grad_norm": 26.153547286987305, "learning_rate": 8.386407858128706e-07, "loss": 0.0, "num_tokens": 3069568.0, "reward": 0.44218748807907104, "reward_std": 0.21409499645233154, "rewards/format_reward/mean": 0.671875, "rewards/format_reward/std": 0.23935678601264954, "rewards/mcq_exact_match_reward/mean": 0.375, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 433.0, "completions/max_terminated_length": 433.0, "completions/mean_length": 23.828125, "completions/mean_terminated_length": 23.828125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08, "frac_reward_zero_std": 0.25, "grad_norm": 33.8153076171875, "learning_rate": 8.262928807620843e-07, "loss": -0.0, "num_tokens": 3148957.0, "reward": 0.2796874940395355, "reward_std": 0.17765839397907257, "rewards/format_reward/mean": 0.765625, "rewards/format_reward/std": 0.2514837086200714, "rewards/mcq_exact_match_reward/mean": 0.203125, "rewards/mcq_exact_match_reward/std": 0.40550529956817627, "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": 51.0, "completions/max_terminated_length": 51.0, "completions/mean_length": 12.0625, "completions/mean_terminated_length": 12.0625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0825, "frac_reward_zero_std": 0.25, "grad_norm": 39.811927795410156, "learning_rate": 8.135881792367685e-07, "loss": 0.0, "num_tokens": 3229377.0, "reward": 0.3851562440395355, "reward_std": 0.21707361936569214, "rewards/format_reward/mean": 0.8828125, "rewards/format_reward/std": 0.21347814798355103, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "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": 526.0, "completions/max_terminated_length": 526.0, "completions/mean_length": 39.390625, "completions/mean_terminated_length": 39.390625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.085, "frac_reward_zero_std": 0.125, "grad_norm": 30.158384323120117, "learning_rate": 8.005405736415125e-07, "loss": 0.0, "num_tokens": 3312722.0, "reward": 0.5562499761581421, "reward_std": 0.31706470251083374, "rewards/format_reward/mean": 0.875, "rewards/format_reward/std": 0.2357022762298584, "rewards/mcq_exact_match_reward/mean": 0.46875, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "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": 374.0, "completions/max_terminated_length": 374.0, "completions/mean_length": 19.453125, "completions/mean_terminated_length": 19.453125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0875, "frac_reward_zero_std": 0.625, "grad_norm": 17.119461059570312, "learning_rate": 7.871643313414718e-07, "loss": 0.0, "num_tokens": 3386271.0, "reward": 0.5546875, "reward_std": 0.0903368592262268, "rewards/format_reward/mean": 0.859375, "rewards/format_reward/std": 0.22658175230026245, "rewards/mcq_exact_match_reward/mean": 0.46875, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "step": 35 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 291.0, "completions/mean_length": 49.28125, "completions/mean_terminated_length": 17.555557250976562, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.09, "frac_reward_zero_std": 0.25, "grad_norm": 22.900962829589844, "learning_rate": 7.734740790612136e-07, "loss": -0.0, "num_tokens": 3458017.0, "reward": 0.3500000238418579, "reward_std": 0.2655738592147827, "rewards/format_reward/mean": 0.84375, "rewards/format_reward/std": 0.279951810836792, "rewards/mcq_exact_match_reward/mean": 0.265625, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "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": 381.0, "completions/max_terminated_length": 381.0, "completions/mean_length": 24.359375, "completions/mean_terminated_length": 24.359375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0925, "frac_reward_zero_std": 0.25, "grad_norm": 16.959518432617188, "learning_rate": 7.594847868906076e-07, "loss": 0.0, "num_tokens": 3523720.0, "reward": 0.5007812976837158, "reward_std": 0.27784913778305054, "rewards/format_reward/mean": 0.9453125, "rewards/format_reward/std": 0.15728822350502014, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "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": 304.0, "completions/max_terminated_length": 304.0, "completions/mean_length": 17.046875, "completions/mean_terminated_length": 17.046875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.095, "frac_reward_zero_std": 0.125, "grad_norm": 42.04970932006836, "learning_rate": 7.452117519152541e-07, "loss": 0.0, "num_tokens": 3607035.0, "reward": 0.8289062976837158, "reward_std": 0.3522304892539978, "rewards/format_reward/mean": 0.9453125, "rewards/format_reward/std": 0.180765300989151, "rewards/mcq_exact_match_reward/mean": 0.734375, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "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": 234.0, "completions/max_terminated_length": 234.0, "completions/mean_length": 19.640625, "completions/mean_terminated_length": 19.640625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.0975, "frac_reward_zero_std": 0.25, "grad_norm": 19.459901809692383, "learning_rate": 7.306705814893439e-07, "loss": 0.0, "num_tokens": 3681708.0, "reward": 0.3921875059604645, "reward_std": 0.28690505027770996, "rewards/format_reward/mean": 0.953125, "rewards/format_reward/std": 0.19352105259895325, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "step": 39 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 814.0, "completions/max_terminated_length": 814.0, "completions/mean_length": 73.3125, "completions/mean_terminated_length": 73.3125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1, "frac_reward_zero_std": 0.375, "grad_norm": 13.290220260620117, "learning_rate": 7.158771761692464e-07, "loss": 0.0, "num_tokens": 3776136.0, "reward": 0.628125011920929, "reward_std": 0.27760595083236694, "rewards/format_reward/mean": 0.96875, "rewards/format_reward/std": 0.17536810040473938, "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": 473.0, "completions/max_terminated_length": 473.0, "completions/mean_length": 32.46875, "completions/mean_terminated_length": 32.46875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1025, "frac_reward_zero_std": 0.625, "grad_norm": 12.637351036071777, "learning_rate": 7.008477123264847e-07, "loss": 0.0, "num_tokens": 3844102.0, "reward": 0.535937488079071, "reward_std": 0.15604791045188904, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.4375, "rewards/mcq_exact_match_reward/std": 0.5, "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": 426.0, "completions/max_terminated_length": 426.0, "completions/mean_length": 44.765625, "completions/mean_terminated_length": 44.765625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.105, "frac_reward_zero_std": 0.375, "grad_norm": 9.738632202148438, "learning_rate": 6.855986244591103e-07, "loss": 0.0, "num_tokens": 3932831.0, "reward": 0.546875, "reward_std": 0.21087342500686646, "rewards/format_reward/mean": 0.9375, "rewards/format_reward/std": 0.24397502839565277, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "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": 654.0, "completions/max_terminated_length": 654.0, "completions/mean_length": 70.40625, "completions/mean_terminated_length": 70.40625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1075, "frac_reward_zero_std": 0.5, "grad_norm": 20.780141830444336, "learning_rate": 6.701465872208216e-07, "loss": 0.0, "num_tokens": 4044281.0, "reward": 0.6140625476837158, "reward_std": 0.21178999543190002, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.515625, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "step": 43 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 524.0, "completions/mean_length": 98.59375, "completions/mean_terminated_length": 67.65079498291016, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.11, "frac_reward_zero_std": 0.25, "grad_norm": 13.968363761901855, "learning_rate": 6.545084971874736e-07, "loss": 0.0, "num_tokens": 4143999.0, "reward": 0.39375001192092896, "reward_std": 0.2535737156867981, "rewards/format_reward/mean": 0.96875, "rewards/format_reward/std": 0.17536810040473938, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 377.0, "completions/mean_length": 79.71875, "completions/mean_terminated_length": 48.476192474365234, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1125, "frac_reward_zero_std": 0.125, "grad_norm": 16.15597915649414, "learning_rate": 6.387014543809223e-07, "loss": 0.0, "num_tokens": 4213101.0, "reward": 0.5484374761581421, "reward_std": 0.31799760460853577, "rewards/format_reward/mean": 0.953125, "rewards/format_reward/std": 0.21304203569889069, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 45 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 453.0, "completions/mean_length": 104.296875, "completions/mean_terminated_length": 41.596771240234375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.115, "frac_reward_zero_std": 0.125, "grad_norm": 34.93645477294922, "learning_rate": 6.227427435703995e-07, "loss": 0.0, "num_tokens": 4294880.0, "reward": 0.518750011920929, "reward_std": 0.33193695545196533, "rewards/format_reward/mean": 0.96875, "rewards/format_reward/std": 0.17536810040473938, "rewards/mcq_exact_match_reward/mean": 0.421875, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "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": 643.0, "completions/max_terminated_length": 643.0, "completions/mean_length": 74.84375, "completions/mean_terminated_length": 74.84375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1175, "frac_reward_zero_std": 0.125, "grad_norm": 22.327064514160156, "learning_rate": 6.066498153718734e-07, "loss": 0.0, "num_tokens": 4379998.0, "reward": 0.6109374761581421, "reward_std": 0.41196733713150024, "rewards/format_reward/mean": 0.953125, "rewards/format_reward/std": 0.21304203569889069, "rewards/mcq_exact_match_reward/mean": 0.515625, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "step": 47 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1137.0, "completions/mean_length": 139.265625, "completions/mean_terminated_length": 108.96826171875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.12, "frac_reward_zero_std": 0.25, "grad_norm": 12.447111129760742, "learning_rate": 5.90440267166055e-07, "loss": 0.0, "num_tokens": 4492815.0, "reward": 0.546093761920929, "reward_std": 0.3132563829421997, "rewards/format_reward/mean": 0.9296875, "rewards/format_reward/std": 0.23345555365085602, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 48 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1780.0, "completions/max_terminated_length": 1780.0, "completions/mean_length": 99.40625, "completions/mean_terminated_length": 99.40625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1225, "frac_reward_zero_std": 0.625, "grad_norm": 5.872361660003662, "learning_rate": 5.741318238559209e-07, "loss": 0.0, "num_tokens": 4575065.0, "reward": 0.21718750894069672, "reward_std": 0.14709635078907013, "rewards/format_reward/mean": 0.921875, "rewards/format_reward/std": 0.20351573824882507, "rewards/mcq_exact_match_reward/mean": 0.125, "rewards/mcq_exact_match_reward/std": 0.3333333432674408, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 857.0, "completions/mean_length": 150.921875, "completions/mean_terminated_length": 89.7258071899414, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.125, "frac_reward_zero_std": 0.375, "grad_norm": 9.439247131347656, "learning_rate": 5.577423184847931e-07, "loss": 0.0, "num_tokens": 4677188.0, "reward": 0.46875, "reward_std": 0.23794105648994446, "rewards/format_reward/mean": 0.9375, "rewards/format_reward/std": 0.24397502839565277, "rewards/mcq_exact_match_reward/mean": 0.375, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 171.0, "completions/mean_length": 54.828125, "completions/mean_terminated_length": 23.19047737121582, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1275, "frac_reward_zero_std": 0.5, "grad_norm": 12.116606712341309, "learning_rate": 5.412896727361662e-07, "loss": 0.0, "num_tokens": 4761753.0, "reward": 0.659375011920929, "reward_std": 0.20608238875865936, "rewards/format_reward/mean": 0.96875, "rewards/format_reward/std": 0.17536810040473938, "rewards/mcq_exact_match_reward/mean": 0.5625, "rewards/mcq_exact_match_reward/std": 0.5, "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": 0.046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 678.0, "completions/mean_length": 145.453125, "completions/mean_terminated_length": 51.8852424621582, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.13, "frac_reward_zero_std": 0.375, "grad_norm": 10.022530555725098, "learning_rate": 5.247918773366111e-07, "loss": 0.0, "num_tokens": 4849406.0, "reward": 0.75, "reward_std": 0.28560173511505127, "rewards/format_reward/mean": 0.9375, "rewards/format_reward/std": 0.24397502839565277, "rewards/mcq_exact_match_reward/mean": 0.65625, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "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": 1209.0, "completions/max_terminated_length": 1209.0, "completions/mean_length": 69.921875, "completions/mean_terminated_length": 69.921875, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.1325, "frac_reward_zero_std": 0.375, "grad_norm": 11.730982780456543, "learning_rate": 5.082669723831793e-07, "loss": 0.0, "num_tokens": 4947497.0, "reward": 0.5046875476837158, "reward_std": 0.23576273024082184, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "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": 0.0625, "completions/max_length": 2048.0, "completions/max_terminated_length": 218.0, "completions/mean_length": 155.90625, "completions/mean_terminated_length": 29.76666831970215, "completions/min_length": 17.0, "completions/min_terminated_length": 17.0, "epoch": 0.135, "frac_reward_zero_std": 0.125, "grad_norm": 13.682527542114258, "learning_rate": 4.917330276168208e-07, "loss": 0.0, "num_tokens": 5038387.0, "reward": 0.7171875238418579, "reward_std": 0.3899826407432556, "rewards/format_reward/mean": 0.921875, "rewards/format_reward/std": 0.27048972249031067, "rewards/mcq_exact_match_reward/mean": 0.625, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 448.0, "completions/mean_length": 97.96875, "completions/mean_terminated_length": 35.06451416015625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1375, "frac_reward_zero_std": 0.625, "grad_norm": 5.281121253967285, "learning_rate": 4.752081226633888e-07, "loss": 0.0, "num_tokens": 5125001.0, "reward": 0.46484375, "reward_std": 0.15240783989429474, "rewards/format_reward/mean": 0.8984375, "rewards/format_reward/std": 0.25479042530059814, "rewards/mcq_exact_match_reward/mean": 0.375, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 0.0625, "completions/max_length": 2048.0, "completions/max_terminated_length": 488.0, "completions/mean_length": 154.375, "completions/mean_terminated_length": 28.13333511352539, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.14, "frac_reward_zero_std": 0.5, "grad_norm": 29.663118362426758, "learning_rate": 4.5871032726383385e-07, "loss": 0.0, "num_tokens": 5235017.0, "reward": 0.38984376192092896, "reward_std": 0.20349781215190887, "rewards/format_reward/mean": 0.9296875, "rewards/format_reward/std": 0.2498759627342224, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "step": 56 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 206.0, "completions/mean_length": 59.65625, "completions/mean_terminated_length": 28.095239639282227, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1425, "frac_reward_zero_std": 0.25, "grad_norm": 11.722626686096191, "learning_rate": 4.4225768151520694e-07, "loss": -0.0, "num_tokens": 5333579.0, "reward": 0.6273437738418579, "reward_std": 0.33814358711242676, "rewards/format_reward/mean": 0.9609375, "rewards/format_reward/std": 0.18483558297157288, "rewards/mcq_exact_match_reward/mean": 0.53125, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "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": 0.09375, "completions/max_length": 2048.0, "completions/max_terminated_length": 290.0, "completions/mean_length": 219.546875, "completions/mean_terminated_length": 30.39655113220215, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.145, "frac_reward_zero_std": 0.375, "grad_norm": 18.92375946044922, "learning_rate": 4.258681761440789e-07, "loss": 0.0, "num_tokens": 5451374.0, "reward": 0.2562500238418579, "reward_std": 0.14224836230278015, "rewards/format_reward/mean": 0.84375, "rewards/format_reward/std": 0.33184191584587097, "rewards/mcq_exact_match_reward/mean": 0.171875, "rewards/mcq_exact_match_reward/std": 0.38025420904159546, "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": 0.0625, "completions/max_length": 2048.0, "completions/max_terminated_length": 370.0, "completions/mean_length": 150.59375, "completions/mean_terminated_length": 24.100000381469727, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.1475, "frac_reward_zero_std": 0.375, "grad_norm": 19.904739379882812, "learning_rate": 4.095597328339452e-07, "loss": 0.0, "num_tokens": 5549340.0, "reward": 0.515625, "reward_std": 0.1971278041601181, "rewards/format_reward/mean": 0.9375, "rewards/format_reward/std": 0.24397502839565277, "rewards/mcq_exact_match_reward/mean": 0.421875, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "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": 0.0625, "completions/max_length": 2048.0, "completions/max_terminated_length": 19.0, "completions/mean_length": 143.609375, "completions/mean_terminated_length": 16.650001525878906, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.15, "frac_reward_zero_std": 0.5, "grad_norm": 17.63539695739746, "learning_rate": 3.9335018462812664e-07, "loss": 0.0, "num_tokens": 5655467.0, "reward": 0.542187511920929, "reward_std": 0.1321321278810501, "rewards/format_reward/mean": 0.890625, "rewards/format_reward/std": 0.2741328477859497, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 60 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 20.0, "completions/mean_length": 176.125, "completions/mean_terminated_length": 17.491525650024414, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1525, "frac_reward_zero_std": 0.5, "grad_norm": 13.483983039855957, "learning_rate": 3.772572564296004e-07, "loss": 0.0, "num_tokens": 5734291.0, "reward": 0.26250001788139343, "reward_std": 0.16397076845169067, "rewards/format_reward/mean": 0.90625, "rewards/format_reward/std": 0.29378482699394226, "rewards/mcq_exact_match_reward/mean": 0.171875, "rewards/mcq_exact_match_reward/std": 0.38025420904159546, "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": 49.0, "completions/max_terminated_length": 49.0, "completions/mean_length": 18.40625, "completions/mean_terminated_length": 18.40625, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.155, "frac_reward_zero_std": 0.875, "grad_norm": 4.588263988494873, "learning_rate": 3.612985456190778e-07, "loss": 0.0, "num_tokens": 5836829.0, "reward": 0.7562500238418579, "reward_std": 0.0578637570142746, "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": 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": 191.0, "completions/max_terminated_length": 191.0, "completions/mean_length": 22.203125, "completions/mean_terminated_length": 22.203125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1575, "frac_reward_zero_std": 0.75, "grad_norm": 11.367886543273926, "learning_rate": 3.454915028125263e-07, "loss": 0.0, "num_tokens": 5927042.0, "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": 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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 204.0, "completions/mean_length": 52.0, "completions/mean_terminated_length": 20.317461013793945, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.16, "frac_reward_zero_std": 0.5, "grad_norm": 17.91376304626465, "learning_rate": 3.2985341277917846e-07, "loss": 0.0, "num_tokens": 6016122.0, "reward": 0.6296875476837158, "reward_std": 0.24654385447502136, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.53125, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "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": 416.0, "completions/max_terminated_length": 416.0, "completions/mean_length": 32.0, "completions/mean_terminated_length": 32.0, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.1625, "frac_reward_zero_std": 0.5, "grad_norm": 7.180004119873047, "learning_rate": 3.1440137554088953e-07, "loss": 0.0, "num_tokens": 6101122.0, "reward": 0.39531251788139343, "reward_std": 0.13700193166732788, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "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": 19.0, "completions/max_terminated_length": 19.0, "completions/mean_length": 17.609375, "completions/mean_terminated_length": 17.609375, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.165, "frac_reward_zero_std": 0.5, "grad_norm": 13.667840957641602, "learning_rate": 2.9915228767351535e-07, "loss": 0.0, "num_tokens": 6167865.0, "reward": 0.6156250238418579, "reward_std": 0.19044627249240875, "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": 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": 19.0, "completions/max_terminated_length": 19.0, "completions/mean_length": 17.296875, "completions/mean_terminated_length": 17.296875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1675, "frac_reward_zero_std": 0.625, "grad_norm": 12.145076751708984, "learning_rate": 2.841228238307536e-07, "loss": 0.0, "num_tokens": 6263852.0, "reward": 0.5531250238418579, "reward_std": 0.15992169082164764, "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": 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": 19.0, "completions/max_terminated_length": 19.0, "completions/mean_length": 17.421875, "completions/mean_terminated_length": 17.421875, "completions/min_length": 16.0, "completions/min_terminated_length": 16.0, "epoch": 0.17, "frac_reward_zero_std": 0.875, "grad_norm": 11.613428115844727, "learning_rate": 2.6932941851065615e-07, "loss": 0.0, "num_tokens": 6359071.0, "reward": 0.7562500238418579, "reward_std": 0.0578637570142746, "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": 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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 486.0, "completions/mean_length": 56.5, "completions/mean_terminated_length": 24.888891220092773, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1725, "frac_reward_zero_std": 0.625, "grad_norm": 9.375327110290527, "learning_rate": 2.547882480847461e-07, "loss": 0.0, "num_tokens": 6449079.0, "reward": 0.675000011920929, "reward_std": 0.1414213627576828, "rewards/format_reward/mean": 0.96875, "rewards/format_reward/std": 0.17536810040473938, "rewards/mcq_exact_match_reward/mean": 0.578125, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "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": 18.0, "completions/max_terminated_length": 18.0, "completions/mean_length": 17.640625, "completions/mean_terminated_length": 17.640625, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.175, "frac_reward_zero_std": 0.75, "grad_norm": 8.627045631408691, "learning_rate": 2.4051521310939254e-07, "loss": 0.0, "num_tokens": 6522256.0, "reward": 0.8796875476837158, "reward_std": 0.11031685769557953, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.78125, "rewards/mcq_exact_match_reward/std": 0.4166666865348816, "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": 570.0, "completions/max_terminated_length": 570.0, "completions/mean_length": 25.40625, "completions/mean_terminated_length": 25.40625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1775, "frac_reward_zero_std": 0.75, "grad_norm": 11.447951316833496, "learning_rate": 2.2652592093878665e-07, "loss": 0.0, "num_tokens": 6611418.0, "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": 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": 23.0, "completions/max_terminated_length": 23.0, "completions/mean_length": 16.0625, "completions/mean_terminated_length": 16.0625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.18, "frac_reward_zero_std": 0.375, "grad_norm": 14.377066612243652, "learning_rate": 2.128356686585282e-07, "loss": 0.0, "num_tokens": 6675974.0, "reward": 0.8171875476837158, "reward_std": 0.23548468947410583, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.71875, "rewards/mcq_exact_match_reward/std": 0.4531635046005249, "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": 19.0, "completions/max_terminated_length": 19.0, "completions/mean_length": 17.84375, "completions/mean_terminated_length": 17.84375, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.1825, "frac_reward_zero_std": 0.625, "grad_norm": 6.553983211517334, "learning_rate": 1.9945942635848745e-07, "loss": 0.0, "num_tokens": 6767620.0, "reward": 0.5062500238418579, "reward_std": 0.17570313811302185, "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": 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": 596.0, "completions/max_terminated_length": 596.0, "completions/mean_length": 30.96875, "completions/mean_terminated_length": 30.96875, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.185, "frac_reward_zero_std": 0.75, "grad_norm": 5.052456855773926, "learning_rate": 1.8641182076323148e-07, "loss": -0.0, "num_tokens": 6841722.0, "reward": 0.6312500238418579, "reward_std": 0.0883883461356163, "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": 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": 23.0, "completions/max_terminated_length": 23.0, "completions/mean_length": 17.46875, "completions/mean_terminated_length": 17.46875, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.1875, "frac_reward_zero_std": 0.875, "grad_norm": 3.0893845558166504, "learning_rate": 1.7370711923791564e-07, "loss": -0.0, "num_tokens": 6920352.0, "reward": 0.6156250238418579, "reward_std": 0.04419417306780815, "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": 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": 558.0, "completions/max_terminated_length": 558.0, "completions/mean_length": 26.109375, "completions/mean_terminated_length": 26.109375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.19, "frac_reward_zero_std": 0.625, "grad_norm": 8.620270729064941, "learning_rate": 1.6135921418712955e-07, "loss": 0.0, "num_tokens": 7014735.0, "reward": 0.7250000238418579, "reward_std": 0.16675157845020294, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.625, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "step": 76 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 20.0, "completions/max_terminated_length": 20.0, "completions/mean_length": 17.765625, "completions/mean_terminated_length": 17.765625, "completions/min_length": 16.0, "completions/min_terminated_length": 16.0, "epoch": 0.1925, "frac_reward_zero_std": 0.625, "grad_norm": 11.2853364944458, "learning_rate": 1.493816078637557e-07, "loss": 0.0, "num_tokens": 7103072.0, "reward": 0.3031250238418579, "reward_std": 0.15992169082164764, "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": 77 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 695.0, "completions/max_terminated_length": 695.0, "completions/mean_length": 38.71875, "completions/mean_terminated_length": 38.71875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.195, "frac_reward_zero_std": 0.75, "grad_norm": 1.1937483549118042, "learning_rate": 1.3778739760445552e-07, "loss": 0.0, "num_tokens": 7204254.0, "reward": 0.48906251788139343, "reward_std": 0.04861358925700188, "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": 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": 19.0, "completions/max_terminated_length": 19.0, "completions/mean_length": 17.40625, "completions/mean_terminated_length": 17.40625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1975, "frac_reward_zero_std": 0.625, "grad_norm": 11.812748908996582, "learning_rate": 1.2658926150792322e-07, "loss": 0.0, "num_tokens": 7300648.0, "reward": 0.7406250238418579, "reward_std": 0.1530819982290268, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.640625, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "step": 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": 19.0, "completions/max_terminated_length": 19.0, "completions/mean_length": 15.90625, "completions/mean_terminated_length": 15.90625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2, "frac_reward_zero_std": 0.625, "grad_norm": 12.863357543945312, "learning_rate": 1.1579944457157059e-07, "loss": 0.0, "num_tokens": 7369634.0, "reward": 0.6156250238418579, "reward_std": 0.189372718334198, "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": 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": 20.0, "completions/max_terminated_length": 20.0, "completions/mean_length": 16.640625, "completions/mean_terminated_length": 16.640625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2025, "frac_reward_zero_std": 0.75, "grad_norm": 14.480189323425293, "learning_rate": 1.0542974530180327e-07, "loss": 0.0, "num_tokens": 7444587.0, "reward": 0.5218750238418579, "reward_std": 0.12255740165710449, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.421875, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "step": 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": 19.0, "completions/max_terminated_length": 19.0, "completions/mean_length": 16.578125, "completions/mean_terminated_length": 16.578125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.205, "frac_reward_zero_std": 0.625, "grad_norm": 16.999113082885742, "learning_rate": 9.549150281252632e-08, "loss": 0.0, "num_tokens": 7539944.0, "reward": 0.4710937738418579, "reward_std": 0.16591501235961914, "rewards/format_reward/mean": 0.9609375, "rewards/format_reward/std": 0.13524486124515533, "rewards/mcq_exact_match_reward/mean": 0.375, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 19.0, "completions/max_terminated_length": 19.0, "completions/mean_length": 17.21875, "completions/mean_terminated_length": 17.21875, "completions/min_length": 16.0, "completions/min_terminated_length": 16.0, "epoch": 0.2075, "frac_reward_zero_std": 0.5, "grad_norm": 12.124747276306152, "learning_rate": 8.599558442598998e-08, "loss": 0.0, "num_tokens": 7613302.0, "reward": 0.48906251788139343, "reward_std": 0.23940719664096832, "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": 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": 19.0, "completions/max_terminated_length": 19.0, "completions/mean_length": 16.65625, "completions/mean_terminated_length": 16.65625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.21, "frac_reward_zero_std": 0.75, "grad_norm": 8.188920974731445, "learning_rate": 7.695237378953224e-08, "loss": -0.0, "num_tokens": 7695208.0, "reward": 0.6000000238418579, "reward_std": 0.0883883461356163, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.5, "rewards/mcq_exact_match_reward/std": 0.5039526224136353, "step": 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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 234.0, "completions/mean_length": 56.40625, "completions/mean_terminated_length": 24.793651580810547, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2125, "frac_reward_zero_std": 0.5, "grad_norm": 9.312822341918945, "learning_rate": 6.837175952121304e-08, "loss": 0.0, "num_tokens": 7765610.0, "reward": 0.5515625476837158, "reward_std": 0.19486570358276367, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 85 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 19.0, "completions/max_terminated_length": 19.0, "completions/mean_length": 17.125, "completions/mean_terminated_length": 17.125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.215, "frac_reward_zero_std": 0.5, "grad_norm": 16.462488174438477, "learning_rate": 6.026312439675551e-08, "loss": 0.0, "num_tokens": 7841922.0, "reward": 0.7093750238418579, "reward_std": 0.2198973000049591, "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": 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": 28.0, "completions/max_terminated_length": 28.0, "completions/mean_length": 17.3125, "completions/mean_terminated_length": 17.3125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2175, "frac_reward_zero_std": 0.5, "grad_norm": 18.94294548034668, "learning_rate": 5.263533508961826e-08, "loss": 0.0, "num_tokens": 7912254.0, "reward": 0.6453125476837158, "reward_std": 0.17329266667366028, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.546875, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 256.0, "completions/mean_length": 51.953125, "completions/mean_terminated_length": 20.26984214782715, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.22, "frac_reward_zero_std": 0.5, "grad_norm": 15.992481231689453, "learning_rate": 4.549673247541874e-08, "loss": 0.0, "num_tokens": 7984515.0, "reward": 0.721875011920929, "reward_std": 0.15509094297885895, "rewards/format_reward/mean": 0.96875, "rewards/format_reward/std": 0.17536810040473938, "rewards/mcq_exact_match_reward/mean": 0.625, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 656.0, "completions/max_terminated_length": 656.0, "completions/mean_length": 28.0625, "completions/mean_terminated_length": 28.0625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2225, "frac_reward_zero_std": 0.875, "grad_norm": 2.776789903640747, "learning_rate": 3.8855122511307626e-08, "loss": 0.0, "num_tokens": 8047727.0, "reward": 0.5218750238418579, "reward_std": 0.0646936446428299, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.421875, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "step": 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": 20.0, "completions/max_terminated_length": 20.0, "completions/mean_length": 16.703125, "completions/mean_terminated_length": 16.703125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.225, "frac_reward_zero_std": 0.625, "grad_norm": 14.825093269348145, "learning_rate": 3.271776770026963e-08, "loss": 0.0, "num_tokens": 8138692.0, "reward": 0.2250000238418579, "reward_std": 0.1462520956993103, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.125, "rewards/mcq_exact_match_reward/std": 0.3333333432674408, "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": 20.0, "completions/max_terminated_length": 20.0, "completions/mean_length": 17.8125, "completions/mean_terminated_length": 17.8125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2275, "frac_reward_zero_std": 0.75, "grad_norm": 10.433076858520508, "learning_rate": 2.7091379149682682e-08, "loss": 0.0, "num_tokens": 8232864.0, "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": 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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 19.0, "completions/mean_length": 49.484375, "completions/mean_terminated_length": 17.761905670166016, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.23, "frac_reward_zero_std": 0.75, "grad_norm": 7.038073539733887, "learning_rate": 2.1982109232821176e-08, "loss": 0.0, "num_tokens": 8350719.0, "reward": 0.6296875476837158, "reward_std": 0.11330723762512207, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.53125, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "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": 146.0, "completions/max_terminated_length": 146.0, "completions/mean_length": 19.796875, "completions/mean_terminated_length": 19.796875, "completions/min_length": 16.0, "completions/min_terminated_length": 16.0, "epoch": 0.2325, "frac_reward_zero_std": 0.75, "grad_norm": 8.651636123657227, "learning_rate": 1.7395544861325718e-08, "loss": 0.0, "num_tokens": 8423642.0, "reward": 0.7718750238418579, "reward_std": 0.11100947856903076, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.671875, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "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": 19.0, "completions/max_terminated_length": 19.0, "completions/mean_length": 16.546875, "completions/mean_terminated_length": 16.546875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.235, "frac_reward_zero_std": 0.75, "grad_norm": 7.556106090545654, "learning_rate": 1.3336701375997127e-08, "loss": -0.0, "num_tokens": 8516605.0, "reward": 0.6312500238418579, "reward_std": 0.0883883461356163, "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": 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": 18.0, "completions/max_terminated_length": 18.0, "completions/mean_length": 17.390625, "completions/mean_terminated_length": 17.390625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2375, "frac_reward_zero_std": 0.625, "grad_norm": 14.45374870300293, "learning_rate": 9.810017062595321e-09, "loss": 0.0, "num_tokens": 8607510.0, "reward": 0.33281251788139343, "reward_std": 0.15534965693950653, "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": 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": 19.0, "completions/max_terminated_length": 19.0, "completions/mean_length": 17.71875, "completions/mean_terminated_length": 17.71875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.24, "frac_reward_zero_std": 0.75, "grad_norm": 6.179401874542236, "learning_rate": 6.819348298638839e-09, "loss": 0.0, "num_tokens": 8698164.0, "reward": 0.5531250238418579, "reward_std": 0.11100947856903076, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 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": 19.0, "completions/max_terminated_length": 19.0, "completions/mean_length": 17.21875, "completions/mean_terminated_length": 17.21875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2425, "frac_reward_zero_std": 0.5, "grad_norm": 20.131906509399414, "learning_rate": 4.367965336512403e-09, "loss": 0.0, "num_tokens": 8795594.0, "reward": 0.4906250238418579, "reward_std": 0.260896235704422, "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": 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": 22.0, "completions/max_terminated_length": 22.0, "completions/mean_length": 17.953125, "completions/mean_terminated_length": 17.953125, "completions/min_length": 16.0, "completions/min_terminated_length": 16.0, "epoch": 0.245, "frac_reward_zero_std": 0.5, "grad_norm": 10.667953491210938, "learning_rate": 2.458548727494292e-09, "loss": 0.0, "num_tokens": 8867999.0, "reward": 0.4125000238418579, "reward_std": 0.2041158676147461, "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": 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": 19.0, "completions/max_terminated_length": 19.0, "completions/mean_length": 16.546875, "completions/mean_terminated_length": 16.546875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2475, "frac_reward_zero_std": 0.625, "grad_norm": 13.83715534210205, "learning_rate": 1.0931863906127325e-09, "loss": 0.0, "num_tokens": 8936210.0, "reward": 0.6156250238418579, "reward_std": 0.15992169082164764, "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": 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": 78.0, "completions/max_terminated_length": 78.0, "completions/mean_length": 18.65625, "completions/mean_terminated_length": 18.65625, "completions/min_length": 16.0, "completions/min_terminated_length": 16.0, "epoch": 0.25, "frac_reward_zero_std": 0.625, "grad_norm": 10.48843002319336, "learning_rate": 2.733713295369755e-10, "loss": 0.0, "num_tokens": 9020348.0, "reward": 0.8656250238418579, "reward_std": 0.17358146607875824, "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": 100 } ], "logging_steps": 1, "max_steps": 100, "num_input_tokens_seen": 9020348, "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 }