{ "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": 1441.0, "completions/mean_length": 533.734375, "completions/mean_terminated_length": 509.6984558105469, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.0025, "frac_reward_zero_std": 0.0, "grad_norm": 9.574075698852539, "learning_rate": 0.0, "loss": 0.0, "num_tokens": 137999.0, "reward": 0.36250001192092896, "reward_std": 0.33260637521743774, "rewards/format_reward/mean": 0.34375, "rewards/format_reward/std": 0.25, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1433.0, "completions/mean_length": 514.625, "completions/mean_terminated_length": 490.2857360839844, "completions/min_length": 11.0, "completions/min_terminated_length": 11.0, "epoch": 0.005, "frac_reward_zero_std": 0.0, "grad_norm": 2.057363748550415, "learning_rate": 2e-07, "loss": 0.0, "num_tokens": 239247.0, "reward": 0.22187499701976776, "reward_std": 0.24802732467651367, "rewards/format_reward/mean": 0.34375, "rewards/format_reward/std": 0.3196600377559662, "rewards/mcq_exact_match_reward/mean": 0.1875, "rewards/mcq_exact_match_reward/std": 0.39339789748191833, "step": 2 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 685.953125, "completions/mean_terminated_length": 642.01611328125, "completions/min_length": 11.0, "completions/min_terminated_length": 11.0, "epoch": 0.0075, "frac_reward_zero_std": 0.25, "grad_norm": 6.393513202667236, "learning_rate": 4e-07, "loss": 0.0, "num_tokens": 368564.0, "reward": 0.11015623807907104, "reward_std": 0.20267972350120544, "rewards/format_reward/mean": 0.3203125, "rewards/format_reward/std": 0.24180518090724945, "rewards/mcq_exact_match_reward/mean": 0.078125, "rewards/mcq_exact_match_reward/std": 0.27048972249031067, "step": 3 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1870.0, "completions/mean_length": 566.828125, "completions/mean_terminated_length": 519.0484008789062, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.01, "frac_reward_zero_std": 0.0, "grad_norm": 5.784980773925781, "learning_rate": 6e-07, "loss": -0.0, "num_tokens": 499017.0, "reward": 0.28828123211860657, "reward_std": 0.3855046331882477, "rewards/format_reward/mean": 0.3828125, "rewards/format_reward/std": 0.21347814798355103, "rewards/mcq_exact_match_reward/mean": 0.25, "rewards/mcq_exact_match_reward/std": 0.4364357888698578, "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": 1655.0, "completions/mean_length": 393.703125, "completions/mean_terminated_length": 340.3387145996094, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0125, "frac_reward_zero_std": 0.125, "grad_norm": 7.174067497253418, "learning_rate": 8e-07, "loss": -0.0, "num_tokens": 624182.0, "reward": 0.17500001192092896, "reward_std": 0.20127461850643158, "rewards/format_reward/mean": 0.34375, "rewards/format_reward/std": 0.279951810836792, "rewards/mcq_exact_match_reward/mean": 0.140625, "rewards/mcq_exact_match_reward/std": 0.3503824472427368, "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": 1544.0, "completions/max_terminated_length": 1544.0, "completions/mean_length": 549.734375, "completions/mean_terminated_length": 549.734375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.015, "frac_reward_zero_std": 0.0, "grad_norm": 18.369905471801758, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 769069.0, "reward": 0.15781249105930328, "reward_std": 0.237949937582016, "rewards/format_reward/mean": 0.328125, "rewards/format_reward/std": 0.23935678601264954, "rewards/mcq_exact_match_reward/mean": 0.125, "rewards/mcq_exact_match_reward/std": 0.3333333432674408, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1866.0, "completions/mean_length": 628.546875, "completions/mean_terminated_length": 582.758056640625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0175, "frac_reward_zero_std": 0.0, "grad_norm": 2.5032742023468018, "learning_rate": 9.99726628670463e-07, "loss": 0.0, "num_tokens": 919208.0, "reward": 0.23671874403953552, "reward_std": 0.3317972421646118, "rewards/format_reward/mean": 0.3359375, "rewards/format_reward/std": 0.2824873626232147, "rewards/mcq_exact_match_reward/mean": 0.203125, "rewards/mcq_exact_match_reward/std": 0.40550529956817627, "step": 7 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1681.0, "completions/max_terminated_length": 1681.0, "completions/mean_length": 513.578125, "completions/mean_terminated_length": 513.578125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.02, "frac_reward_zero_std": 0.0, "grad_norm": 8.070008277893066, "learning_rate": 9.989068136093872e-07, "loss": -0.0, "num_tokens": 1041813.0, "reward": 0.33906248211860657, "reward_std": 0.4019244909286499, "rewards/format_reward/mean": 0.421875, "rewards/format_reward/std": 0.27048972249031067, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "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": 0.046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1318.0, "completions/mean_length": 607.453125, "completions/mean_terminated_length": 536.6065063476562, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0225, "frac_reward_zero_std": 0.125, "grad_norm": 2.9918811321258545, "learning_rate": 9.975414512725056e-07, "loss": -0.0, "num_tokens": 1206514.0, "reward": 0.18984374403953552, "reward_std": 0.3112410306930542, "rewards/format_reward/mean": 0.3359375, "rewards/format_reward/std": 0.2366211861371994, "rewards/mcq_exact_match_reward/mean": 0.15625, "rewards/mcq_exact_match_reward/std": 0.36596253514289856, "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": 0.0625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2022.0, "completions/mean_length": 654.25, "completions/mean_terminated_length": 561.3333740234375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.025, "frac_reward_zero_std": 0.0, "grad_norm": 4.438917636871338, "learning_rate": 9.956320346634875e-07, "loss": -0.0, "num_tokens": 1345970.0, "reward": 0.3515625, "reward_std": 0.43372035026550293, "rewards/format_reward/mean": 0.390625, "rewards/format_reward/std": 0.2592533528804779, "rewards/mcq_exact_match_reward/mean": 0.3125, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "step": 10 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1661.0, "completions/max_terminated_length": 1661.0, "completions/mean_length": 385.953125, "completions/mean_terminated_length": 385.953125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0275, "frac_reward_zero_std": 0.0, "grad_norm": 8.773837089538574, "learning_rate": 9.931806517013612e-07, "loss": -0.0, "num_tokens": 1463111.0, "reward": 0.4703124761581421, "reward_std": 0.483210027217865, "rewards/format_reward/mean": 0.484375, "rewards/format_reward/std": 0.2514837086200714, "rewards/mcq_exact_match_reward/mean": 0.421875, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "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": 0.03125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1265.0, "completions/mean_length": 410.375, "completions/mean_terminated_length": 357.5483703613281, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.03, "frac_reward_zero_std": 0.25, "grad_norm": 6.2783918380737305, "learning_rate": 9.901899829374047e-07, "loss": -0.0, "num_tokens": 1570039.0, "reward": 0.23281247913837433, "reward_std": 0.2448979616165161, "rewards/format_reward/mean": 0.453125, "rewards/format_reward/std": 0.23091863095760345, "rewards/mcq_exact_match_reward/mean": 0.1875, "rewards/mcq_exact_match_reward/std": 0.39339789748191833, "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": 1565.0, "completions/mean_length": 465.4375, "completions/mean_terminated_length": 440.3174743652344, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0325, "frac_reward_zero_std": 0.0, "grad_norm": 17.623409271240234, "learning_rate": 9.866632986240029e-07, "loss": -0.0, "num_tokens": 1686171.0, "reward": 0.3609374761581421, "reward_std": 0.4090258777141571, "rewards/format_reward/mean": 0.484375, "rewards/format_reward/std": 0.23517554998397827, "rewards/mcq_exact_match_reward/mean": 0.3125, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "step": 13 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1421.0, "completions/mean_length": 321.546875, "completions/mean_terminated_length": 294.14288330078125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.035, "frac_reward_zero_std": 0.125, "grad_norm": 12.31415843963623, "learning_rate": 9.826044551386742e-07, "loss": 0.0, "num_tokens": 1798950.0, "reward": 0.34453123807907104, "reward_std": 0.3583349585533142, "rewards/format_reward/mean": 0.4765625, "rewards/format_reward/std": 0.25865477323532104, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "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": 1573.0, "completions/max_terminated_length": 1573.0, "completions/mean_length": 396.40625, "completions/mean_terminated_length": 396.40625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0375, "frac_reward_zero_std": 0.0, "grad_norm": 19.253488540649414, "learning_rate": 9.780178907671788e-07, "loss": -0.0, "num_tokens": 1916400.0, "reward": 0.359375, "reward_std": 0.2717418670654297, "rewards/format_reward/mean": 0.46875, "rewards/format_reward/std": 0.33184191584587097, "rewards/mcq_exact_match_reward/mean": 0.3125, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "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": 1225.0, "completions/max_terminated_length": 1225.0, "completions/mean_length": 230.765625, "completions/mean_terminated_length": 230.765625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.04, "frac_reward_zero_std": 0.5, "grad_norm": 7.886924743652344, "learning_rate": 9.729086208503173e-07, "loss": -0.0, "num_tokens": 2033721.0, "reward": 0.6734374761581421, "reward_std": 0.14883117377758026, "rewards/format_reward/mean": 0.484375, "rewards/format_reward/std": 0.08768405020236969, "rewards/mcq_exact_match_reward/mean": 0.625, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 1281.0, "completions/max_terminated_length": 1281.0, "completions/mean_length": 342.96875, "completions/mean_terminated_length": 342.96875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0425, "frac_reward_zero_std": 0.0, "grad_norm": 14.499305725097656, "learning_rate": 9.672822322997304e-07, "loss": -0.0, "num_tokens": 2119255.0, "reward": 0.39374998211860657, "reward_std": 0.35225731134414673, "rewards/format_reward/mean": 0.5, "rewards/format_reward/std": 0.2519763112068176, "rewards/mcq_exact_match_reward/mean": 0.34375, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "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": 1413.0, "completions/max_terminated_length": 1413.0, "completions/mean_length": 127.03125, "completions/mean_terminated_length": 127.03125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.045, "frac_reward_zero_std": 0.375, "grad_norm": 16.520790100097656, "learning_rate": 9.611448774886923e-07, "loss": -0.0, "num_tokens": 2218393.0, "reward": 0.22890624403953552, "reward_std": 0.2134948968887329, "rewards/format_reward/mean": 0.5703125, "rewards/format_reward/std": 0.19654127955436707, "rewards/mcq_exact_match_reward/mean": 0.171875, "rewards/mcq_exact_match_reward/std": 0.38025420904159546, "step": 18 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1190.0, "completions/max_terminated_length": 1190.0, "completions/mean_length": 104.8125, "completions/mean_terminated_length": 104.8125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0475, "frac_reward_zero_std": 0.375, "grad_norm": 18.01238250732422, "learning_rate": 9.545032675245813e-07, "loss": 0.0, "num_tokens": 2340869.0, "reward": 0.5492187142372131, "reward_std": 0.25567346811294556, "rewards/format_reward/mean": 0.4921875, "rewards/format_reward/std": 0.1406387835741043, "rewards/mcq_exact_match_reward/mean": 0.5, "rewards/mcq_exact_match_reward/std": 0.5039526224136353, "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": 863.0, "completions/max_terminated_length": 863.0, "completions/mean_length": 81.359375, "completions/mean_terminated_length": 81.359375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.05, "frac_reward_zero_std": 0.25, "grad_norm": 25.483932495117188, "learning_rate": 9.473646649103817e-07, "loss": -0.0, "num_tokens": 2443764.0, "reward": 0.36171871423721313, "reward_std": 0.3018960654735565, "rewards/format_reward/mean": 0.4921875, "rewards/format_reward/std": 0.1406387835741043, "rewards/mcq_exact_match_reward/mean": 0.3125, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "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": 643.0, "completions/max_terminated_length": 643.0, "completions/mean_length": 81.828125, "completions/mean_terminated_length": 81.828125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0525, "frac_reward_zero_std": 0.625, "grad_norm": 6.843085765838623, "learning_rate": 9.397368756032444e-07, "loss": -0.0, "num_tokens": 2556569.0, "reward": 0.520312488079071, "reward_std": 0.1663079857826233, "rewards/format_reward/mean": 0.515625, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.46875, "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": 284.0, "completions/max_terminated_length": 284.0, "completions/mean_length": 29.0, "completions/mean_terminated_length": 29.0, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.055, "frac_reward_zero_std": 0.375, "grad_norm": 17.255512237548828, "learning_rate": 9.316282404787869e-07, "loss": 0.0, "num_tokens": 2650289.0, "reward": 0.2828124761581421, "reward_std": 0.1596856713294983, "rewards/format_reward/mean": 0.484375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "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": 711.0, "completions/max_terminated_length": 711.0, "completions/mean_length": 112.296875, "completions/mean_terminated_length": 112.296875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0575, "frac_reward_zero_std": 0.375, "grad_norm": 15.66704273223877, "learning_rate": 9.230476262104676e-07, "loss": -0.0, "num_tokens": 2738812.0, "reward": 0.55078125, "reward_std": 0.2940145432949066, "rewards/format_reward/mean": 0.5078125, "rewards/format_reward/std": 0.1406387835741043, "rewards/mcq_exact_match_reward/mean": 0.5, "rewards/mcq_exact_match_reward/std": 0.5039526224136353, "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": 601.0, "completions/max_terminated_length": 601.0, "completions/mean_length": 33.671875, "completions/mean_terminated_length": 33.671875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.06, "frac_reward_zero_std": 0.875, "grad_norm": 5.158745765686035, "learning_rate": 9.1400441557401e-07, "loss": 0.0, "num_tokens": 2839007.0, "reward": 0.5187499523162842, "reward_std": 0.0578637532889843, "rewards/format_reward/mean": 0.5, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.46875, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "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": 708.0, "completions/max_terminated_length": 708.0, "completions/mean_length": 143.5625, "completions/mean_terminated_length": 143.5625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0625, "frac_reward_zero_std": 0.5, "grad_norm": 9.724034309387207, "learning_rate": 9.045084971874737e-07, "loss": -0.0, "num_tokens": 2925923.0, "reward": 0.36562496423721313, "reward_std": 0.11465182900428772, "rewards/format_reward/mean": 0.53125, "rewards/format_reward/std": 0.12198751419782639, "rewards/mcq_exact_match_reward/mean": 0.3125, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "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": 1025.0, "completions/max_terminated_length": 1025.0, "completions/mean_length": 98.84375, "completions/mean_terminated_length": 98.84375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.065, "frac_reward_zero_std": 0.25, "grad_norm": 25.84486198425293, "learning_rate": 8.945702546981968e-07, "loss": -0.0, "num_tokens": 3022441.0, "reward": 0.45624998211860657, "reward_std": 0.3131307363510132, "rewards/format_reward/mean": 0.5, "rewards/format_reward/std": 0.17817416787147522, "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": 59.0, "completions/max_terminated_length": 59.0, "completions/mean_length": 7.578125, "completions/mean_terminated_length": 7.578125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0675, "frac_reward_zero_std": 0.375, "grad_norm": 24.051807403564453, "learning_rate": 8.842005554284295e-07, "loss": -0.0, "num_tokens": 3109870.0, "reward": 0.4976562261581421, "reward_std": 0.2336549162864685, "rewards/format_reward/mean": 0.4453125, "rewards/format_reward/std": 0.15728822350502014, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "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": 433.0, "completions/max_terminated_length": 433.0, "completions/mean_length": 19.1875, "completions/mean_terminated_length": 19.1875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.07, "frac_reward_zero_std": 0.25, "grad_norm": 42.07900619506836, "learning_rate": 8.734107384920769e-07, "loss": 0.0, "num_tokens": 3204578.0, "reward": 0.5656249523162842, "reward_std": 0.2993340790271759, "rewards/format_reward/mean": 0.5, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.515625, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "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": 6.0, "completions/max_terminated_length": 6.0, "completions/mean_length": 6.0, "completions/mean_terminated_length": 6.0, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0725, "frac_reward_zero_std": 0.625, "grad_norm": 22.37701988220215, "learning_rate": 8.622126023955445e-07, "loss": 0.0, "num_tokens": 3283794.0, "reward": 0.5968749523162842, "reward_std": 0.1530819833278656, "rewards/format_reward/mean": 0.5, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.546875, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "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": 749.0, "completions/max_terminated_length": 749.0, "completions/mean_length": 117.28125, "completions/mean_terminated_length": 117.28125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.075, "frac_reward_zero_std": 0.125, "grad_norm": 35.932193756103516, "learning_rate": 8.506183921362442e-07, "loss": -0.0, "num_tokens": 3362228.0, "reward": 0.4468749761581421, "reward_std": 0.334439218044281, "rewards/format_reward/mean": 0.5625, "rewards/format_reward/std": 0.18898223340511322, "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": 586.0, "completions/max_terminated_length": 586.0, "completions/mean_length": 17.84375, "completions/mean_terminated_length": 17.84375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0775, "frac_reward_zero_std": 0.625, "grad_norm": 14.128104209899902, "learning_rate": 8.386407858128706e-07, "loss": 0.0, "num_tokens": 3444178.0, "reward": 0.6921874284744263, "reward_std": 0.10495112091302872, "rewards/format_reward/mean": 0.515625, "rewards/format_reward/std": 0.08768405020236969, "rewards/mcq_exact_match_reward/mean": 0.640625, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "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": 596.0, "completions/max_terminated_length": 596.0, "completions/mean_length": 50.65625, "completions/mean_terminated_length": 50.65625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.08, "frac_reward_zero_std": 0.375, "grad_norm": 26.173019409179688, "learning_rate": 8.262928807620843e-07, "loss": -0.0, "num_tokens": 3527332.0, "reward": 0.4562499523162842, "reward_std": 0.28247907757759094, "rewards/format_reward/mean": 0.5, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "step": 32 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 6.0, "completions/max_terminated_length": 6.0, "completions/mean_length": 6.0, "completions/mean_terminated_length": 6.0, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0825, "frac_reward_zero_std": 0.625, "grad_norm": 19.123197555541992, "learning_rate": 8.135881792367685e-07, "loss": -0.0, "num_tokens": 3623060.0, "reward": 0.5187499523162842, "reward_std": 0.1872510462999344, "rewards/format_reward/mean": 0.5, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.46875, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "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": 842.0, "completions/max_terminated_length": 842.0, "completions/mean_length": 81.34375, "completions/mean_terminated_length": 81.34375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.085, "frac_reward_zero_std": 0.375, "grad_norm": 25.286540985107422, "learning_rate": 8.005405736415125e-07, "loss": 0.0, "num_tokens": 3731522.0, "reward": 0.41328126192092896, "reward_std": 0.24796469509601593, "rewards/format_reward/mean": 0.5390625, "rewards/format_reward/std": 0.13524486124515533, "rewards/mcq_exact_match_reward/mean": 0.359375, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "step": 34 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 589.0, "completions/max_terminated_length": 589.0, "completions/mean_length": 27.9375, "completions/mean_terminated_length": 27.9375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0875, "frac_reward_zero_std": 0.625, "grad_norm": 15.118162155151367, "learning_rate": 7.871643313414718e-07, "loss": -0.0, "num_tokens": 3817430.0, "reward": 0.39374998211860657, "reward_std": 0.1462520956993103, "rewards/format_reward/mean": 0.5, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.34375, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "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": 6.0, "completions/max_terminated_length": 6.0, "completions/mean_length": 6.0, "completions/mean_terminated_length": 6.0, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.09, "frac_reward_zero_std": 0.625, "grad_norm": 21.149654388427734, "learning_rate": 7.734740790612136e-07, "loss": -0.0, "num_tokens": 3937078.0, "reward": 0.4406249523162842, "reward_std": 0.1530819982290268, "rewards/format_reward/mean": 0.5, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.390625, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "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": 254.0, "completions/max_terminated_length": 254.0, "completions/mean_length": 15.515625, "completions/mean_terminated_length": 15.515625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0925, "frac_reward_zero_std": 0.75, "grad_norm": 12.659499168395996, "learning_rate": 7.594847868906076e-07, "loss": 0.0, "num_tokens": 4022103.0, "reward": 0.45546871423721313, "reward_std": 0.1255451738834381, "rewards/format_reward/mean": 0.4921875, "rewards/format_reward/std": 0.0625, "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": 706.0, "completions/max_terminated_length": 706.0, "completions/mean_length": 156.703125, "completions/mean_terminated_length": 156.703125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.095, "frac_reward_zero_std": 0.375, "grad_norm": 24.284320831298828, "learning_rate": 7.452117519152541e-07, "loss": -0.0, "num_tokens": 4111860.0, "reward": 0.6460937261581421, "reward_std": 0.30066248774528503, "rewards/format_reward/mean": 0.5234375, "rewards/format_reward/std": 0.20758795738220215, "rewards/mcq_exact_match_reward/mean": 0.59375, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "step": 38 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 508.0, "completions/max_terminated_length": 508.0, "completions/mean_length": 51.203125, "completions/mean_terminated_length": 51.203125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.0975, "frac_reward_zero_std": 0.5, "grad_norm": 13.789477348327637, "learning_rate": 7.306705814893439e-07, "loss": 0.0, "num_tokens": 4230641.0, "reward": 0.21484375, "reward_std": 0.09593167901039124, "rewards/format_reward/mean": 0.5859375, "rewards/format_reward/std": 0.209963858127594, "rewards/mcq_exact_match_reward/mean": 0.15625, "rewards/mcq_exact_match_reward/std": 0.36596253514289856, "step": 39 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 643.0, "completions/max_terminated_length": 643.0, "completions/mean_length": 38.53125, "completions/mean_terminated_length": 38.53125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1, "frac_reward_zero_std": 0.875, "grad_norm": 4.174818992614746, "learning_rate": 7.158771761692464e-07, "loss": -0.0, "num_tokens": 4334243.0, "reward": 0.23984374105930328, "reward_std": 0.06772264838218689, "rewards/format_reward/mean": 0.5234375, "rewards/format_reward/std": 0.10652101784944534, "rewards/mcq_exact_match_reward/mean": 0.1875, "rewards/mcq_exact_match_reward/std": 0.39339789748191833, "step": 40 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 472.0, "completions/max_terminated_length": 472.0, "completions/mean_length": 15.53125, "completions/mean_terminated_length": 15.53125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1025, "frac_reward_zero_std": 0.75, "grad_norm": 19.140796661376953, "learning_rate": 7.008477123264847e-07, "loss": -0.0, "num_tokens": 4430149.0, "reward": 0.6499999761581421, "reward_std": 0.10888782143592834, "rewards/format_reward/mean": 0.5625, "rewards/format_reward/std": 0.1666666716337204, "rewards/mcq_exact_match_reward/mean": 0.59375, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "step": 41 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 783.0, "completions/max_terminated_length": 783.0, "completions/mean_length": 32.1875, "completions/mean_terminated_length": 32.1875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.105, "frac_reward_zero_std": 0.5, "grad_norm": 26.529802322387695, "learning_rate": 6.855986244591103e-07, "loss": 0.0, "num_tokens": 4510649.0, "reward": 0.6078125238418579, "reward_std": 0.17176641523838043, "rewards/format_reward/mean": 0.609375, "rewards/format_reward/std": 0.2083333432674408, "rewards/mcq_exact_match_reward/mean": 0.546875, "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": 402.0, "completions/max_terminated_length": 402.0, "completions/mean_length": 17.953125, "completions/mean_terminated_length": 17.953125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1075, "frac_reward_zero_std": 0.75, "grad_norm": 32.29198455810547, "learning_rate": 6.701465872208216e-07, "loss": 0.0, "num_tokens": 4623750.0, "reward": 0.4921874403953552, "reward_std": 0.06970848888158798, "rewards/format_reward/mean": 0.546875, "rewards/format_reward/std": 0.14689241349697113, "rewards/mcq_exact_match_reward/mean": 0.4375, "rewards/mcq_exact_match_reward/std": 0.5, "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": 603.0, "completions/max_terminated_length": 603.0, "completions/mean_length": 20.890625, "completions/mean_terminated_length": 20.890625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.11, "frac_reward_zero_std": 0.625, "grad_norm": 19.41641616821289, "learning_rate": 6.545084971874736e-07, "loss": 0.0, "num_tokens": 4717423.0, "reward": 0.26093751192092896, "reward_std": 0.14697228372097015, "rewards/format_reward/mean": 0.734375, "rewards/format_reward/std": 0.2514837086200714, "rewards/mcq_exact_match_reward/mean": 0.1875, "rewards/mcq_exact_match_reward/std": 0.39339789748191833, "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": 478.0, "completions/max_terminated_length": 478.0, "completions/mean_length": 35.546875, "completions/mean_terminated_length": 35.546875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1125, "frac_reward_zero_std": 0.625, "grad_norm": 47.5869255065918, "learning_rate": 6.387014543809223e-07, "loss": -0.0, "num_tokens": 4798898.0, "reward": 0.793749988079071, "reward_std": 0.19670185446739197, "rewards/format_reward/mean": 0.59375, "rewards/format_reward/std": 0.19669894874095917, "rewards/mcq_exact_match_reward/mean": 0.734375, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "step": 45 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 6.0, "completions/max_terminated_length": 6.0, "completions/mean_length": 6.0, "completions/mean_terminated_length": 6.0, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.115, "frac_reward_zero_std": 0.875, "grad_norm": 14.315831184387207, "learning_rate": 6.227427435703995e-07, "loss": 0.0, "num_tokens": 4905098.0, "reward": 0.5656249523162842, "reward_std": 0.04419416934251785, "rewards/format_reward/mean": 0.5, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.515625, "rewards/mcq_exact_match_reward/std": 0.5037065148353577, "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": 765.0, "completions/max_terminated_length": 765.0, "completions/mean_length": 26.28125, "completions/mean_terminated_length": 26.28125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1175, "frac_reward_zero_std": 0.5, "grad_norm": 16.962886810302734, "learning_rate": 6.066498153718734e-07, "loss": -0.0, "num_tokens": 4993452.0, "reward": 0.6570311784744263, "reward_std": 0.16428497433662415, "rewards/format_reward/mean": 0.4765625, "rewards/format_reward/std": 0.10652101784944534, "rewards/mcq_exact_match_reward/mean": 0.609375, "rewards/mcq_exact_match_reward/std": 0.4917473793029785, "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": 8.0, "completions/max_terminated_length": 8.0, "completions/mean_length": 6.0625, "completions/mean_terminated_length": 6.0625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.12, "frac_reward_zero_std": 0.625, "grad_norm": 22.988372802734375, "learning_rate": 5.90440267166055e-07, "loss": -0.0, "num_tokens": 5099120.0, "reward": 0.34609371423721313, "reward_std": 0.15529169142246246, "rewards/format_reward/mean": 0.4921875, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "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": 12.0, "completions/max_terminated_length": 12.0, "completions/mean_length": 6.75, "completions/mean_terminated_length": 6.75, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1225, "frac_reward_zero_std": 1.0, "grad_norm": 0.0, "learning_rate": 5.741318238559209e-07, "loss": 0.0, "num_tokens": 5189152.0, "reward": 0.6812499761581421, "reward_std": 0.0, "rewards/format_reward/mean": 0.5625, "rewards/format_reward/std": 0.1666666716337204, "rewards/mcq_exact_match_reward/mean": 0.625, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 186.0, "completions/max_terminated_length": 186.0, "completions/mean_length": 10.90625, "completions/mean_terminated_length": 10.90625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.125, "frac_reward_zero_std": 0.75, "grad_norm": 13.93482780456543, "learning_rate": 5.577423184847931e-07, "loss": 0.0, "num_tokens": 5269898.0, "reward": 0.657031238079071, "reward_std": 0.06007347255945206, "rewards/format_reward/mean": 0.6328125, "rewards/format_reward/std": 0.22257846593856812, "rewards/mcq_exact_match_reward/mean": 0.59375, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "step": 50 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 433.0, "completions/max_terminated_length": 433.0, "completions/mean_length": 26.359375, "completions/mean_terminated_length": 26.359375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1275, "frac_reward_zero_std": 0.625, "grad_norm": 21.169849395751953, "learning_rate": 5.412896727361662e-07, "loss": -0.0, "num_tokens": 5370137.0, "reward": 0.609375, "reward_std": 0.16887323558330536, "rewards/format_reward/mean": 0.625, "rewards/format_reward/std": 0.2182178944349289, "rewards/mcq_exact_match_reward/mean": 0.546875, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 51 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 6.0, "completions/max_terminated_length": 6.0, "completions/mean_length": 6.0, "completions/mean_terminated_length": 6.0, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.13, "frac_reward_zero_std": 0.25, "grad_norm": 44.459861755371094, "learning_rate": 5.247918773366111e-07, "loss": 0.0, "num_tokens": 5474185.0, "reward": 0.5187499523162842, "reward_std": 0.31300365924835205, "rewards/format_reward/mean": 0.5, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.46875, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "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": 12.0, "completions/max_terminated_length": 12.0, "completions/mean_length": 7.875, "completions/mean_terminated_length": 7.875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1325, "frac_reward_zero_std": 0.5, "grad_norm": 34.96181869506836, "learning_rate": 5.082669723831793e-07, "loss": 0.0, "num_tokens": 5570305.0, "reward": 0.7687499523162842, "reward_std": 0.10784431546926498, "rewards/format_reward/mean": 0.65625, "rewards/format_reward/std": 0.233588308095932, "rewards/mcq_exact_match_reward/mean": 0.703125, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "step": 53 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 12.0, "completions/max_terminated_length": 12.0, "completions/mean_length": 6.09375, "completions/mean_terminated_length": 6.09375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.135, "frac_reward_zero_std": 0.625, "grad_norm": 23.179283142089844, "learning_rate": 4.917330276168208e-07, "loss": 0.0, "num_tokens": 5681327.0, "reward": 0.6289061903953552, "reward_std": 0.11321918666362762, "rewards/format_reward/mean": 0.5078125, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.578125, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "step": 54 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 13.0, "completions/max_terminated_length": 13.0, "completions/mean_length": 8.0625, "completions/mean_terminated_length": 8.0625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1375, "frac_reward_zero_std": 0.5, "grad_norm": 45.4562873840332, "learning_rate": 4.752081226633888e-07, "loss": 0.0, "num_tokens": 5763475.0, "reward": 0.78125, "reward_std": 0.11570000648498535, "rewards/format_reward/mean": 0.625, "rewards/format_reward/std": 0.28171807527542114, "rewards/mcq_exact_match_reward/mean": 0.71875, "rewards/mcq_exact_match_reward/std": 0.4531635046005249, "step": 55 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 12.0, "completions/max_terminated_length": 12.0, "completions/mean_length": 7.59375, "completions/mean_terminated_length": 7.59375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.14, "frac_reward_zero_std": 0.625, "grad_norm": 20.14704132080078, "learning_rate": 4.5871032726383385e-07, "loss": 0.0, "num_tokens": 5865409.0, "reward": 0.40703123807907104, "reward_std": 0.11109752953052521, "rewards/format_reward/mean": 0.6328125, "rewards/format_reward/std": 0.22257846593856812, "rewards/mcq_exact_match_reward/mean": 0.34375, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "step": 56 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 525.0, "completions/max_terminated_length": 525.0, "completions/mean_length": 18.9375, "completions/mean_terminated_length": 18.9375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1425, "frac_reward_zero_std": 0.5, "grad_norm": 25.70990753173828, "learning_rate": 4.4225768151520694e-07, "loss": -0.0, "num_tokens": 5955429.0, "reward": 0.5640624761581421, "reward_std": 0.09276594966650009, "rewards/format_reward/mean": 0.640625, "rewards/format_reward/std": 0.24346621334552765, "rewards/mcq_exact_match_reward/mean": 0.5, "rewards/mcq_exact_match_reward/std": 0.5039526224136353, "step": 57 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 19.0, "completions/max_terminated_length": 19.0, "completions/mean_length": 9.015625, "completions/mean_terminated_length": 9.015625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.145, "frac_reward_zero_std": 0.625, "grad_norm": 17.63555335998535, "learning_rate": 4.258681761440789e-07, "loss": -0.0, "num_tokens": 6034854.0, "reward": 0.4921875, "reward_std": 0.10495111346244812, "rewards/format_reward/mean": 0.703125, "rewards/format_reward/std": 0.24750742316246033, "rewards/mcq_exact_match_reward/mean": 0.421875, "rewards/mcq_exact_match_reward/std": 0.49776285886764526, "step": 58 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1164.0, "completions/max_terminated_length": 1164.0, "completions/mean_length": 41.515625, "completions/mean_terminated_length": 41.515625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1475, "frac_reward_zero_std": 0.375, "grad_norm": 40.475502014160156, "learning_rate": 4.095597328339452e-07, "loss": 0.0, "num_tokens": 6130751.0, "reward": 0.39765626192092896, "reward_std": 0.11546652019023895, "rewards/format_reward/mean": 0.8515625, "rewards/format_reward/std": 0.2302463799715042, "rewards/mcq_exact_match_reward/mean": 0.3125, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "step": 59 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 551.0, "completions/max_terminated_length": 551.0, "completions/mean_length": 29.171875, "completions/mean_terminated_length": 29.171875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.15, "frac_reward_zero_std": 0.625, "grad_norm": 36.07637405395508, "learning_rate": 3.9335018462812664e-07, "loss": 0.0, "num_tokens": 6218122.0, "reward": 0.714062511920929, "reward_std": 0.1759525090456009, "rewards/format_reward/mean": 0.890625, "rewards/format_reward/std": 0.2083333432674408, "rewards/mcq_exact_match_reward/mean": 0.625, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "step": 60 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 389.0, "completions/max_terminated_length": 389.0, "completions/mean_length": 29.484375, "completions/mean_terminated_length": 29.484375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1525, "frac_reward_zero_std": 0.375, "grad_norm": 16.634305953979492, "learning_rate": 3.772572564296004e-07, "loss": -0.0, "num_tokens": 6323425.0, "reward": 0.32343751192092896, "reward_std": 0.17329266667366028, "rewards/format_reward/mean": 0.890625, "rewards/format_reward/std": 0.2083333432674408, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "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": 899.0, "completions/max_terminated_length": 899.0, "completions/mean_length": 69.53125, "completions/mean_terminated_length": 69.53125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.155, "frac_reward_zero_std": 0.5, "grad_norm": 11.543915748596191, "learning_rate": 3.612985456190778e-07, "loss": 0.0, "num_tokens": 6405131.0, "reward": 0.5296875238418579, "reward_std": 0.19812041521072388, "rewards/format_reward/mean": 0.921875, "rewards/format_reward/std": 0.20351573824882507, "rewards/mcq_exact_match_reward/mean": 0.4375, "rewards/mcq_exact_match_reward/std": 0.5, "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": 375.0, "completions/max_terminated_length": 375.0, "completions/mean_length": 26.484375, "completions/mean_terminated_length": 26.484375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1575, "frac_reward_zero_std": 0.25, "grad_norm": 74.34740447998047, "learning_rate": 3.454915028125263e-07, "loss": 0.0, "num_tokens": 6505386.0, "reward": 0.7632812857627869, "reward_std": 0.07966035604476929, "rewards/format_reward/mean": 0.9140625, "rewards/format_reward/std": 0.209963858127594, "rewards/mcq_exact_match_reward/mean": 0.671875, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "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": 873.0, "completions/mean_length": 69.625, "completions/mean_terminated_length": 38.222225189208984, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.16, "frac_reward_zero_std": 0.75, "grad_norm": 14.204571723937988, "learning_rate": 3.2985341277917846e-07, "loss": 0.0, "num_tokens": 6615122.0, "reward": 0.32500001788139343, "reward_std": 0.0530330091714859, "rewards/format_reward/mean": 0.90625, "rewards/format_reward/std": 0.233588308095932, "rewards/mcq_exact_match_reward/mean": 0.234375, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "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": 412.0, "completions/max_terminated_length": 412.0, "completions/mean_length": 21.53125, "completions/mean_terminated_length": 21.53125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1625, "frac_reward_zero_std": 0.625, "grad_norm": 20.502248764038086, "learning_rate": 3.1440137554088953e-07, "loss": 0.0, "num_tokens": 6726572.0, "reward": 0.7562500238418579, "reward_std": 0.1462521106004715, "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": 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": 681.0, "completions/max_terminated_length": 681.0, "completions/mean_length": 34.203125, "completions/mean_terminated_length": 34.203125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.165, "frac_reward_zero_std": 0.625, "grad_norm": 12.760836601257324, "learning_rate": 2.9915228767351535e-07, "loss": 0.0, "num_tokens": 6797241.0, "reward": 0.7703125476837158, "reward_std": 0.15534965693950653, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.671875, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "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": 300.0, "completions/max_terminated_length": 300.0, "completions/mean_length": 19.390625, "completions/mean_terminated_length": 19.390625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1675, "frac_reward_zero_std": 0.75, "grad_norm": 21.35260009765625, "learning_rate": 2.841228238307536e-07, "loss": 0.0, "num_tokens": 6869322.0, "reward": 0.9257813096046448, "reward_std": 0.10262773931026459, "rewards/format_reward/mean": 0.9765625, "rewards/format_reward/std": 0.10652101784944534, "rewards/mcq_exact_match_reward/mean": 0.828125, "rewards/mcq_exact_match_reward/std": 0.38025420904159546, "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": 607.0, "completions/max_terminated_length": 607.0, "completions/mean_length": 39.265625, "completions/mean_terminated_length": 39.265625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.17, "frac_reward_zero_std": 0.75, "grad_norm": 6.614924430847168, "learning_rate": 2.6932941851065615e-07, "loss": 0.0, "num_tokens": 6961595.0, "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": 68 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 495.0, "completions/max_terminated_length": 495.0, "completions/mean_length": 24.4375, "completions/mean_terminated_length": 24.4375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1725, "frac_reward_zero_std": 0.375, "grad_norm": 24.30486297607422, "learning_rate": 2.547882480847461e-07, "loss": 0.0, "num_tokens": 7065951.0, "reward": 0.5414062738418579, "reward_std": 0.18418928980827332, "rewards/format_reward/mean": 0.8828125, "rewards/format_reward/std": 0.3052735924720764, "rewards/mcq_exact_match_reward/mean": 0.453125, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 69 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1072.0, "completions/max_terminated_length": 1072.0, "completions/mean_length": 41.4375, "completions/mean_terminated_length": 41.4375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.175, "frac_reward_zero_std": 1.0, "grad_norm": 0.0, "learning_rate": 2.4051521310939254e-07, "loss": 0.0, "num_tokens": 7151091.0, "reward": 0.6000000238418579, "reward_std": 0.0, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.5, "rewards/mcq_exact_match_reward/std": 0.5039526224136353, "step": 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": 904.0, "completions/max_terminated_length": 904.0, "completions/mean_length": 61.953125, "completions/mean_terminated_length": 61.953125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1775, "frac_reward_zero_std": 0.75, "grad_norm": 7.29455041885376, "learning_rate": 2.2652592093878665e-07, "loss": 0.0, "num_tokens": 7266952.0, "reward": 0.45781251788139343, "reward_std": 0.13330145180225372, "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": 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": 1618.0, "completions/max_terminated_length": 1618.0, "completions/mean_length": 69.28125, "completions/mean_terminated_length": 69.28125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.18, "frac_reward_zero_std": 0.625, "grad_norm": 14.70644760131836, "learning_rate": 2.128356686585282e-07, "loss": 0.0, "num_tokens": 7347554.0, "reward": 0.5632812976837158, "reward_std": 0.14814922213554382, "rewards/format_reward/mean": 0.9453125, "rewards/format_reward/std": 0.180765300989151, "rewards/mcq_exact_match_reward/mean": 0.46875, "rewards/mcq_exact_match_reward/std": 0.5029674172401428, "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": 18.0, "completions/max_terminated_length": 18.0, "completions/mean_length": 12.125, "completions/mean_terminated_length": 12.125, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "epoch": 0.1825, "frac_reward_zero_std": 0.75, "grad_norm": 15.100133895874023, "learning_rate": 1.9945942635848745e-07, "loss": 0.0, "num_tokens": 7415778.0, "reward": 0.6140625476837158, "reward_std": 0.10647734999656677, "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": 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": 529.0, "completions/max_terminated_length": 529.0, "completions/mean_length": 21.84375, "completions/mean_terminated_length": 21.84375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.185, "frac_reward_zero_std": 0.75, "grad_norm": 10.5975980758667, "learning_rate": 1.8641182076323148e-07, "loss": 0.0, "num_tokens": 7512208.0, "reward": 0.5687500238418579, "reward_std": 0.1246790662407875, "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": 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": 373.0, "completions/max_terminated_length": 373.0, "completions/mean_length": 31.015625, "completions/mean_terminated_length": 31.015625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1875, "frac_reward_zero_std": 0.375, "grad_norm": 32.69182586669922, "learning_rate": 1.7370711923791564e-07, "loss": -0.0, "num_tokens": 7606025.0, "reward": 0.643750011920929, "reward_std": 0.2797520160675049, "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": 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": 223.0, "completions/max_terminated_length": 223.0, "completions/mean_length": 19.734375, "completions/mean_terminated_length": 19.734375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.19, "frac_reward_zero_std": 0.25, "grad_norm": 16.534635543823242, "learning_rate": 1.6135921418712955e-07, "loss": -0.0, "num_tokens": 7705680.0, "reward": 0.44687503576278687, "reward_std": 0.2517387270927429, "rewards/format_reward/mean": 0.875, "rewards/format_reward/std": 0.2519763112068176, "rewards/mcq_exact_match_reward/mean": 0.359375, "rewards/mcq_exact_match_reward/std": 0.4836103618144989, "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": 989.0, "completions/max_terminated_length": 989.0, "completions/mean_length": 27.390625, "completions/mean_terminated_length": 27.390625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.1925, "frac_reward_zero_std": 0.75, "grad_norm": 6.076826572418213, "learning_rate": 1.493816078637557e-07, "loss": -0.0, "num_tokens": 7803609.0, "reward": 0.8648437857627869, "reward_std": 0.046403881162405014, "rewards/format_reward/mean": 0.9921875, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.765625, "rewards/mcq_exact_match_reward/std": 0.42695629596710205, "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": 249.0, "completions/max_terminated_length": 249.0, "completions/mean_length": 20.328125, "completions/mean_terminated_length": 20.328125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.195, "frac_reward_zero_std": 0.25, "grad_norm": 44.02492141723633, "learning_rate": 1.3778739760445552e-07, "loss": 0.0, "num_tokens": 7899878.0, "reward": 0.5921875238418579, "reward_std": 0.22322005033493042, "rewards/format_reward/mean": 0.921875, "rewards/format_reward/std": 0.18298126757144928, "rewards/mcq_exact_match_reward/mean": 0.5, "rewards/mcq_exact_match_reward/std": 0.5039526224136353, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 321.0, "completions/mean_length": 51.859375, "completions/mean_terminated_length": 20.174604415893555, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1975, "frac_reward_zero_std": 0.75, "grad_norm": 6.103572845458984, "learning_rate": 1.2658926150792322e-07, "loss": 0.0, "num_tokens": 7999205.0, "reward": 0.5906250476837158, "reward_std": 0.0088388342410326, "rewards/format_reward/mean": 0.90625, "rewards/format_reward/std": 0.233588308095932, "rewards/mcq_exact_match_reward/mean": 0.5, "rewards/mcq_exact_match_reward/std": 0.5039526224136353, "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": 520.0, "completions/max_terminated_length": 520.0, "completions/mean_length": 41.609375, "completions/mean_terminated_length": 41.609375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2, "frac_reward_zero_std": 0.75, "grad_norm": 10.459931373596191, "learning_rate": 1.1579944457157059e-07, "loss": 0.0, "num_tokens": 8091900.0, "reward": 0.7406250238418579, "reward_std": 0.10205793380737305, "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": 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": 1210.0, "completions/max_terminated_length": 1210.0, "completions/mean_length": 46.171875, "completions/mean_terminated_length": 46.171875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2025, "frac_reward_zero_std": 0.625, "grad_norm": 12.259103775024414, "learning_rate": 1.0542974530180327e-07, "loss": 0.0, "num_tokens": 8203911.0, "reward": 0.8031250238418579, "reward_std": 0.13258251547813416, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.703125, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "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": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1723.0, "completions/mean_length": 105.875, "completions/mean_terminated_length": 75.04762268066406, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.205, "frac_reward_zero_std": 0.5, "grad_norm": 32.62446212768555, "learning_rate": 9.549150281252632e-08, "loss": 0.0, "num_tokens": 8298847.0, "reward": 0.6453125476837158, "reward_std": 0.21536517143249512, "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": 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": 349.0, "completions/max_terminated_length": 349.0, "completions/mean_length": 22.140625, "completions/mean_terminated_length": 22.140625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2075, "frac_reward_zero_std": 0.75, "grad_norm": 20.074138641357422, "learning_rate": 8.599558442598998e-08, "loss": 0.0, "num_tokens": 8382224.0, "reward": 0.8968750238418579, "reward_std": 0.10205793380737305, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.796875, "rewards/mcq_exact_match_reward/std": 0.40550529956817627, "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": 693.0, "completions/max_terminated_length": 693.0, "completions/mean_length": 25.75, "completions/mean_terminated_length": 25.75, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.21, "frac_reward_zero_std": 0.75, "grad_norm": 9.44982624053955, "learning_rate": 7.695237378953224e-08, "loss": 0.0, "num_tokens": 8477496.0, "reward": 0.4906250238418579, "reward_std": 0.1315089464187622, "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": 84 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 81.0, "completions/max_terminated_length": 81.0, "completions/mean_length": 13.140625, "completions/mean_terminated_length": 13.140625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.2125, "frac_reward_zero_std": 0.75, "grad_norm": 9.163570404052734, "learning_rate": 6.837175952121304e-08, "loss": 0.0, "num_tokens": 8550401.0, "reward": 0.421875, "reward_std": 0.11100947856903076, "rewards/format_reward/mean": 0.9375, "rewards/format_reward/std": 0.1666666716337204, "rewards/mcq_exact_match_reward/mean": 0.328125, "rewards/mcq_exact_match_reward/std": 0.4732423722743988, "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": 1087.0, "completions/max_terminated_length": 1087.0, "completions/mean_length": 31.140625, "completions/mean_terminated_length": 31.140625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.215, "frac_reward_zero_std": 0.875, "grad_norm": 6.9531965255737305, "learning_rate": 6.026312439675551e-08, "loss": 0.0, "num_tokens": 8637714.0, "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": 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": 416.0, "completions/max_terminated_length": 416.0, "completions/mean_length": 19.875, "completions/mean_terminated_length": 19.875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2175, "frac_reward_zero_std": 0.5, "grad_norm": 16.834619522094727, "learning_rate": 5.263533508961826e-08, "loss": 0.0, "num_tokens": 8752626.0, "reward": 0.6609375476837158, "reward_std": 0.20169559121131897, "rewards/format_reward/mean": 0.984375, "rewards/format_reward/std": 0.125, "rewards/mcq_exact_match_reward/mean": 0.5625, "rewards/mcq_exact_match_reward/std": 0.5, "step": 87 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 244.0, "completions/max_terminated_length": 244.0, "completions/mean_length": 25.421875, "completions/mean_terminated_length": 25.421875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.22, "frac_reward_zero_std": 0.5, "grad_norm": 20.588544845581055, "learning_rate": 4.549673247541874e-08, "loss": 0.0, "num_tokens": 8847109.0, "reward": 0.47187501192092896, "reward_std": 0.1720152497291565, "rewards/format_reward/mean": 0.96875, "rewards/format_reward/std": 0.17536810040473938, "rewards/mcq_exact_match_reward/mean": 0.375, "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": 833.0, "completions/max_terminated_length": 833.0, "completions/mean_length": 73.6875, "completions/mean_terminated_length": 73.6875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2225, "frac_reward_zero_std": 0.5, "grad_norm": 16.177391052246094, "learning_rate": 3.8855122511307626e-08, "loss": 0.0, "num_tokens": 8938049.0, "reward": 0.47187501192092896, "reward_std": 0.22823089361190796, "rewards/format_reward/mean": 0.96875, "rewards/format_reward/std": 0.17536810040473938, "rewards/mcq_exact_match_reward/mean": 0.375, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "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": 805.0, "completions/max_terminated_length": 805.0, "completions/mean_length": 35.359375, "completions/mean_terminated_length": 35.359375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.225, "frac_reward_zero_std": 0.5, "grad_norm": 19.034664154052734, "learning_rate": 3.271776770026963e-08, "loss": 0.0, "num_tokens": 9030152.0, "reward": 0.7875000238418579, "reward_std": 0.22461533546447754, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.6875, "rewards/mcq_exact_match_reward/std": 0.467176616191864, "step": 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": 301.0, "completions/max_terminated_length": 301.0, "completions/mean_length": 24.453125, "completions/mean_terminated_length": 24.453125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2275, "frac_reward_zero_std": 0.875, "grad_norm": 3.6413960456848145, "learning_rate": 2.7091379149682682e-08, "loss": 0.0, "num_tokens": 9110845.0, "reward": 0.8343750238418579, "reward_std": 0.04419417306780815, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.734375, "rewards/mcq_exact_match_reward/std": 0.44515693187713623, "step": 91 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 400.0, "completions/max_terminated_length": 400.0, "completions/mean_length": 21.5625, "completions/mean_terminated_length": 21.5625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.23, "frac_reward_zero_std": 0.75, "grad_norm": 11.584431648254395, "learning_rate": 2.1982109232821176e-08, "loss": -0.0, "num_tokens": 9212873.0, "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": 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": 2028.0, "completions/max_terminated_length": 2028.0, "completions/mean_length": 69.546875, "completions/mean_terminated_length": 69.546875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2325, "frac_reward_zero_std": 0.75, "grad_norm": 12.158230781555176, "learning_rate": 1.7395544861325718e-08, "loss": 0.0, "num_tokens": 9304844.0, "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": 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": 1206.0, "completions/max_terminated_length": 1206.0, "completions/mean_length": 50.3125, "completions/mean_terminated_length": 50.3125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.235, "frac_reward_zero_std": 0.625, "grad_norm": 20.270219802856445, "learning_rate": 1.3336701375997127e-08, "loss": -0.0, "num_tokens": 9391712.0, "reward": 0.3968750238418579, "reward_std": 0.1530819982290268, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.296875, "rewards/mcq_exact_match_reward/std": 0.4604927599430084, "step": 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": 92.0, "completions/max_terminated_length": 92.0, "completions/mean_length": 15.296875, "completions/mean_terminated_length": 15.296875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2375, "frac_reward_zero_std": 0.75, "grad_norm": 7.157800197601318, "learning_rate": 9.810017062595321e-09, "loss": 0.0, "num_tokens": 9496939.0, "reward": 0.4437500238418579, "reward_std": 0.0883883461356163, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.34375, "rewards/mcq_exact_match_reward/std": 0.4787135720252991, "step": 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": 486.0, "completions/max_terminated_length": 486.0, "completions/mean_length": 26.1875, "completions/mean_terminated_length": 26.1875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.24, "frac_reward_zero_std": 0.5, "grad_norm": 21.17302131652832, "learning_rate": 6.819348298638839e-09, "loss": -0.0, "num_tokens": 9600967.0, "reward": 0.7242187857627869, "reward_std": 0.1779128462076187, "rewards/format_reward/mean": 0.9921875, "rewards/format_reward/std": 0.0625, "rewards/mcq_exact_match_reward/mean": 0.625, "rewards/mcq_exact_match_reward/std": 0.48795005679130554, "step": 96 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1324.0, "completions/mean_length": 93.265625, "completions/mean_terminated_length": 62.23809814453125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2425, "frac_reward_zero_std": 0.25, "grad_norm": 13.476895332336426, "learning_rate": 4.367965336512403e-09, "loss": 0.0, "num_tokens": 9701576.0, "reward": 0.45781251788139343, "reward_std": 0.2890201807022095, "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": 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": 369.0, "completions/max_terminated_length": 369.0, "completions/mean_length": 40.421875, "completions/mean_terminated_length": 40.421875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.245, "frac_reward_zero_std": 0.25, "grad_norm": 22.92038345336914, "learning_rate": 2.458548727494292e-09, "loss": 0.0, "num_tokens": 9810283.0, "reward": 0.49687501788139343, "reward_std": 0.318790078163147, "rewards/format_reward/mean": 0.90625, "rewards/format_reward/std": 0.29378482699394226, "rewards/mcq_exact_match_reward/mean": 0.40625, "rewards/mcq_exact_match_reward/std": 0.49501484632492065, "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": 122.0, "completions/max_terminated_length": 122.0, "completions/mean_length": 16.484375, "completions/mean_terminated_length": 16.484375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.2475, "frac_reward_zero_std": 0.75, "grad_norm": 5.6133646965026855, "learning_rate": 1.0931863906127325e-09, "loss": 0.0, "num_tokens": 9910370.0, "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": 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": 590.0, "completions/max_terminated_length": 590.0, "completions/mean_length": 44.375, "completions/mean_terminated_length": 44.375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.25, "frac_reward_zero_std": 0.5, "grad_norm": 17.08899688720703, "learning_rate": 2.733713295369755e-10, "loss": 0.0, "num_tokens": 9991058.0, "reward": 0.6468750238418579, "reward_std": 0.2561880052089691, "rewards/format_reward/mean": 1.0, "rewards/format_reward/std": 0.0, "rewards/mcq_exact_match_reward/mean": 0.546875, "rewards/mcq_exact_match_reward/std": 0.501733124256134, "step": 100 } ], "logging_steps": 1, "max_steps": 100, "num_input_tokens_seen": 9991058, "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 }