| { |
| "best_metric": null, |
| "best_model_checkpoint": null, |
| "epoch": 0.004997251511668582, |
| "eval_steps": 500, |
| "global_step": 1200, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 116.9375, |
| "correct_distill_rate": 0.0375, |
| "distill_active_tokens": 8.625, |
| "distill_lambda": 0.5225, |
| "distill_loss": 0.10027681589126587, |
| "distill_pairs": 267.375, |
| "distill_skipped_no_correct_rate": 0.7, |
| "eligible_block_rate": 0.3, |
| "epoch": 4.1643762597238186e-05, |
| "grad_norm": 0.0473579503595829, |
| "grpo_loss": 0.002949414495378733, |
| "kl": 0.07373543381690979, |
| "learning_rate": 2e-06, |
| "loss": 0.0551, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.34749999865889547, |
| "reward_std": 0.04166190326213837, |
| "rewards/countdown_reward_func": 0.34749999865889547, |
| "step": 10, |
| "teacher_correct_rate": 0.3, |
| "teacher_step_mean": 9.6, |
| "teacher_step_min": 9.6, |
| "zero_std_ratio": 0.9 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.5875, |
| "correct_distill_rate": 0.0375, |
| "distill_active_tokens": 8.75, |
| "distill_lambda": 0.5725, |
| "distill_loss": 0.09738296568393708, |
| "distill_pairs": 271.25, |
| "distill_skipped_no_correct_rate": 0.7, |
| "eligible_block_rate": 0.3, |
| "epoch": 8.328752519447637e-05, |
| "grad_norm": 0.04033520445227623, |
| "grpo_loss": 0.0028141275281086562, |
| "kl": 0.07035325840115547, |
| "learning_rate": 2e-06, |
| "loss": 0.0575, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.34749999865889547, |
| "reward_std": 0.04166190326213837, |
| "rewards/countdown_reward_func": 0.34749999865889547, |
| "step": 20, |
| "teacher_correct_rate": 0.3, |
| "teacher_step_mean": 9.6, |
| "teacher_step_min": 9.6, |
| "zero_std_ratio": 0.9 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 119.3, |
| "correct_distill_rate": 0.025, |
| "distill_active_tokens": 5.95, |
| "distill_lambda": 0.6224999999999999, |
| "distill_loss": 0.06997707486152649, |
| "distill_pairs": 184.45, |
| "distill_skipped_no_correct_rate": 0.8, |
| "eligible_block_rate": 0.2, |
| "epoch": 0.00012493128779171456, |
| "grad_norm": 0.0612516887485981, |
| "grpo_loss": 0.002280982898082584, |
| "kl": 0.057024573907256126, |
| "learning_rate": 2e-06, |
| "loss": 0.0469, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.2800000011920929, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.2800000011920929, |
| "step": 30, |
| "teacher_correct_rate": 0.2, |
| "teacher_step_mean": 6.4, |
| "teacher_step_min": 6.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 119.4875, |
| "correct_distill_rate": 0.0125, |
| "distill_active_tokens": 3.025, |
| "distill_lambda": 0.6725, |
| "distill_loss": 0.028908628225326537, |
| "distill_pairs": 93.775, |
| "distill_skipped_no_correct_rate": 0.9, |
| "eligible_block_rate": 0.1, |
| "epoch": 0.00016657505038895275, |
| "grad_norm": 0.027009356766939163, |
| "grpo_loss": 0.001997075858525932, |
| "kl": 0.049926896765828135, |
| "learning_rate": 2e-06, |
| "loss": 0.0209, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.1900000013411045, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.1900000013411045, |
| "step": 40, |
| "teacher_correct_rate": 0.1, |
| "teacher_step_mean": 3.2, |
| "teacher_step_min": 3.2, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 119.95, |
| "correct_distill_rate": 0.0625, |
| "distill_active_tokens": 15.125, |
| "distill_lambda": 0.7224999999999999, |
| "distill_loss": 0.1480610191822052, |
| "distill_pairs": 468.875, |
| "distill_skipped_no_correct_rate": 0.5, |
| "eligible_block_rate": 0.4, |
| "epoch": 0.00020821881298619093, |
| "grad_norm": 1.0446033477783203, |
| "grpo_loss": 0.002224944659974426, |
| "kl": 0.055623612552881244, |
| "learning_rate": 2e-06, |
| "loss": 0.1084, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.4712500013411045, |
| "reward_std": 0.03181980550289154, |
| "rewards/countdown_reward_func": 0.4712500013411045, |
| "step": 50, |
| "teacher_correct_rate": 0.4, |
| "teacher_step_mean": 12.8, |
| "teacher_step_min": 12.8, |
| "zero_std_ratio": 0.9 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.9375, |
| "correct_distill_rate": 0.05, |
| "distill_active_tokens": 11.825, |
| "distill_lambda": 0.7725000000000001, |
| "distill_loss": 0.11083452999591828, |
| "distill_pairs": 366.575, |
| "distill_skipped_no_correct_rate": 0.6, |
| "eligible_block_rate": 0.4, |
| "epoch": 0.0002498625755834291, |
| "grad_norm": 1.3056390285491943, |
| "grpo_loss": 0.0020815988653339447, |
| "kl": 0.05686977878212929, |
| "learning_rate": 2e-06, |
| "loss": 0.0876, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.3812499985098839, |
| "reward_std": 0.12158873081207275, |
| "rewards/countdown_reward_func": 0.3812499985098839, |
| "step": 60, |
| "teacher_correct_rate": 0.4, |
| "teacher_step_mean": 12.8, |
| "teacher_step_min": 12.8, |
| "zero_std_ratio": 0.7 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 119.375, |
| "correct_distill_rate": 0.05, |
| "distill_active_tokens": 11.925, |
| "distill_lambda": 0.8225, |
| "distill_loss": 0.09993059784173966, |
| "distill_pairs": 369.675, |
| "distill_skipped_no_correct_rate": 0.6, |
| "eligible_block_rate": 0.4, |
| "epoch": 0.0002915063381806673, |
| "grad_norm": 2.4835469722747803, |
| "grpo_loss": 0.0024276475189253686, |
| "kl": 0.06069119200110436, |
| "learning_rate": 2e-06, |
| "loss": 0.0844, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.46000000089406967, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.46000000089406967, |
| "step": 70, |
| "teacher_correct_rate": 0.4, |
| "teacher_step_mean": 12.8, |
| "teacher_step_min": 12.8, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.75, |
| "correct_distill_rate": 0.0375, |
| "distill_active_tokens": 8.975, |
| "distill_lambda": 0.8724999999999999, |
| "distill_loss": 0.06559722125530243, |
| "distill_pairs": 278.225, |
| "distill_skipped_no_correct_rate": 0.7, |
| "eligible_block_rate": 0.3, |
| "epoch": 0.0003331501007779055, |
| "grad_norm": 1.195908546447754, |
| "grpo_loss": 0.002371857478283346, |
| "kl": 0.059296438470482826, |
| "learning_rate": 2e-06, |
| "loss": 0.0603, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.3700000010430813, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.3700000010430813, |
| "step": 80, |
| "teacher_correct_rate": 0.3, |
| "teacher_step_mean": 9.6, |
| "teacher_step_min": 9.6, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 119.2, |
| "correct_distill_rate": 0.025, |
| "distill_active_tokens": 5.875, |
| "distill_lambda": 0.9225, |
| "distill_loss": 0.05624785572290421, |
| "distill_pairs": 182.125, |
| "distill_skipped_no_correct_rate": 0.8, |
| "eligible_block_rate": 0.2, |
| "epoch": 0.0003747938633751437, |
| "grad_norm": 0.04264846071600914, |
| "grpo_loss": 0.002276410860940814, |
| "kl": 0.05691027082502842, |
| "learning_rate": 2e-06, |
| "loss": 0.0534, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.2800000011920929, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.2800000011920929, |
| "step": 90, |
| "teacher_correct_rate": 0.2, |
| "teacher_step_mean": 6.4, |
| "teacher_step_min": 6.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 119.425, |
| "correct_distill_rate": 0.0375, |
| "distill_active_tokens": 9.025, |
| "distill_lambda": 0.9724999999999998, |
| "distill_loss": 0.06245248168706894, |
| "distill_pairs": 279.775, |
| "distill_skipped_no_correct_rate": 0.7, |
| "eligible_block_rate": 0.3, |
| "epoch": 0.00041643762597238186, |
| "grad_norm": 2.6133651733398438, |
| "grpo_loss": 0.0026280033285729586, |
| "kl": 0.0657000843435526, |
| "learning_rate": 2e-06, |
| "loss": 0.0638, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.3700000010430813, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.3700000010430813, |
| "step": 100, |
| "teacher_correct_rate": 0.3, |
| "teacher_step_mean": 9.6, |
| "teacher_step_min": 9.6, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 119.3375, |
| "correct_distill_rate": 0.0125, |
| "distill_active_tokens": 2.975, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.021732477843761443, |
| "distill_pairs": 92.225, |
| "distill_skipped_no_correct_rate": 0.9, |
| "eligible_block_rate": 0.1, |
| "epoch": 0.00045808138856962005, |
| "grad_norm": 0.02853594906628132, |
| "grpo_loss": 0.002150285360403359, |
| "kl": 0.053757134824991226, |
| "learning_rate": 2e-06, |
| "loss": 0.0239, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.1900000013411045, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.1900000013411045, |
| "step": 110, |
| "teacher_correct_rate": 0.1, |
| "teacher_step_mean": 3.2, |
| "teacher_step_min": 3.2, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.3, |
| "correct_distill_rate": 0.025, |
| "distill_active_tokens": 5.9, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.04337462484836578, |
| "distill_pairs": 182.9, |
| "distill_skipped_no_correct_rate": 0.8, |
| "eligible_block_rate": 0.2, |
| "epoch": 0.0004997251511668582, |
| "grad_norm": 0.05574803426861763, |
| "grpo_loss": 0.0019564196816645564, |
| "kl": 0.0489104937762022, |
| "learning_rate": 2e-06, |
| "loss": 0.0453, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.2800000011920929, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.2800000011920929, |
| "step": 120, |
| "teacher_correct_rate": 0.2, |
| "teacher_step_mean": 6.4, |
| "teacher_step_min": 6.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.6, |
| "correct_distill_rate": 0.0375, |
| "distill_active_tokens": 8.975, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.06203582286834717, |
| "distill_pairs": 278.225, |
| "distill_skipped_no_correct_rate": 0.7, |
| "eligible_block_rate": 0.3, |
| "epoch": 0.0005413689137640964, |
| "grad_norm": 0.05419427528977394, |
| "grpo_loss": 0.0022394997533410786, |
| "kl": 0.05598749555647373, |
| "learning_rate": 2e-06, |
| "loss": 0.0643, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.3700000010430813, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.3700000010430813, |
| "step": 130, |
| "teacher_correct_rate": 0.3, |
| "teacher_step_mean": 9.6, |
| "teacher_step_min": 9.6, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 119.7, |
| "correct_distill_rate": 0.0125, |
| "distill_active_tokens": 2.975, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.020414377748966216, |
| "distill_pairs": 92.225, |
| "distill_skipped_no_correct_rate": 0.9, |
| "eligible_block_rate": 0.1, |
| "epoch": 0.0005830126763613346, |
| "grad_norm": 0.056853942573070526, |
| "grpo_loss": 0.0022451088996604087, |
| "kl": 0.05612772218883037, |
| "learning_rate": 2e-06, |
| "loss": 0.0227, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.1900000013411045, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.1900000013411045, |
| "step": 140, |
| "teacher_correct_rate": 0.1, |
| "teacher_step_mean": 3.2, |
| "teacher_step_min": 3.2, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.1375, |
| "correct_distill_rate": 0.0375, |
| "distill_active_tokens": 9.0, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.07003793567419052, |
| "distill_pairs": 279.0, |
| "distill_skipped_no_correct_rate": 0.7, |
| "eligible_block_rate": 0.3, |
| "epoch": 0.0006246564389585727, |
| "grad_norm": 2.7419216632843018, |
| "grpo_loss": 0.00220655903685838, |
| "kl": 0.055163977295160295, |
| "learning_rate": 2e-06, |
| "loss": 0.0722, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.3700000010430813, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.3700000010430813, |
| "step": 150, |
| "teacher_correct_rate": 0.3, |
| "teacher_step_mean": 9.6, |
| "teacher_step_min": 9.6, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 119.2, |
| "correct_distill_rate": 0.0375, |
| "distill_active_tokens": 8.95, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.0763092502951622, |
| "distill_pairs": 277.45, |
| "distill_skipped_no_correct_rate": 0.7, |
| "eligible_block_rate": 0.3, |
| "epoch": 0.000666300201555811, |
| "grad_norm": 1.848865032196045, |
| "grpo_loss": 0.001968178804963827, |
| "kl": 0.04920447170734406, |
| "learning_rate": 2e-06, |
| "loss": 0.0783, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.3700000010430813, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.3700000010430813, |
| "step": 160, |
| "teacher_correct_rate": 0.3, |
| "teacher_step_mean": 9.6, |
| "teacher_step_min": 9.6, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.625, |
| "correct_distill_rate": 0.0, |
| "distill_active_tokens": 0.0, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.0, |
| "distill_pairs": 0.0, |
| "distill_skipped_no_correct_rate": 1.0, |
| "eligible_block_rate": 0.0, |
| "epoch": 0.0007079439641530491, |
| "grad_norm": 0.029138779267668724, |
| "grpo_loss": 0.002021158393472433, |
| "kl": 0.05052896067500114, |
| "learning_rate": 2e-06, |
| "loss": 0.002, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.10000000149011612, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.10000000149011612, |
| "step": 170, |
| "teacher_correct_rate": 0.0, |
| "teacher_step_mean": 0.0, |
| "teacher_step_min": 0.0, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.375, |
| "correct_distill_rate": 0.05, |
| "distill_active_tokens": 11.9, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.06838815659284592, |
| "distill_pairs": 368.9, |
| "distill_skipped_no_correct_rate": 0.6, |
| "eligible_block_rate": 0.4, |
| "epoch": 0.0007495877267502874, |
| "grad_norm": 0.033792804926633835, |
| "grpo_loss": 0.00208354004425928, |
| "kl": 0.052088503167033194, |
| "learning_rate": 2e-06, |
| "loss": 0.0705, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.46000000089406967, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.46000000089406967, |
| "step": 180, |
| "teacher_correct_rate": 0.4, |
| "teacher_step_mean": 12.8, |
| "teacher_step_min": 12.8, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 116.5, |
| "correct_distill_rate": 0.075, |
| "distill_active_tokens": 17.325, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.11272647678852081, |
| "distill_pairs": 537.075, |
| "distill_skipped_no_correct_rate": 0.4, |
| "eligible_block_rate": 0.5, |
| "epoch": 0.0007912314893475255, |
| "grad_norm": 2.0204617977142334, |
| "grpo_loss": 0.0022038140101358294, |
| "kl": 0.05007304437458515, |
| "learning_rate": 2e-06, |
| "loss": 0.1149, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.5837500005960464, |
| "reward_std": 0.07992682754993438, |
| "rewards/countdown_reward_func": 0.5837500005960464, |
| "step": 190, |
| "teacher_correct_rate": 0.5, |
| "teacher_step_mean": 16.0, |
| "teacher_step_min": 16.0, |
| "zero_std_ratio": 0.8 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 119.6, |
| "correct_distill_rate": 0.0125, |
| "distill_active_tokens": 3.025, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.02132212072610855, |
| "distill_pairs": 93.775, |
| "distill_skipped_no_correct_rate": 0.9, |
| "eligible_block_rate": 0.1, |
| "epoch": 0.0008328752519447637, |
| "grad_norm": 0.06612089276313782, |
| "grpo_loss": 0.0023776727728545667, |
| "kl": 0.059441820904612544, |
| "learning_rate": 2e-06, |
| "loss": 0.0237, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.1900000013411045, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.1900000013411045, |
| "step": 200, |
| "teacher_correct_rate": 0.1, |
| "teacher_step_mean": 3.2, |
| "teacher_step_min": 3.2, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.725, |
| "correct_distill_rate": 0.0125, |
| "distill_active_tokens": 2.95, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.016931504011154175, |
| "distill_pairs": 91.45, |
| "distill_skipped_no_correct_rate": 0.9, |
| "eligible_block_rate": 0.1, |
| "epoch": 0.0008745190145420019, |
| "grad_norm": 0.052074432373046875, |
| "grpo_loss": 0.0021197052556090057, |
| "kl": 0.05299263261258602, |
| "learning_rate": 2e-06, |
| "loss": 0.0191, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.1900000013411045, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.1900000013411045, |
| "step": 210, |
| "teacher_correct_rate": 0.1, |
| "teacher_step_mean": 3.2, |
| "teacher_step_min": 3.2, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 116.7125, |
| "correct_distill_rate": 0.025, |
| "distill_active_tokens": 5.875, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.04089897572994232, |
| "distill_pairs": 182.125, |
| "distill_skipped_no_correct_rate": 0.8, |
| "eligible_block_rate": 0.2, |
| "epoch": 0.0009161627771392401, |
| "grad_norm": 0.03555711731314659, |
| "grpo_loss": 0.0022087479126639666, |
| "kl": 0.05521869845688343, |
| "learning_rate": 2e-06, |
| "loss": 0.0431, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.2800000011920929, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.2800000011920929, |
| "step": 220, |
| "teacher_correct_rate": 0.2, |
| "teacher_step_mean": 6.4, |
| "teacher_step_min": 6.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.15, |
| "correct_distill_rate": 0.0375, |
| "distill_active_tokens": 9.025, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.041565853357315066, |
| "distill_pairs": 279.775, |
| "distill_skipped_no_correct_rate": 0.7, |
| "eligible_block_rate": 0.3, |
| "epoch": 0.0009578065397364782, |
| "grad_norm": 0.01943708024919033, |
| "grpo_loss": 0.0019857927109114826, |
| "kl": 0.04964487478137016, |
| "learning_rate": 2e-06, |
| "loss": 0.0436, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.3587499998509884, |
| "reward_std": 0.03181980550289154, |
| "rewards/countdown_reward_func": 0.3587499998509884, |
| "step": 230, |
| "teacher_correct_rate": 0.3, |
| "teacher_step_mean": 9.6, |
| "teacher_step_min": 9.6, |
| "zero_std_ratio": 0.9 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 116.9375, |
| "correct_distill_rate": 0.0125, |
| "distill_active_tokens": 2.875, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.021218661963939667, |
| "distill_pairs": 89.125, |
| "distill_skipped_no_correct_rate": 0.9, |
| "eligible_block_rate": 0.1, |
| "epoch": 0.0009994503023337165, |
| "grad_norm": 0.02304602973163128, |
| "grpo_loss": 0.0021223828196525575, |
| "kl": 0.05183143056929111, |
| "learning_rate": 2e-06, |
| "loss": 0.0233, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.1787500001490116, |
| "reward_std": 0.03181980550289154, |
| "rewards/countdown_reward_func": 0.1787500001490116, |
| "step": 240, |
| "teacher_correct_rate": 0.1, |
| "teacher_step_mean": 3.2, |
| "teacher_step_min": 3.2, |
| "zero_std_ratio": 0.9 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.1875, |
| "correct_distill_rate": 0.0125, |
| "distill_active_tokens": 3.025, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.01580086350440979, |
| "distill_pairs": 93.775, |
| "distill_skipped_no_correct_rate": 0.9, |
| "eligible_block_rate": 0.1, |
| "epoch": 0.0010410940649309547, |
| "grad_norm": 1.9513165950775146, |
| "grpo_loss": 0.001872677868232131, |
| "kl": 0.04708518460392952, |
| "learning_rate": 2e-06, |
| "loss": 0.0177, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.1875000014901161, |
| "reward_std": 0.004629100486636162, |
| "rewards/countdown_reward_func": 0.1875000014901161, |
| "step": 250, |
| "teacher_correct_rate": 0.1, |
| "teacher_step_mean": 3.2, |
| "teacher_step_min": 3.2, |
| "zero_std_ratio": 0.9 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.3, |
| "correct_distill_rate": 0.05, |
| "distill_active_tokens": 11.85, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.07058436125516891, |
| "distill_pairs": 367.35, |
| "distill_skipped_no_correct_rate": 0.6, |
| "eligible_block_rate": 0.4, |
| "epoch": 0.0010827378275281927, |
| "grad_norm": 0.02116476371884346, |
| "grpo_loss": 0.0019912143354304133, |
| "kl": 0.0497804019600153, |
| "learning_rate": 2e-06, |
| "loss": 0.0726, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.44874999970197677, |
| "reward_std": 0.03181980550289154, |
| "rewards/countdown_reward_func": 0.44874999970197677, |
| "step": 260, |
| "teacher_correct_rate": 0.4, |
| "teacher_step_mean": 12.8, |
| "teacher_step_min": 12.8, |
| "zero_std_ratio": 0.9 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 116.7875, |
| "correct_distill_rate": 0.025, |
| "distill_active_tokens": 5.775, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.06586634516716003, |
| "distill_pairs": 179.025, |
| "distill_skipped_no_correct_rate": 0.8, |
| "eligible_block_rate": 0.2, |
| "epoch": 0.001124381590125431, |
| "grad_norm": 0.027689794078469276, |
| "grpo_loss": 0.0020254326751455664, |
| "kl": 0.050635816529393196, |
| "learning_rate": 2e-06, |
| "loss": 0.0679, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.2800000011920929, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.2800000011920929, |
| "step": 270, |
| "teacher_correct_rate": 0.2, |
| "teacher_step_mean": 6.4, |
| "teacher_step_min": 6.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.4, |
| "correct_distill_rate": 0.0375, |
| "distill_active_tokens": 8.85, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.0542089655995369, |
| "distill_pairs": 274.35, |
| "distill_skipped_no_correct_rate": 0.7, |
| "eligible_block_rate": 0.3, |
| "epoch": 0.0011660253527226692, |
| "grad_norm": 2.306246757507324, |
| "grpo_loss": 0.002030787628609687, |
| "kl": 0.05076969228684902, |
| "learning_rate": 2e-06, |
| "loss": 0.0562, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.3700000010430813, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.3700000010430813, |
| "step": 280, |
| "teacher_correct_rate": 0.3, |
| "teacher_step_mean": 9.6, |
| "teacher_step_min": 9.6, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.475, |
| "correct_distill_rate": 0.0, |
| "distill_active_tokens": 0.0, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.0, |
| "distill_pairs": 0.0, |
| "distill_skipped_no_correct_rate": 1.0, |
| "eligible_block_rate": 0.0, |
| "epoch": 0.0012076691153199075, |
| "grad_norm": 0.018009664490818977, |
| "grpo_loss": 0.0018333346699364484, |
| "kl": 0.04583336897194386, |
| "learning_rate": 2e-06, |
| "loss": 0.0018, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.10000000149011612, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.10000000149011612, |
| "step": 290, |
| "teacher_correct_rate": 0.0, |
| "teacher_step_mean": 0.0, |
| "teacher_step_min": 0.0, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.7, |
| "correct_distill_rate": 0.0375, |
| "distill_active_tokens": 8.75, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.05150433331727981, |
| "distill_pairs": 271.25, |
| "distill_skipped_no_correct_rate": 0.7, |
| "eligible_block_rate": 0.3, |
| "epoch": 0.0012493128779171455, |
| "grad_norm": 0.02741893380880356, |
| "grpo_loss": 0.001777921407483518, |
| "kl": 0.044448037073016165, |
| "learning_rate": 2e-06, |
| "loss": 0.0533, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.3700000010430813, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.3700000010430813, |
| "step": 300, |
| "teacher_correct_rate": 0.3, |
| "teacher_step_mean": 9.6, |
| "teacher_step_min": 9.6, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.1375, |
| "correct_distill_rate": 0.0375, |
| "distill_active_tokens": 8.675, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.05892942100763321, |
| "distill_pairs": 268.925, |
| "distill_skipped_no_correct_rate": 0.7, |
| "eligible_block_rate": 0.3, |
| "epoch": 0.0012909566405143837, |
| "grad_norm": 2.2237205505371094, |
| "grpo_loss": 0.001850801520049572, |
| "kl": 0.04869457334280014, |
| "learning_rate": 2e-06, |
| "loss": 0.0608, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.3362500034272671, |
| "reward_std": 0.04657942354679108, |
| "rewards/countdown_reward_func": 0.3362500034272671, |
| "step": 310, |
| "teacher_correct_rate": 0.3, |
| "teacher_step_mean": 9.6, |
| "teacher_step_min": 9.6, |
| "zero_std_ratio": 0.9 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.325, |
| "correct_distill_rate": 0.05, |
| "distill_active_tokens": 11.725, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.06612278297543525, |
| "distill_pairs": 363.475, |
| "distill_skipped_no_correct_rate": 0.6, |
| "eligible_block_rate": 0.4, |
| "epoch": 0.001332600403111622, |
| "grad_norm": 4.420324325561523, |
| "grpo_loss": 0.001976323639973998, |
| "kl": 0.04940809085965157, |
| "learning_rate": 2e-06, |
| "loss": 0.0681, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.46000000089406967, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.46000000089406967, |
| "step": 320, |
| "teacher_correct_rate": 0.4, |
| "teacher_step_mean": 12.8, |
| "teacher_step_min": 12.8, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.6, |
| "correct_distill_rate": 0.0625, |
| "distill_active_tokens": 14.7, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.09555412977933883, |
| "distill_pairs": 455.7, |
| "distill_skipped_no_correct_rate": 0.5, |
| "eligible_block_rate": 0.5, |
| "epoch": 0.0013742441657088602, |
| "grad_norm": 0.02176631987094879, |
| "grpo_loss": 0.0019719922565855086, |
| "kl": 0.04929983280599117, |
| "learning_rate": 2e-06, |
| "loss": 0.0975, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.5387499995529652, |
| "reward_std": 0.03181980550289154, |
| "rewards/countdown_reward_func": 0.5387499995529652, |
| "step": 330, |
| "teacher_correct_rate": 0.5, |
| "teacher_step_mean": 16.0, |
| "teacher_step_min": 16.0, |
| "zero_std_ratio": 0.9 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.1, |
| "correct_distill_rate": 0.025, |
| "distill_active_tokens": 5.95, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.02215418815612793, |
| "distill_pairs": 184.45, |
| "distill_skipped_no_correct_rate": 0.8, |
| "eligible_block_rate": 0.2, |
| "epoch": 0.0014158879283060982, |
| "grad_norm": 0.024003691971302032, |
| "grpo_loss": 0.001744992018211633, |
| "kl": 0.04362480118870735, |
| "learning_rate": 2e-06, |
| "loss": 0.0239, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.2800000011920929, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.2800000011920929, |
| "step": 340, |
| "teacher_correct_rate": 0.2, |
| "teacher_step_mean": 6.4, |
| "teacher_step_min": 6.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.475, |
| "correct_distill_rate": 0.0375, |
| "distill_active_tokens": 8.95, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.043416925519704816, |
| "distill_pairs": 277.45, |
| "distill_skipped_no_correct_rate": 0.7, |
| "eligible_block_rate": 0.3, |
| "epoch": 0.0014575316909033365, |
| "grad_norm": 6.062681674957275, |
| "grpo_loss": 0.0018388222553767264, |
| "kl": 0.0459705576300621, |
| "learning_rate": 2e-06, |
| "loss": 0.0453, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.3700000010430813, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.3700000010430813, |
| "step": 350, |
| "teacher_correct_rate": 0.3, |
| "teacher_step_mean": 9.6, |
| "teacher_step_min": 9.6, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 119.575, |
| "correct_distill_rate": 0.0375, |
| "distill_active_tokens": 9.025, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.06833959594368935, |
| "distill_pairs": 279.775, |
| "distill_skipped_no_correct_rate": 0.7, |
| "eligible_block_rate": 0.3, |
| "epoch": 0.0014991754535005747, |
| "grad_norm": 0.034300971776247025, |
| "grpo_loss": 0.0018975758226588368, |
| "kl": 0.047439396381378174, |
| "learning_rate": 2e-06, |
| "loss": 0.0702, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.3700000010430813, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.3700000010430813, |
| "step": 360, |
| "teacher_correct_rate": 0.3, |
| "teacher_step_mean": 9.6, |
| "teacher_step_min": 9.6, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.9625, |
| "correct_distill_rate": 0.025, |
| "distill_active_tokens": 6.0, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.02234559804201126, |
| "distill_pairs": 186.0, |
| "distill_skipped_no_correct_rate": 0.8, |
| "eligible_block_rate": 0.2, |
| "epoch": 0.001540819216097813, |
| "grad_norm": 0.03386453539133072, |
| "grpo_loss": 0.0020326102850958703, |
| "kl": 0.05081525854766369, |
| "learning_rate": 2e-06, |
| "loss": 0.0244, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.2800000011920929, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.2800000011920929, |
| "step": 370, |
| "teacher_correct_rate": 0.2, |
| "teacher_step_mean": 6.4, |
| "teacher_step_min": 6.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.0, |
| "correct_distill_rate": 0.0125, |
| "distill_active_tokens": 2.875, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.015381811559200287, |
| "distill_pairs": 89.125, |
| "distill_skipped_no_correct_rate": 0.9, |
| "eligible_block_rate": 0.1, |
| "epoch": 0.001582462978695051, |
| "grad_norm": 0.019087934866547585, |
| "grpo_loss": 0.001766977144870907, |
| "kl": 0.04417442940175533, |
| "learning_rate": 2e-06, |
| "loss": 0.0171, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.1900000013411045, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.1900000013411045, |
| "step": 380, |
| "teacher_correct_rate": 0.1, |
| "teacher_step_mean": 3.2, |
| "teacher_step_min": 3.2, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 119.8, |
| "correct_distill_rate": 0.025, |
| "distill_active_tokens": 6.0, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.03494762182235718, |
| "distill_pairs": 186.0, |
| "distill_skipped_no_correct_rate": 0.8, |
| "eligible_block_rate": 0.2, |
| "epoch": 0.0016241067412922892, |
| "grad_norm": 0.024580754339694977, |
| "grpo_loss": 0.001987570594064891, |
| "kl": 0.049689266458153726, |
| "learning_rate": 2e-06, |
| "loss": 0.0369, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.2800000011920929, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.2800000011920929, |
| "step": 390, |
| "teacher_correct_rate": 0.2, |
| "teacher_step_mean": 6.4, |
| "teacher_step_min": 6.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 119.7, |
| "correct_distill_rate": 0.0625, |
| "distill_active_tokens": 14.9, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.06795874610543251, |
| "distill_pairs": 461.9, |
| "distill_skipped_no_correct_rate": 0.5, |
| "eligible_block_rate": 0.5, |
| "epoch": 0.0016657505038895275, |
| "grad_norm": 0.026707541197538376, |
| "grpo_loss": 0.0019142629811540247, |
| "kl": 0.04785657674074173, |
| "learning_rate": 2e-06, |
| "loss": 0.0699, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.5500000007450581, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.5500000007450581, |
| "step": 400, |
| "teacher_correct_rate": 0.5, |
| "teacher_step_mean": 16.0, |
| "teacher_step_min": 16.0, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.7, |
| "correct_distill_rate": 0.0625, |
| "distill_active_tokens": 14.925, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.05689811110496521, |
| "distill_pairs": 462.675, |
| "distill_skipped_no_correct_rate": 0.5, |
| "eligible_block_rate": 0.5, |
| "epoch": 0.0017073942664867657, |
| "grad_norm": 3.4637458324432373, |
| "grpo_loss": 0.002052889100741595, |
| "kl": 0.0513222299516201, |
| "learning_rate": 2e-06, |
| "loss": 0.059, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.5500000007450581, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.5500000007450581, |
| "step": 410, |
| "teacher_correct_rate": 0.5, |
| "teacher_step_mean": 16.0, |
| "teacher_step_min": 16.0, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.7875, |
| "correct_distill_rate": 0.025, |
| "distill_active_tokens": 5.85, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.028889547288417815, |
| "distill_pairs": 181.35, |
| "distill_skipped_no_correct_rate": 0.8, |
| "eligible_block_rate": 0.2, |
| "epoch": 0.0017490380290840037, |
| "grad_norm": 4.4815545082092285, |
| "grpo_loss": 0.0018659217399545013, |
| "kl": 0.04664804488420486, |
| "learning_rate": 2e-06, |
| "loss": 0.0308, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.2800000011920929, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.2800000011920929, |
| "step": 420, |
| "teacher_correct_rate": 0.2, |
| "teacher_step_mean": 6.4, |
| "teacher_step_min": 6.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 120.2, |
| "correct_distill_rate": 0.05, |
| "distill_active_tokens": 12.1, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.06863014847040176, |
| "distill_pairs": 375.1, |
| "distill_skipped_no_correct_rate": 0.6, |
| "eligible_block_rate": 0.4, |
| "epoch": 0.001790681791681242, |
| "grad_norm": 0.025247307494282722, |
| "grpo_loss": 0.0015974719659425317, |
| "kl": 0.03993680104613304, |
| "learning_rate": 2e-06, |
| "loss": 0.0702, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.46000000089406967, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.46000000089406967, |
| "step": 430, |
| "teacher_correct_rate": 0.4, |
| "teacher_step_mean": 12.8, |
| "teacher_step_min": 12.8, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 119.2, |
| "correct_distill_rate": 0.0375, |
| "distill_active_tokens": 9.075, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.033978326618671416, |
| "distill_pairs": 281.325, |
| "distill_skipped_no_correct_rate": 0.7, |
| "eligible_block_rate": 0.3, |
| "epoch": 0.0018323255542784802, |
| "grad_norm": 0.019356798380613327, |
| "grpo_loss": 0.0017639909172430634, |
| "kl": 0.044099774211645126, |
| "learning_rate": 2e-06, |
| "loss": 0.0357, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.3700000010430813, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.3700000010430813, |
| "step": 440, |
| "teacher_correct_rate": 0.3, |
| "teacher_step_mean": 9.6, |
| "teacher_step_min": 9.6, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 119.3875, |
| "correct_distill_rate": 0.05, |
| "distill_active_tokens": 11.825, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.07094771042466164, |
| "distill_pairs": 366.575, |
| "distill_skipped_no_correct_rate": 0.6, |
| "eligible_block_rate": 0.4, |
| "epoch": 0.0018739693168757184, |
| "grad_norm": 0.04520213603973389, |
| "grpo_loss": 0.004100967245176434, |
| "kl": 0.10252418667078018, |
| "learning_rate": 2e-06, |
| "loss": 0.075, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.46000000089406967, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.46000000089406967, |
| "step": 450, |
| "teacher_correct_rate": 0.4, |
| "teacher_step_mean": 12.8, |
| "teacher_step_min": 12.8, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.075, |
| "correct_distill_rate": 0.05, |
| "distill_active_tokens": 11.8, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.07750831246376037, |
| "distill_pairs": 365.8, |
| "distill_skipped_no_correct_rate": 0.6, |
| "eligible_block_rate": 0.4, |
| "epoch": 0.0019156130794729565, |
| "grad_norm": 1.541292428970337, |
| "grpo_loss": 0.0016718463855795562, |
| "kl": 0.041796160861849785, |
| "learning_rate": 2e-06, |
| "loss": 0.0792, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.46000000089406967, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.46000000089406967, |
| "step": 460, |
| "teacher_correct_rate": 0.4, |
| "teacher_step_mean": 12.8, |
| "teacher_step_min": 12.8, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.6, |
| "correct_distill_rate": 0.0125, |
| "distill_active_tokens": 3.025, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.01007867455482483, |
| "distill_pairs": 93.775, |
| "distill_skipped_no_correct_rate": 0.9, |
| "eligible_block_rate": 0.1, |
| "epoch": 0.001957256842070195, |
| "grad_norm": 0.029470475390553474, |
| "grpo_loss": 0.0019210274331271648, |
| "kl": 0.04802568592131138, |
| "learning_rate": 2e-06, |
| "loss": 0.012, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.1900000013411045, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.1900000013411045, |
| "step": 470, |
| "teacher_correct_rate": 0.1, |
| "teacher_step_mean": 3.2, |
| "teacher_step_min": 3.2, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.8875, |
| "correct_distill_rate": 0.05, |
| "distill_active_tokens": 11.725, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.06811503991484642, |
| "distill_pairs": 363.475, |
| "distill_skipped_no_correct_rate": 0.6, |
| "eligible_block_rate": 0.4, |
| "epoch": 0.001998900604667433, |
| "grad_norm": 0.01782740280032158, |
| "grpo_loss": 0.0018472394440323114, |
| "kl": 0.046180986985564235, |
| "learning_rate": 2e-06, |
| "loss": 0.07, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.46000000089406967, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.46000000089406967, |
| "step": 480, |
| "teacher_correct_rate": 0.4, |
| "teacher_step_mean": 12.8, |
| "teacher_step_min": 12.8, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.7, |
| "correct_distill_rate": 0.0, |
| "distill_active_tokens": 0.0, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.0, |
| "distill_pairs": 0.0, |
| "distill_skipped_no_correct_rate": 1.0, |
| "eligible_block_rate": 0.0, |
| "epoch": 0.002040544367264671, |
| "grad_norm": 0.01768697053194046, |
| "grpo_loss": 0.0018223325139842927, |
| "kl": 0.04555831328034401, |
| "learning_rate": 2e-06, |
| "loss": 0.0018, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.10000000149011612, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.10000000149011612, |
| "step": 490, |
| "teacher_correct_rate": 0.0, |
| "teacher_step_mean": 0.0, |
| "teacher_step_min": 0.0, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.7875, |
| "correct_distill_rate": 0.025, |
| "distill_active_tokens": 5.95, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.020546692609786987, |
| "distill_pairs": 184.45, |
| "distill_skipped_no_correct_rate": 0.8, |
| "eligible_block_rate": 0.2, |
| "epoch": 0.0020821881298619094, |
| "grad_norm": 0.02694624662399292, |
| "grpo_loss": 0.0015416789567098022, |
| "kl": 0.03854197524487972, |
| "learning_rate": 2e-06, |
| "loss": 0.0221, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.2800000011920929, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.2800000011920929, |
| "step": 500, |
| "teacher_correct_rate": 0.2, |
| "teacher_step_mean": 6.4, |
| "teacher_step_min": 6.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 119.275, |
| "correct_distill_rate": 0.025, |
| "distill_active_tokens": 6.0, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.023047928512096406, |
| "distill_pairs": 186.0, |
| "distill_skipped_no_correct_rate": 0.8, |
| "eligible_block_rate": 0.2, |
| "epoch": 0.0021238318924591475, |
| "grad_norm": 0.040342044085264206, |
| "grpo_loss": 0.0021890802076086403, |
| "kl": 0.05472700819373131, |
| "learning_rate": 2e-06, |
| "loss": 0.0252, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.2800000011920929, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.2800000011920929, |
| "step": 510, |
| "teacher_correct_rate": 0.2, |
| "teacher_step_mean": 6.4, |
| "teacher_step_min": 6.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.375, |
| "correct_distill_rate": 0.0, |
| "distill_active_tokens": 0.0, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.0, |
| "distill_pairs": 0.0, |
| "distill_skipped_no_correct_rate": 1.0, |
| "eligible_block_rate": 0.0, |
| "epoch": 0.0021654756550563855, |
| "grad_norm": 0.03764764964580536, |
| "grpo_loss": 0.0017872425843961537, |
| "kl": 0.04481751471757889, |
| "learning_rate": 2e-06, |
| "loss": 0.0018, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.09875000193715096, |
| "reward_std": 0.00353553369641304, |
| "rewards/countdown_reward_func": 0.09875000193715096, |
| "step": 520, |
| "teacher_correct_rate": 0.0, |
| "teacher_step_mean": 0.0, |
| "teacher_step_min": 0.0, |
| "zero_std_ratio": 0.9 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.45, |
| "correct_distill_rate": 0.0375, |
| "distill_active_tokens": 8.825, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.03324308693408966, |
| "distill_pairs": 273.575, |
| "distill_skipped_no_correct_rate": 0.7, |
| "eligible_block_rate": 0.3, |
| "epoch": 0.002207119417653624, |
| "grad_norm": 0.01843937672674656, |
| "grpo_loss": 0.002046559366863221, |
| "kl": 0.051163984835147856, |
| "learning_rate": 2e-06, |
| "loss": 0.0353, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.3700000010430813, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.3700000010430813, |
| "step": 530, |
| "teacher_correct_rate": 0.3, |
| "teacher_step_mean": 9.6, |
| "teacher_step_min": 9.6, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.8, |
| "correct_distill_rate": 0.05, |
| "distill_active_tokens": 12.0, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.03926486186683178, |
| "distill_pairs": 372.0, |
| "distill_skipped_no_correct_rate": 0.6, |
| "eligible_block_rate": 0.4, |
| "epoch": 0.002248763180250862, |
| "grad_norm": 0.04481740668416023, |
| "grpo_loss": 0.0018742202082648873, |
| "kl": 0.046855506114661695, |
| "learning_rate": 2e-06, |
| "loss": 0.0411, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.46000000089406967, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.46000000089406967, |
| "step": 540, |
| "teacher_correct_rate": 0.4, |
| "teacher_step_mean": 12.8, |
| "teacher_step_min": 12.8, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.1, |
| "correct_distill_rate": 0.025, |
| "distill_active_tokens": 5.9, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.022375880926847457, |
| "distill_pairs": 182.9, |
| "distill_skipped_no_correct_rate": 0.8, |
| "eligible_block_rate": 0.2, |
| "epoch": 0.0022904069428481004, |
| "grad_norm": 0.024466684088110924, |
| "grpo_loss": 0.0016954999533481895, |
| "kl": 0.04238750077784061, |
| "learning_rate": 2e-06, |
| "loss": 0.0241, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.2800000011920929, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.2800000011920929, |
| "step": 550, |
| "teacher_correct_rate": 0.2, |
| "teacher_step_mean": 6.4, |
| "teacher_step_min": 6.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 119.0, |
| "correct_distill_rate": 0.0125, |
| "distill_active_tokens": 3.025, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.012097294628620147, |
| "distill_pairs": 93.775, |
| "distill_skipped_no_correct_rate": 0.9, |
| "eligible_block_rate": 0.1, |
| "epoch": 0.0023320507054453384, |
| "grad_norm": 0.07262931019067764, |
| "grpo_loss": 0.0021076064556837084, |
| "kl": 0.052690163254737854, |
| "learning_rate": 2e-06, |
| "loss": 0.0142, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.1900000013411045, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.1900000013411045, |
| "step": 560, |
| "teacher_correct_rate": 0.1, |
| "teacher_step_mean": 3.2, |
| "teacher_step_min": 3.2, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.475, |
| "correct_distill_rate": 0.025, |
| "distill_active_tokens": 5.7, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.048476952314376834, |
| "distill_pairs": 176.7, |
| "distill_skipped_no_correct_rate": 0.8, |
| "eligible_block_rate": 0.2, |
| "epoch": 0.0023736944680425765, |
| "grad_norm": 0.024663519114255905, |
| "grpo_loss": 0.0018766604829579593, |
| "kl": 0.04691651277244091, |
| "learning_rate": 2e-06, |
| "loss": 0.0504, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.2800000011920929, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.2800000011920929, |
| "step": 570, |
| "teacher_correct_rate": 0.2, |
| "teacher_step_mean": 6.4, |
| "teacher_step_min": 6.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.7875, |
| "correct_distill_rate": 0.0375, |
| "distill_active_tokens": 8.925, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.0312973752617836, |
| "distill_pairs": 276.675, |
| "distill_skipped_no_correct_rate": 0.7, |
| "eligible_block_rate": 0.3, |
| "epoch": 0.002415338230639815, |
| "grad_norm": 0.0238138809800148, |
| "grpo_loss": 0.0018116210994776338, |
| "kl": 0.045290527679026124, |
| "learning_rate": 2e-06, |
| "loss": 0.0331, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.3700000010430813, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.3700000010430813, |
| "step": 580, |
| "teacher_correct_rate": 0.3, |
| "teacher_step_mean": 9.6, |
| "teacher_step_min": 9.6, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 119.5875, |
| "correct_distill_rate": 0.0375, |
| "distill_active_tokens": 8.9, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.028200805932283402, |
| "distill_pairs": 275.9, |
| "distill_skipped_no_correct_rate": 0.7, |
| "eligible_block_rate": 0.3, |
| "epoch": 0.002456981993237053, |
| "grad_norm": 0.04106259346008301, |
| "grpo_loss": 0.001833037578035146, |
| "kl": 0.04582594148814678, |
| "learning_rate": 2e-06, |
| "loss": 0.03, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.3700000010430813, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.3700000010430813, |
| "step": 590, |
| "teacher_correct_rate": 0.3, |
| "teacher_step_mean": 9.6, |
| "teacher_step_min": 9.6, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.85, |
| "correct_distill_rate": 0.0125, |
| "distill_active_tokens": 2.975, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.006647808104753494, |
| "distill_pairs": 92.225, |
| "distill_skipped_no_correct_rate": 0.9, |
| "eligible_block_rate": 0.1, |
| "epoch": 0.002498625755834291, |
| "grad_norm": 0.01524669211357832, |
| "grpo_loss": 0.0015321651473641396, |
| "kl": 0.038304129429161546, |
| "learning_rate": 2e-06, |
| "loss": 0.0082, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.1900000013411045, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.1900000013411045, |
| "step": 600, |
| "teacher_correct_rate": 0.1, |
| "teacher_step_mean": 3.2, |
| "teacher_step_min": 3.2, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 116.8, |
| "correct_distill_rate": 0.0, |
| "distill_active_tokens": 0.0, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.0, |
| "distill_pairs": 0.0, |
| "distill_skipped_no_correct_rate": 1.0, |
| "eligible_block_rate": 0.0, |
| "epoch": 0.0025402695184315294, |
| "grad_norm": 0.027340073138475418, |
| "grpo_loss": 0.001800740952603519, |
| "kl": 0.0450185252353549, |
| "learning_rate": 2e-06, |
| "loss": 0.0018, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.10000000149011612, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.10000000149011612, |
| "step": 610, |
| "teacher_correct_rate": 0.0, |
| "teacher_step_mean": 0.0, |
| "teacher_step_min": 0.0, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 119.4, |
| "correct_distill_rate": 0.0, |
| "distill_active_tokens": 0.0, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.0, |
| "distill_pairs": 0.0, |
| "distill_skipped_no_correct_rate": 1.0, |
| "eligible_block_rate": 0.0, |
| "epoch": 0.0025819132810287674, |
| "grad_norm": 0.024580704048275948, |
| "grpo_loss": 0.001650526630692184, |
| "kl": 0.041263166815042496, |
| "learning_rate": 2e-06, |
| "loss": 0.0017, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.10000000149011612, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.10000000149011612, |
| "step": 620, |
| "teacher_correct_rate": 0.0, |
| "teacher_step_mean": 0.0, |
| "teacher_step_min": 0.0, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 116.925, |
| "correct_distill_rate": 0.0375, |
| "distill_active_tokens": 8.825, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.03538615107536316, |
| "distill_pairs": 273.575, |
| "distill_skipped_no_correct_rate": 0.7, |
| "eligible_block_rate": 0.3, |
| "epoch": 0.002623557043626006, |
| "grad_norm": 0.9122187495231628, |
| "grpo_loss": 0.001667627552524209, |
| "kl": 0.04169069044291973, |
| "learning_rate": 2e-06, |
| "loss": 0.0371, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.3700000010430813, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.3700000010430813, |
| "step": 630, |
| "teacher_correct_rate": 0.3, |
| "teacher_step_mean": 9.6, |
| "teacher_step_min": 9.6, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 115.3, |
| "correct_distill_rate": 0.0125, |
| "distill_active_tokens": 2.875, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.026404333114624024, |
| "distill_pairs": 89.125, |
| "distill_skipped_no_correct_rate": 0.9, |
| "eligible_block_rate": 0.1, |
| "epoch": 0.002665200806223244, |
| "grad_norm": 0.03214803710579872, |
| "grpo_loss": 0.0018081217887811363, |
| "kl": 0.04520304724574089, |
| "learning_rate": 2e-06, |
| "loss": 0.0282, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.1900000013411045, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.1900000013411045, |
| "step": 640, |
| "teacher_correct_rate": 0.1, |
| "teacher_step_mean": 3.2, |
| "teacher_step_min": 3.2, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 115.875, |
| "correct_distill_rate": 0.05, |
| "distill_active_tokens": 11.5, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.04926532208919525, |
| "distill_pairs": 356.5, |
| "distill_skipped_no_correct_rate": 0.6, |
| "eligible_block_rate": 0.4, |
| "epoch": 0.002706844568820482, |
| "grad_norm": 0.0179664958268404, |
| "grpo_loss": 0.0015351969166658818, |
| "kl": 0.03837992306798697, |
| "learning_rate": 2e-06, |
| "loss": 0.0508, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.46000000089406967, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.46000000089406967, |
| "step": 650, |
| "teacher_correct_rate": 0.4, |
| "teacher_step_mean": 12.8, |
| "teacher_step_min": 12.8, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 119.5, |
| "correct_distill_rate": 0.05, |
| "distill_active_tokens": 11.925, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.05791616961359978, |
| "distill_pairs": 369.675, |
| "distill_skipped_no_correct_rate": 0.6, |
| "eligible_block_rate": 0.4, |
| "epoch": 0.0027484883314177204, |
| "grad_norm": 1.7235947847366333, |
| "grpo_loss": 0.001487891166470945, |
| "kl": 0.03719727993011475, |
| "learning_rate": 2e-06, |
| "loss": 0.0594, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.46000000089406967, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.46000000089406967, |
| "step": 660, |
| "teacher_correct_rate": 0.4, |
| "teacher_step_mean": 12.8, |
| "teacher_step_min": 12.8, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.6, |
| "correct_distill_rate": 0.0375, |
| "distill_active_tokens": 8.825, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.054658699035644534, |
| "distill_pairs": 273.575, |
| "distill_skipped_no_correct_rate": 0.7, |
| "eligible_block_rate": 0.3, |
| "epoch": 0.0027901320940149584, |
| "grad_norm": 0.018401429057121277, |
| "grpo_loss": 0.0016094543039798736, |
| "kl": 0.04023635908961296, |
| "learning_rate": 2e-06, |
| "loss": 0.0563, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.3700000010430813, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.3700000010430813, |
| "step": 670, |
| "teacher_correct_rate": 0.3, |
| "teacher_step_mean": 9.6, |
| "teacher_step_min": 9.6, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.3, |
| "correct_distill_rate": 0.0625, |
| "distill_active_tokens": 14.9, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.06135166436433792, |
| "distill_pairs": 461.9, |
| "distill_skipped_no_correct_rate": 0.5, |
| "eligible_block_rate": 0.5, |
| "epoch": 0.0028317758566121965, |
| "grad_norm": 4.852332592010498, |
| "grpo_loss": 0.001661125652026385, |
| "kl": 0.04152814298868179, |
| "learning_rate": 2e-06, |
| "loss": 0.063, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.5500000007450581, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.5500000007450581, |
| "step": 680, |
| "teacher_correct_rate": 0.5, |
| "teacher_step_mean": 16.0, |
| "teacher_step_min": 16.0, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.625, |
| "correct_distill_rate": 0.025, |
| "distill_active_tokens": 5.725, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.021581918746232987, |
| "distill_pairs": 177.475, |
| "distill_skipped_no_correct_rate": 0.8, |
| "eligible_block_rate": 0.2, |
| "epoch": 0.002873419619209435, |
| "grad_norm": 0.04149038717150688, |
| "grpo_loss": 0.0016454326454550029, |
| "kl": 0.041135816089808944, |
| "learning_rate": 2e-06, |
| "loss": 0.0232, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.2800000011920929, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.2800000011920929, |
| "step": 690, |
| "teacher_correct_rate": 0.2, |
| "teacher_step_mean": 6.4, |
| "teacher_step_min": 6.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.375, |
| "correct_distill_rate": 0.0375, |
| "distill_active_tokens": 8.925, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.037727190554142, |
| "distill_pairs": 276.675, |
| "distill_skipped_no_correct_rate": 0.7, |
| "eligible_block_rate": 0.3, |
| "epoch": 0.002915063381806673, |
| "grad_norm": 2.8539164066314697, |
| "grpo_loss": 0.0019391254638321698, |
| "kl": 0.0484781363978982, |
| "learning_rate": 2e-06, |
| "loss": 0.0397, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.3700000010430813, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.3700000010430813, |
| "step": 700, |
| "teacher_correct_rate": 0.3, |
| "teacher_step_mean": 9.6, |
| "teacher_step_min": 9.6, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 119.7875, |
| "correct_distill_rate": 0.05, |
| "distill_active_tokens": 12.0, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.03495294824242592, |
| "distill_pairs": 372.0, |
| "distill_skipped_no_correct_rate": 0.6, |
| "eligible_block_rate": 0.4, |
| "epoch": 0.0029567071444039114, |
| "grad_norm": 1.0490450859069824, |
| "grpo_loss": 0.001755313656758517, |
| "kl": 0.04388284385204315, |
| "learning_rate": 2e-06, |
| "loss": 0.0367, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.46000000089406967, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.46000000089406967, |
| "step": 710, |
| "teacher_correct_rate": 0.4, |
| "teacher_step_mean": 12.8, |
| "teacher_step_min": 12.8, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.825, |
| "correct_distill_rate": 0.025, |
| "distill_active_tokens": 5.825, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.030229657888412476, |
| "distill_pairs": 180.575, |
| "distill_skipped_no_correct_rate": 0.8, |
| "eligible_block_rate": 0.2, |
| "epoch": 0.0029983509070011494, |
| "grad_norm": 0.025108857080340385, |
| "grpo_loss": 0.0015562113956548273, |
| "kl": 0.03890528529882431, |
| "learning_rate": 2e-06, |
| "loss": 0.0318, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.2800000011920929, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.2800000011920929, |
| "step": 720, |
| "teacher_correct_rate": 0.2, |
| "teacher_step_mean": 6.4, |
| "teacher_step_min": 6.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.9, |
| "correct_distill_rate": 0.0375, |
| "distill_active_tokens": 9.025, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.022043896466493608, |
| "distill_pairs": 279.775, |
| "distill_skipped_no_correct_rate": 0.7, |
| "eligible_block_rate": 0.3, |
| "epoch": 0.0030399946695983874, |
| "grad_norm": 1.4314059019088745, |
| "grpo_loss": 0.0015278301434591413, |
| "kl": 0.038195754028856756, |
| "learning_rate": 2e-06, |
| "loss": 0.0236, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.3700000010430813, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.3700000010430813, |
| "step": 730, |
| "teacher_correct_rate": 0.3, |
| "teacher_step_mean": 9.6, |
| "teacher_step_min": 9.6, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 120.0, |
| "correct_distill_rate": 0.0125, |
| "distill_active_tokens": 2.975, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.01439603567123413, |
| "distill_pairs": 92.225, |
| "distill_skipped_no_correct_rate": 0.9, |
| "eligible_block_rate": 0.1, |
| "epoch": 0.003081638432195626, |
| "grad_norm": 0.046130210161209106, |
| "grpo_loss": 0.0016216201009228825, |
| "kl": 0.040540555492043495, |
| "learning_rate": 2e-06, |
| "loss": 0.016, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.1787500001490116, |
| "reward_std": 0.03181980550289154, |
| "rewards/countdown_reward_func": 0.1787500001490116, |
| "step": 740, |
| "teacher_correct_rate": 0.1, |
| "teacher_step_mean": 3.2, |
| "teacher_step_min": 3.2, |
| "zero_std_ratio": 0.9 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.9, |
| "correct_distill_rate": 0.05, |
| "distill_active_tokens": 11.825, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.053345417231321336, |
| "distill_pairs": 366.575, |
| "distill_skipped_no_correct_rate": 0.6, |
| "eligible_block_rate": 0.4, |
| "epoch": 0.003123282194792864, |
| "grad_norm": 4.38554573059082, |
| "grpo_loss": 0.001586603233590722, |
| "kl": 0.039665082842111586, |
| "learning_rate": 2e-06, |
| "loss": 0.0549, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.46000000089406967, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.46000000089406967, |
| "step": 750, |
| "teacher_correct_rate": 0.4, |
| "teacher_step_mean": 12.8, |
| "teacher_step_min": 12.8, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 116.5125, |
| "correct_distill_rate": 0.0125, |
| "distill_active_tokens": 3.025, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.008107659220695496, |
| "distill_pairs": 93.775, |
| "distill_skipped_no_correct_rate": 0.9, |
| "eligible_block_rate": 0.1, |
| "epoch": 0.003164925957390102, |
| "grad_norm": 2.5829641819000244, |
| "grpo_loss": 0.001549235754646361, |
| "kl": 0.03873089589178562, |
| "learning_rate": 2e-06, |
| "loss": 0.0097, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.1900000013411045, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.1900000013411045, |
| "step": 760, |
| "teacher_correct_rate": 0.1, |
| "teacher_step_mean": 3.2, |
| "teacher_step_min": 3.2, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 119.1, |
| "correct_distill_rate": 0.0125, |
| "distill_active_tokens": 2.975, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.006333620101213455, |
| "distill_pairs": 92.225, |
| "distill_skipped_no_correct_rate": 0.9, |
| "eligible_block_rate": 0.1, |
| "epoch": 0.0032065697199873404, |
| "grad_norm": 0.02771187573671341, |
| "grpo_loss": 0.0016519801341928543, |
| "kl": 0.041299504786729814, |
| "learning_rate": 2e-06, |
| "loss": 0.008, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.1900000013411045, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.1900000013411045, |
| "step": 770, |
| "teacher_correct_rate": 0.1, |
| "teacher_step_mean": 3.2, |
| "teacher_step_min": 3.2, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 119.4, |
| "correct_distill_rate": 0.0625, |
| "distill_active_tokens": 14.925, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.04773559235036373, |
| "distill_pairs": 462.675, |
| "distill_skipped_no_correct_rate": 0.5, |
| "eligible_block_rate": 0.5, |
| "epoch": 0.0032482134825845784, |
| "grad_norm": 0.031229112297296524, |
| "grpo_loss": 0.0015942462487146258, |
| "kl": 0.03985615726560354, |
| "learning_rate": 2e-06, |
| "loss": 0.0493, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.5500000007450581, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.5500000007450581, |
| "step": 780, |
| "teacher_correct_rate": 0.5, |
| "teacher_step_mean": 16.0, |
| "teacher_step_min": 16.0, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.3, |
| "correct_distill_rate": 0.025, |
| "distill_active_tokens": 6.05, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.01303027682006359, |
| "distill_pairs": 187.55, |
| "distill_skipped_no_correct_rate": 0.8, |
| "eligible_block_rate": 0.2, |
| "epoch": 0.0032898572451818165, |
| "grad_norm": 0.027933601289987564, |
| "grpo_loss": 0.0018914732965640724, |
| "kl": 0.047286834195256235, |
| "learning_rate": 2e-06, |
| "loss": 0.0149, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.2800000011920929, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.2800000011920929, |
| "step": 790, |
| "teacher_correct_rate": 0.2, |
| "teacher_step_mean": 6.4, |
| "teacher_step_min": 6.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.7, |
| "correct_distill_rate": 0.025, |
| "distill_active_tokens": 5.85, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.01853492483496666, |
| "distill_pairs": 181.35, |
| "distill_skipped_no_correct_rate": 0.8, |
| "eligible_block_rate": 0.2, |
| "epoch": 0.003331501007779055, |
| "grad_norm": 0.02773735113441944, |
| "grpo_loss": 0.0016249989392235875, |
| "kl": 0.04062497429549694, |
| "learning_rate": 2e-06, |
| "loss": 0.0202, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.2800000011920929, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.2800000011920929, |
| "step": 800, |
| "teacher_correct_rate": 0.2, |
| "teacher_step_mean": 6.4, |
| "teacher_step_min": 6.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 116.4, |
| "correct_distill_rate": 0.05, |
| "distill_active_tokens": 11.7, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.06442554593086243, |
| "distill_pairs": 362.7, |
| "distill_skipped_no_correct_rate": 0.6, |
| "eligible_block_rate": 0.4, |
| "epoch": 0.003373144770376293, |
| "grad_norm": 0.05563170090317726, |
| "grpo_loss": 0.0015791050856932998, |
| "kl": 0.039477628841996196, |
| "learning_rate": 2e-06, |
| "loss": 0.066, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.46000000089406967, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.46000000089406967, |
| "step": 810, |
| "teacher_correct_rate": 0.4, |
| "teacher_step_mean": 12.8, |
| "teacher_step_min": 12.8, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 119.3, |
| "correct_distill_rate": 0.0375, |
| "distill_active_tokens": 8.825, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.05658463910222054, |
| "distill_pairs": 273.575, |
| "distill_skipped_no_correct_rate": 0.7, |
| "eligible_block_rate": 0.3, |
| "epoch": 0.0034147885329735314, |
| "grad_norm": 0.02054646797478199, |
| "grpo_loss": 0.0016433336539193987, |
| "kl": 0.041083342395722866, |
| "learning_rate": 2e-06, |
| "loss": 0.0582, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.3700000010430813, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.3700000010430813, |
| "step": 820, |
| "teacher_correct_rate": 0.3, |
| "teacher_step_mean": 9.6, |
| "teacher_step_min": 9.6, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.6, |
| "correct_distill_rate": 0.0375, |
| "distill_active_tokens": 8.95, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.029812106117606162, |
| "distill_pairs": 277.45, |
| "distill_skipped_no_correct_rate": 0.7, |
| "eligible_block_rate": 0.3, |
| "epoch": 0.0034564322955707694, |
| "grad_norm": 0.016008736565709114, |
| "grpo_loss": 0.0014529126987326891, |
| "kl": 0.036322818510234355, |
| "learning_rate": 2e-06, |
| "loss": 0.0313, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.3700000010430813, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.3700000010430813, |
| "step": 830, |
| "teacher_correct_rate": 0.3, |
| "teacher_step_mean": 9.6, |
| "teacher_step_min": 9.6, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.8, |
| "correct_distill_rate": 0.0, |
| "distill_active_tokens": 0.0, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.0, |
| "distill_pairs": 0.0, |
| "distill_skipped_no_correct_rate": 1.0, |
| "eligible_block_rate": 0.0, |
| "epoch": 0.0034980760581680074, |
| "grad_norm": 0.05104987323284149, |
| "grpo_loss": 0.0016627219505608083, |
| "kl": 0.041568050533533095, |
| "learning_rate": 2e-06, |
| "loss": 0.0017, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.10000000149011612, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.10000000149011612, |
| "step": 840, |
| "teacher_correct_rate": 0.0, |
| "teacher_step_mean": 0.0, |
| "teacher_step_min": 0.0, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.8125, |
| "correct_distill_rate": 0.0125, |
| "distill_active_tokens": 3.025, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.012090180814266206, |
| "distill_pairs": 93.775, |
| "distill_skipped_no_correct_rate": 0.9, |
| "eligible_block_rate": 0.1, |
| "epoch": 0.003539719820765246, |
| "grad_norm": 0.03538968786597252, |
| "grpo_loss": 0.0014836499351076782, |
| "kl": 0.03709125053137541, |
| "learning_rate": 2e-06, |
| "loss": 0.0136, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.1900000013411045, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.1900000013411045, |
| "step": 850, |
| "teacher_correct_rate": 0.1, |
| "teacher_step_mean": 3.2, |
| "teacher_step_min": 3.2, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 116.9, |
| "correct_distill_rate": 0.0375, |
| "distill_active_tokens": 8.75, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.03502166718244552, |
| "distill_pairs": 271.25, |
| "distill_skipped_no_correct_rate": 0.7, |
| "eligible_block_rate": 0.3, |
| "epoch": 0.003581363583362484, |
| "grad_norm": 3.450979232788086, |
| "grpo_loss": 0.0014392685843631625, |
| "kl": 0.03598171547055244, |
| "learning_rate": 2e-06, |
| "loss": 0.0365, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.3700000010430813, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.3700000010430813, |
| "step": 860, |
| "teacher_correct_rate": 0.3, |
| "teacher_step_mean": 9.6, |
| "teacher_step_min": 9.6, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 115.275, |
| "correct_distill_rate": 0.025, |
| "distill_active_tokens": 5.825, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.025030899792909622, |
| "distill_pairs": 180.575, |
| "distill_skipped_no_correct_rate": 0.8, |
| "eligible_block_rate": 0.2, |
| "epoch": 0.003623007345959722, |
| "grad_norm": 0.027195338159799576, |
| "grpo_loss": 0.0015214605256915093, |
| "kl": 0.038036514259874823, |
| "learning_rate": 2e-06, |
| "loss": 0.0266, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.2800000011920929, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.2800000011920929, |
| "step": 870, |
| "teacher_correct_rate": 0.2, |
| "teacher_step_mean": 6.4, |
| "teacher_step_min": 6.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 119.3, |
| "correct_distill_rate": 0.025, |
| "distill_active_tokens": 5.9, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.017651048302650452, |
| "distill_pairs": 182.9, |
| "distill_skipped_no_correct_rate": 0.8, |
| "eligible_block_rate": 0.2, |
| "epoch": 0.0036646511085569604, |
| "grad_norm": 2.248119831085205, |
| "grpo_loss": 0.0016151666874065994, |
| "kl": 0.04037916790693998, |
| "learning_rate": 2e-06, |
| "loss": 0.0193, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.2800000011920929, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.2800000011920929, |
| "step": 880, |
| "teacher_correct_rate": 0.2, |
| "teacher_step_mean": 6.4, |
| "teacher_step_min": 6.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.7, |
| "correct_distill_rate": 0.025, |
| "distill_active_tokens": 6.0, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.012588660418987273, |
| "distill_pairs": 186.0, |
| "distill_skipped_no_correct_rate": 0.8, |
| "eligible_block_rate": 0.2, |
| "epoch": 0.0037062948711541984, |
| "grad_norm": 0.03247544541954994, |
| "grpo_loss": 0.001700232131406665, |
| "kl": 0.04250580370426178, |
| "learning_rate": 2e-06, |
| "loss": 0.0143, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.2800000011920929, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.2800000011920929, |
| "step": 890, |
| "teacher_correct_rate": 0.2, |
| "teacher_step_mean": 6.4, |
| "teacher_step_min": 6.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.5, |
| "correct_distill_rate": 0.025, |
| "distill_active_tokens": 6.05, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.013561930507421494, |
| "distill_pairs": 187.55, |
| "distill_skipped_no_correct_rate": 0.8, |
| "eligible_block_rate": 0.2, |
| "epoch": 0.003747938633751437, |
| "grad_norm": 0.02233463153243065, |
| "grpo_loss": 0.0013939386466518045, |
| "kl": 0.03484846781939268, |
| "learning_rate": 2e-06, |
| "loss": 0.015, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.2800000011920929, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.2800000011920929, |
| "step": 900, |
| "teacher_correct_rate": 0.2, |
| "teacher_step_mean": 6.4, |
| "teacher_step_min": 6.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 119.3875, |
| "correct_distill_rate": 0.025, |
| "distill_active_tokens": 5.95, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.02756602391600609, |
| "distill_pairs": 184.45, |
| "distill_skipped_no_correct_rate": 0.8, |
| "eligible_block_rate": 0.2, |
| "epoch": 0.003789582396348675, |
| "grad_norm": 0.02028791606426239, |
| "grpo_loss": 0.0015327666769735515, |
| "kl": 0.03831916730850935, |
| "learning_rate": 2e-06, |
| "loss": 0.0291, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.2800000011920929, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.2800000011920929, |
| "step": 910, |
| "teacher_correct_rate": 0.2, |
| "teacher_step_mean": 6.4, |
| "teacher_step_min": 6.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.4, |
| "correct_distill_rate": 0.025, |
| "distill_active_tokens": 6.05, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.014009146764874459, |
| "distill_pairs": 187.55, |
| "distill_skipped_no_correct_rate": 0.8, |
| "eligible_block_rate": 0.2, |
| "epoch": 0.003831226158945913, |
| "grad_norm": 1.03496253490448, |
| "grpo_loss": 0.0015499247354455292, |
| "kl": 0.03874811865389347, |
| "learning_rate": 2e-06, |
| "loss": 0.0156, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.2800000011920929, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.2800000011920929, |
| "step": 920, |
| "teacher_correct_rate": 0.2, |
| "teacher_step_mean": 6.4, |
| "teacher_step_min": 6.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.4, |
| "correct_distill_rate": 0.0, |
| "distill_active_tokens": 0.0, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.0, |
| "distill_pairs": 0.0, |
| "distill_skipped_no_correct_rate": 1.0, |
| "eligible_block_rate": 0.0, |
| "epoch": 0.0038728699215431514, |
| "grad_norm": 0.024818778038024902, |
| "grpo_loss": 0.0013495881110429763, |
| "kl": 0.033739703521132466, |
| "learning_rate": 2e-06, |
| "loss": 0.0013, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.10000000149011612, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.10000000149011612, |
| "step": 930, |
| "teacher_correct_rate": 0.0, |
| "teacher_step_mean": 0.0, |
| "teacher_step_min": 0.0, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 119.1, |
| "correct_distill_rate": 0.0375, |
| "distill_active_tokens": 8.95, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.0384965255856514, |
| "distill_pairs": 277.45, |
| "distill_skipped_no_correct_rate": 0.7, |
| "eligible_block_rate": 0.3, |
| "epoch": 0.00391451368414039, |
| "grad_norm": 1.1330617666244507, |
| "grpo_loss": 0.0014331423211842776, |
| "kl": 0.03582855872809887, |
| "learning_rate": 2e-06, |
| "loss": 0.0399, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.3700000010430813, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.3700000010430813, |
| "step": 940, |
| "teacher_correct_rate": 0.3, |
| "teacher_step_mean": 9.6, |
| "teacher_step_min": 9.6, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.35, |
| "correct_distill_rate": 0.025, |
| "distill_active_tokens": 5.975, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.03094845525920391, |
| "distill_pairs": 185.225, |
| "distill_skipped_no_correct_rate": 0.8, |
| "eligible_block_rate": 0.2, |
| "epoch": 0.003956157446737628, |
| "grad_norm": 0.03506700322031975, |
| "grpo_loss": 0.0013764206785708665, |
| "kl": 0.03441051747649908, |
| "learning_rate": 2e-06, |
| "loss": 0.0323, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.2800000011920929, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.2800000011920929, |
| "step": 950, |
| "teacher_correct_rate": 0.2, |
| "teacher_step_mean": 6.4, |
| "teacher_step_min": 6.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 119.1125, |
| "correct_distill_rate": 0.0125, |
| "distill_active_tokens": 2.85, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.023532481491565706, |
| "distill_pairs": 88.35, |
| "distill_skipped_no_correct_rate": 0.9, |
| "eligible_block_rate": 0.1, |
| "epoch": 0.003997801209334866, |
| "grad_norm": 0.02786865271627903, |
| "grpo_loss": 0.0014480484649538993, |
| "kl": 0.03620121255517006, |
| "learning_rate": 2e-06, |
| "loss": 0.025, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.1900000013411045, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.1900000013411045, |
| "step": 960, |
| "teacher_correct_rate": 0.1, |
| "teacher_step_mean": 3.2, |
| "teacher_step_min": 3.2, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.575, |
| "correct_distill_rate": 0.0125, |
| "distill_active_tokens": 2.95, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.014441192150115967, |
| "distill_pairs": 91.45, |
| "distill_skipped_no_correct_rate": 0.9, |
| "eligible_block_rate": 0.1, |
| "epoch": 0.004039444971932104, |
| "grad_norm": 0.04904806613922119, |
| "grpo_loss": 0.0012988316360861064, |
| "kl": 0.03247079141438007, |
| "learning_rate": 2e-06, |
| "loss": 0.0157, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.1900000013411045, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.1900000013411045, |
| "step": 970, |
| "teacher_correct_rate": 0.1, |
| "teacher_step_mean": 3.2, |
| "teacher_step_min": 3.2, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.4125, |
| "correct_distill_rate": 0.0875, |
| "distill_active_tokens": 20.85, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.06841219961643219, |
| "distill_pairs": 646.35, |
| "distill_skipped_no_correct_rate": 0.3, |
| "eligible_block_rate": 0.7, |
| "epoch": 0.004081088734529342, |
| "grad_norm": 3.678745985031128, |
| "grpo_loss": 0.0012415537785273045, |
| "kl": 0.031038844585418703, |
| "learning_rate": 2e-06, |
| "loss": 0.0697, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.7300000004470348, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.7300000004470348, |
| "step": 980, |
| "teacher_correct_rate": 0.7, |
| "teacher_step_mean": 22.4, |
| "teacher_step_min": 22.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.075, |
| "correct_distill_rate": 0.025, |
| "distill_active_tokens": 5.85, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.01982411667704582, |
| "distill_pairs": 181.35, |
| "distill_skipped_no_correct_rate": 0.8, |
| "eligible_block_rate": 0.2, |
| "epoch": 0.00412273249712658, |
| "grad_norm": 0.03635792061686516, |
| "grpo_loss": 0.0013503476278856398, |
| "kl": 0.03375869132578373, |
| "learning_rate": 2e-06, |
| "loss": 0.0212, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.2800000011920929, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.2800000011920929, |
| "step": 990, |
| "teacher_correct_rate": 0.2, |
| "teacher_step_mean": 6.4, |
| "teacher_step_min": 6.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 119.8, |
| "correct_distill_rate": 0.025, |
| "distill_active_tokens": 5.8, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.02109660059213638, |
| "distill_pairs": 179.8, |
| "distill_skipped_no_correct_rate": 0.8, |
| "eligible_block_rate": 0.2, |
| "epoch": 0.004164376259723819, |
| "grad_norm": 0.018494676798582077, |
| "grpo_loss": 0.001708054111804813, |
| "kl": 0.04270135462284088, |
| "learning_rate": 2e-06, |
| "loss": 0.0228, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.2800000011920929, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.2800000011920929, |
| "step": 1000, |
| "teacher_correct_rate": 0.2, |
| "teacher_step_mean": 6.4, |
| "teacher_step_min": 6.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.425, |
| "correct_distill_rate": 0.0625, |
| "distill_active_tokens": 15.0, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.07608207762241363, |
| "distill_pairs": 465.0, |
| "distill_skipped_no_correct_rate": 0.5, |
| "eligible_block_rate": 0.5, |
| "epoch": 0.004206020022321057, |
| "grad_norm": 0.02504514902830124, |
| "grpo_loss": 0.0015504946815781296, |
| "kl": 0.03876236826181412, |
| "learning_rate": 2e-06, |
| "loss": 0.0776, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.5500000007450581, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.5500000007450581, |
| "step": 1010, |
| "teacher_correct_rate": 0.5, |
| "teacher_step_mean": 16.0, |
| "teacher_step_min": 16.0, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.8, |
| "correct_distill_rate": 0.0125, |
| "distill_active_tokens": 2.975, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.0050029665231704715, |
| "distill_pairs": 92.225, |
| "distill_skipped_no_correct_rate": 0.9, |
| "eligible_block_rate": 0.1, |
| "epoch": 0.004247663784918295, |
| "grad_norm": 0.02285120077431202, |
| "grpo_loss": 0.0015343933133408426, |
| "kl": 0.038359833508729936, |
| "learning_rate": 2e-06, |
| "loss": 0.0065, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.1900000013411045, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.1900000013411045, |
| "step": 1020, |
| "teacher_correct_rate": 0.1, |
| "teacher_step_mean": 3.2, |
| "teacher_step_min": 3.2, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.8, |
| "correct_distill_rate": 0.0, |
| "distill_active_tokens": 0.0, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.0, |
| "distill_pairs": 0.0, |
| "distill_skipped_no_correct_rate": 1.0, |
| "eligible_block_rate": 0.0, |
| "epoch": 0.004289307547515533, |
| "grad_norm": 0.0297338105738163, |
| "grpo_loss": 0.0014691868796944617, |
| "kl": 0.03672967404127121, |
| "learning_rate": 2e-06, |
| "loss": 0.0015, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.10000000149011612, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.10000000149011612, |
| "step": 1030, |
| "teacher_correct_rate": 0.0, |
| "teacher_step_mean": 0.0, |
| "teacher_step_min": 0.0, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.9, |
| "correct_distill_rate": 0.0375, |
| "distill_active_tokens": 8.825, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.02233266532421112, |
| "distill_pairs": 273.575, |
| "distill_skipped_no_correct_rate": 0.7, |
| "eligible_block_rate": 0.3, |
| "epoch": 0.004330951310112771, |
| "grad_norm": 2.297927141189575, |
| "grpo_loss": 0.0014402684988453983, |
| "kl": 0.03600671347230673, |
| "learning_rate": 2e-06, |
| "loss": 0.0238, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.3700000010430813, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.3700000010430813, |
| "step": 1040, |
| "teacher_correct_rate": 0.3, |
| "teacher_step_mean": 9.6, |
| "teacher_step_min": 9.6, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.4625, |
| "correct_distill_rate": 0.05, |
| "distill_active_tokens": 11.775, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.04610191583633423, |
| "distill_pairs": 365.025, |
| "distill_skipped_no_correct_rate": 0.6, |
| "eligible_block_rate": 0.4, |
| "epoch": 0.00437259507271001, |
| "grad_norm": 2.0214428901672363, |
| "grpo_loss": 0.0012075707025360316, |
| "kl": 0.030189268477261066, |
| "learning_rate": 2e-06, |
| "loss": 0.0473, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.46000000089406967, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.46000000089406967, |
| "step": 1050, |
| "teacher_correct_rate": 0.4, |
| "teacher_step_mean": 12.8, |
| "teacher_step_min": 12.8, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 116.6625, |
| "correct_distill_rate": 0.025, |
| "distill_active_tokens": 5.825, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.023518009483814238, |
| "distill_pairs": 180.575, |
| "distill_skipped_no_correct_rate": 0.8, |
| "eligible_block_rate": 0.2, |
| "epoch": 0.004414238835307248, |
| "grad_norm": 0.027771051973104477, |
| "grpo_loss": 0.0013593727722764014, |
| "kl": 0.033984319493174556, |
| "learning_rate": 2e-06, |
| "loss": 0.0249, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.2800000011920929, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.2800000011920929, |
| "step": 1060, |
| "teacher_correct_rate": 0.2, |
| "teacher_step_mean": 6.4, |
| "teacher_step_min": 6.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.625, |
| "correct_distill_rate": 0.0, |
| "distill_active_tokens": 0.0, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.0, |
| "distill_pairs": 0.0, |
| "distill_skipped_no_correct_rate": 1.0, |
| "eligible_block_rate": 0.0, |
| "epoch": 0.004455882597904486, |
| "grad_norm": 0.018284277990460396, |
| "grpo_loss": 0.0013712895917706192, |
| "kl": 0.03428224083036184, |
| "learning_rate": 2e-06, |
| "loss": 0.0014, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.10000000149011612, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.10000000149011612, |
| "step": 1070, |
| "teacher_correct_rate": 0.0, |
| "teacher_step_mean": 0.0, |
| "teacher_step_min": 0.0, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.9375, |
| "correct_distill_rate": 0.025, |
| "distill_active_tokens": 5.975, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.029092486202716827, |
| "distill_pairs": 185.225, |
| "distill_skipped_no_correct_rate": 0.8, |
| "eligible_block_rate": 0.2, |
| "epoch": 0.004497526360501724, |
| "grad_norm": 0.07809923589229584, |
| "grpo_loss": 0.0012806750717572867, |
| "kl": 0.0320168761536479, |
| "learning_rate": 2e-06, |
| "loss": 0.0304, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.2800000011920929, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.2800000011920929, |
| "step": 1080, |
| "teacher_correct_rate": 0.2, |
| "teacher_step_mean": 6.4, |
| "teacher_step_min": 6.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.425, |
| "correct_distill_rate": 0.0125, |
| "distill_active_tokens": 3.025, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.023815013468265533, |
| "distill_pairs": 93.775, |
| "distill_skipped_no_correct_rate": 0.9, |
| "eligible_block_rate": 0.1, |
| "epoch": 0.004539170123098962, |
| "grad_norm": 0.017430124804377556, |
| "grpo_loss": 0.0013626973144710065, |
| "kl": 0.03406743332743645, |
| "learning_rate": 2e-06, |
| "loss": 0.0252, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.1900000013411045, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.1900000013411045, |
| "step": 1090, |
| "teacher_correct_rate": 0.1, |
| "teacher_step_mean": 3.2, |
| "teacher_step_min": 3.2, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 119.2125, |
| "correct_distill_rate": 0.0375, |
| "distill_active_tokens": 8.975, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.05073585957288742, |
| "distill_pairs": 278.225, |
| "distill_skipped_no_correct_rate": 0.7, |
| "eligible_block_rate": 0.3, |
| "epoch": 0.004580813885696201, |
| "grad_norm": 0.027992134913802147, |
| "grpo_loss": 0.0013235102873295546, |
| "kl": 0.033087757602334025, |
| "learning_rate": 2e-06, |
| "loss": 0.0521, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.3700000010430813, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.3700000010430813, |
| "step": 1100, |
| "teacher_correct_rate": 0.3, |
| "teacher_step_mean": 9.6, |
| "teacher_step_min": 9.6, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.9375, |
| "correct_distill_rate": 0.0375, |
| "distill_active_tokens": 8.8, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.03319371715188026, |
| "distill_pairs": 272.8, |
| "distill_skipped_no_correct_rate": 0.7, |
| "eligible_block_rate": 0.3, |
| "epoch": 0.004622457648293439, |
| "grad_norm": 1.1471742391586304, |
| "grpo_loss": 0.0013235443155281245, |
| "kl": 0.033088609017431735, |
| "learning_rate": 2e-06, |
| "loss": 0.0345, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.3700000010430813, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.3700000010430813, |
| "step": 1110, |
| "teacher_correct_rate": 0.3, |
| "teacher_step_mean": 9.6, |
| "teacher_step_min": 9.6, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.5625, |
| "correct_distill_rate": 0.025, |
| "distill_active_tokens": 6.0, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.022497203201055527, |
| "distill_pairs": 186.0, |
| "distill_skipped_no_correct_rate": 0.8, |
| "eligible_block_rate": 0.2, |
| "epoch": 0.004664101410890677, |
| "grad_norm": 5.557712554931641, |
| "grpo_loss": 0.0013524662121199072, |
| "kl": 0.033811656199395655, |
| "learning_rate": 2e-06, |
| "loss": 0.0238, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.2800000011920929, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.2800000011920929, |
| "step": 1120, |
| "teacher_correct_rate": 0.2, |
| "teacher_step_mean": 6.4, |
| "teacher_step_min": 6.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.75, |
| "correct_distill_rate": 0.0125, |
| "distill_active_tokens": 2.9, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.015088455379009246, |
| "distill_pairs": 89.9, |
| "distill_skipped_no_correct_rate": 0.9, |
| "eligible_block_rate": 0.1, |
| "epoch": 0.004705745173487915, |
| "grad_norm": 0.016883619129657745, |
| "grpo_loss": 0.0013344469596631825, |
| "kl": 0.033361174166202545, |
| "learning_rate": 2e-06, |
| "loss": 0.0164, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.1900000013411045, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.1900000013411045, |
| "step": 1130, |
| "teacher_correct_rate": 0.1, |
| "teacher_step_mean": 3.2, |
| "teacher_step_min": 3.2, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.025, |
| "correct_distill_rate": 0.0625, |
| "distill_active_tokens": 14.825, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.05899478681385517, |
| "distill_pairs": 459.575, |
| "distill_skipped_no_correct_rate": 0.5, |
| "eligible_block_rate": 0.5, |
| "epoch": 0.004747388936085153, |
| "grad_norm": 7.67894172668457, |
| "grpo_loss": 0.00178499590838328, |
| "kl": 0.04462493751198053, |
| "learning_rate": 2e-06, |
| "loss": 0.0608, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.5387499995529652, |
| "reward_std": 0.03181980550289154, |
| "rewards/countdown_reward_func": 0.5387499995529652, |
| "step": 1140, |
| "teacher_correct_rate": 0.5, |
| "teacher_step_mean": 16.0, |
| "teacher_step_min": 16.0, |
| "zero_std_ratio": 0.9 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.75, |
| "correct_distill_rate": 0.025, |
| "distill_active_tokens": 6.0, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.013463569432497024, |
| "distill_pairs": 186.0, |
| "distill_skipped_no_correct_rate": 0.8, |
| "eligible_block_rate": 0.2, |
| "epoch": 0.004789032698682391, |
| "grad_norm": 0.04088316485285759, |
| "grpo_loss": 0.0014945810195058583, |
| "kl": 0.03736452590674162, |
| "learning_rate": 2e-06, |
| "loss": 0.015, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.2800000011920929, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.2800000011920929, |
| "step": 1150, |
| "teacher_correct_rate": 0.2, |
| "teacher_step_mean": 6.4, |
| "teacher_step_min": 6.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.5, |
| "correct_distill_rate": 0.075, |
| "distill_active_tokens": 17.525, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.048146254196763036, |
| "distill_pairs": 543.275, |
| "distill_skipped_no_correct_rate": 0.4, |
| "eligible_block_rate": 0.6, |
| "epoch": 0.00483067646127963, |
| "grad_norm": 0.037850912660360336, |
| "grpo_loss": 0.0014006719691678881, |
| "kl": 0.03501679990440607, |
| "learning_rate": 2e-06, |
| "loss": 0.0495, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.6400000005960464, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.6400000005960464, |
| "step": 1160, |
| "teacher_correct_rate": 0.6, |
| "teacher_step_mean": 19.2, |
| "teacher_step_min": 19.2, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 118.4125, |
| "correct_distill_rate": 0.0125, |
| "distill_active_tokens": 3.0, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.012423863261938095, |
| "distill_pairs": 93.0, |
| "distill_skipped_no_correct_rate": 0.9, |
| "eligible_block_rate": 0.1, |
| "epoch": 0.004872320223876868, |
| "grad_norm": 1.806624174118042, |
| "grpo_loss": 0.0013075167778879404, |
| "kl": 0.032687919400632384, |
| "learning_rate": 2e-06, |
| "loss": 0.0137, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.1900000013411045, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.1900000013411045, |
| "step": 1170, |
| "teacher_correct_rate": 0.1, |
| "teacher_step_mean": 3.2, |
| "teacher_step_min": 3.2, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 117.3875, |
| "correct_distill_rate": 0.025, |
| "distill_active_tokens": 6.0, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.038350760191679004, |
| "distill_pairs": 186.0, |
| "distill_skipped_no_correct_rate": 0.8, |
| "eligible_block_rate": 0.2, |
| "epoch": 0.004913963986474106, |
| "grad_norm": 0.020910833030939102, |
| "grpo_loss": 0.001398240774869919, |
| "kl": 0.034956020675599574, |
| "learning_rate": 2e-06, |
| "loss": 0.0397, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.2800000011920929, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.2800000011920929, |
| "step": 1180, |
| "teacher_correct_rate": 0.2, |
| "teacher_step_mean": 6.4, |
| "teacher_step_min": 6.4, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 119.225, |
| "correct_distill_rate": 0.0125, |
| "distill_active_tokens": 3.025, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.010749422758817673, |
| "distill_pairs": 93.775, |
| "distill_skipped_no_correct_rate": 0.9, |
| "eligible_block_rate": 0.1, |
| "epoch": 0.004955607749071344, |
| "grad_norm": 0.03307757154107094, |
| "grpo_loss": 0.0012086151575203985, |
| "kl": 0.03021537996828556, |
| "learning_rate": 2e-06, |
| "loss": 0.012, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.1900000013411045, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.1900000013411045, |
| "step": 1190, |
| "teacher_correct_rate": 0.1, |
| "teacher_step_mean": 3.2, |
| "teacher_step_min": 3.2, |
| "zero_std_ratio": 1.0 |
| }, |
| { |
| "clip_ratio": 0.0, |
| "completion_length": 120.075, |
| "correct_distill_rate": 0.0125, |
| "distill_active_tokens": 3.025, |
| "distill_lambda": 1.0, |
| "distill_loss": 0.005179817602038383, |
| "distill_pairs": 93.775, |
| "distill_skipped_no_correct_rate": 0.9, |
| "eligible_block_rate": 0.1, |
| "epoch": 0.004997251511668582, |
| "grad_norm": 0.027619360014796257, |
| "grpo_loss": 0.0011856672586873173, |
| "kl": 0.029641682840883733, |
| "learning_rate": 2e-06, |
| "loss": 0.0064, |
| "productive_jump_progress_mean": 0.0, |
| "productive_jump_score_mean": 0.0, |
| "productive_jump_selected_rate": 0.0, |
| "productive_jump_span_mean": 0.0, |
| "productive_jump_tail_entropy_mean": 0.0, |
| "reward": 0.1900000013411045, |
| "reward_std": 0.0, |
| "rewards/countdown_reward_func": 0.1900000013411045, |
| "step": 1200, |
| "teacher_correct_rate": 0.1, |
| "teacher_step_mean": 3.2, |
| "teacher_step_min": 3.2, |
| "zero_std_ratio": 1.0 |
| } |
| ], |
| "logging_steps": 10, |
| "max_steps": 1200, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 1, |
| "save_steps": 200, |
| "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": 2, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
