{
  "model": "Qwen/Qwen3.5-0.8B",
  "revision": "2fc06364715b967f1860aea9cf38778875588b17",
  "lora_rank": 8,
  "max_prompt_tokens": 8192,
  "stages": [
    "head_goon"
  ],
  "ppo_epochs": 2,
  "decisions_per_stage": 512,
  "minimum_episodes_per_stage": 8,
  "gradient_batch_size": 64,
  "actor_learning_rate": 3e-06,
  "critic_learning_rate": 0.001,
  "critic_epochs": 200,
  "clip_epsilon": 0.2,
  "entropy_coefficient": 0.01,
  "target_kl": 0.02,
  "gamma_per_frame": 0.9997,
  "lambda_per_20_frames": 0.99,
  "max_grad_norm": 1.0,
  "temperature": 1.0,
  "minimum_action_frames": 6,
  "max_episode_frames": 5400,
  "evaluation_episode_wall_seconds": 90,
  "critic_scale": 100.0,
  "critic_normalization": "layernorm",
  "critic_init_seed": 17,
  "boss_hit_reward": 20.0,
  "bounce_multiplier": 0.5,
  "iterations": 16,
  "inference_batch_size": 1,
  "gradient_microbatch": 1,
  "seed": 101620261
}