{
  "seeds": [0, 1, 2],
  "grid_size": 9,
  "horizon": 28,
  "train_episodes": 500,
  "evaluation_episodes": 200,
  "gamma": 0.98,
  "learning_rate": 0.001,
  "batch_size": 128,
  "her_k": 4,
  "her_strategies": ["final", "future"],
  "reward_variants": ["sparse", "dense"],
  "ood_slip_probability": 0.35,
  "domain_randomization_range": [0.0, 0.35],
  "bootstrap_resamples": 10000
}