Reinforcement Learning on MiniGrid BlockedUnlockPickup
80Success Rate (SR)Hand-Crafted
Evaluation Results
| Method | Links | |
|---|---|---|
| Hand-CraftedTraining Episodes=3,0002026.05 | 80 | |
| LLM One-ShotTraining Episodes=3,0002026.05 | 70 | |
| LLM IterativeTraining Episodes=3,0002026.05 | 60 | |
| No ShapingTraining Episodes=3,0002026.05 | 0 | |
| RNDTraining Episodes=3,0002026.05 | 0 |