Few-shot task expansion on ALFWorld unseen domains 5-shot
71.88Success RateTMoW
Evaluation results
| Method | Links | ||
|---|---|---|---|
| TMoWshot=52026.01 | 71.88 | 25.73 | |
| SayCanPayshot=52026.01 | 49.34 | 37.55 | |
| LLM+FTshot=52026.01 | 48.86 | 35.14 | |
| FLAREshot=52026.01 | 11.46 | 44.4 | |
| LLM-Plannershot=52026.01 | 7.39 | 46.79 |