Embodied Task Planning on ALFWorld (134 evaluation environments)
10.42Cumulative ReturnLWM-Planner
Evaluation Results
| Method | Links | |
|---|---|---|
| LWM-Plannerruns=single-seed, confidence_interval=95%2025.06 | 10.42 | |
| ReAct + FECruns=single-seed, confidence_interval=95%2025.06 | 8.53 | |
| ReActruns=single-seed, confidence_interval=95%2025.06 | 5 | |
| Reflexionruns=single-seed, confidence_interval=95%2025.06 | 4.36 | |
| Randomruns=single-seed, confidence_interval=95%2025.06 | 0 |