Task Completion on ALFWorld Seen
8.6Average StepsEAGLET
Evaluation Results
| Method | Links | |
|---|---|---|
| EAGLETExecutor Agent=GPT-52025.10 | 8.6 | |
| EAGLETExecutor Agent=GPT-4.12025.10 | 9.4 | |
| MPOExecutor Agent=GPT-52025.10 | 9.7 | |
| w/o GuidanceExecutor Agent=GPT-52025.10 | 10.4 | |
| MPOExecutor Agent=GPT-4.12025.10 | 10.6 | |
| w/o GuidanceExecutor Agent=GPT-4.12025.10 | 10.8 |