Interactive Task Solving on WebShop
68.36Average Task PerformanceATG
Evaluation Results
| Method | Links | |
|---|---|---|
| ATGBackbone=Llama-3-8B2026.07 | 68.36 | |
| ATGBackbone=Gemma-7B2026.07 | 64.93 | |
| ReActBackbone=GPT-42026.07 | 64.34 | |
| ATGBackbone=Mistral-7B2026.07 | 62.75 | |
| ReActBackbone=GPT-3.5-Turbo2026.07 | 43.57 | |
| PoGBackbone=Llama-3-8B2026.07 | 31.66 | |
| ToTBackbone=Llama-3-8B2026.07 | 27.4 | |
| CAMELBackbone=Llama-3-8B2026.07 | 24.85 | |
| PoGBackbone=Mistral-7B2026.07 | 24.18 | |
| ReflexionBackbone=Llama-3-8B2026.07 | 22.73 | |
| ToTBackbone=Mistral-7B2026.07 | 20.31 | |
| ReActBackbone=Llama-3-8B2026.07 | 19.32 | |
| CAMELBackbone=Mistral-7B2026.07 | 18.05 | |
| ReflexionBackbone=Mistral-7B2026.07 | 16.64 | |
| ReActBackbone=Mistral-7B2026.07 | 14.63 | |
| PoGBackbone=Gemma-7B2026.07 | 12.56 | |
| ToTBackbone=Gemma-7B2026.07 | 10.84 | |
| CAMELBackbone=Gemma-7B2026.07 | 9.48 | |
| ReflexionBackbone=Gemma-7B2026.07 | 7.71 | |
| ReActBackbone=Gemma-7B2026.07 | 5.93 |