Web Task on WebShop
69.2Average RewardCo-Evolving Agents
Evaluation Results
| Method | Links | |
|---|---|---|
| Co-Evolving AgentsModel=Llama-2-13B-chat2025.11 | 69.2 | |
| Co-Evolving AgentsAdaptation=Fine-tuning, Backbone=Llama-2-7B-Chat2025.11 | 66.7 | |
| ETOModel=Llama-2-13B-chat2025.11 | 65 | |
| Llama-2-7B-Chat + PPOAdaptation=Fine-tuning, Backbone=Llama-2-7B-Chat2025.11 | 64.2 | |
| GPT-4Adaptation=In-context2025.11 | 63.2 | |
| Llama-2-7B-Chat + ETOAdaptation=Fine-tuning, Backbone=Llama-2-7B-Chat2025.11 | 63 | |
| GPT-3.5-TurboAdaptation=In-context2025.11 | 62.4 | |
| Llama-2-7B-Chat + RFTAdaptation=Fine-tuning, Backbone=Llama-2-7B-Chat2025.11 | 61.3 | |
| SFTModel=Llama-2-13B-chat2025.11 | 59.3 | |
| Llama-2-7B-Chat + SFTAdaptation=Fine-tuning, Backbone=Llama-2-7B-Chat2025.11 | 59.2 | |
| LUMOS-IAll-13Bshot=2-/3-shot2023.11 | 50.3 | |
| Claude-instant2023.11 | 49.7 | |
| LUMOS-IQA-13Bshot=2-/3-shot2023.11 | 47.3 | |
| OpenChat-v3.2-13B2023.11 | 46.9 | |
| LUMOS-IWeb-13Bshot=2-/3-shot2023.11 | 46.2 | |
| LUMOS-IMath-13Bshot=2-/3-shot2023.11 | 45.7 | |
| LUMOS-IMM-13Bshot=2-/3-shot2023.11 | 43.8 | |
| Vicuna-v1.5-13B2023.11 | 41.7 | |
| Vicuna-v1.3-33B2023.11 | 23.9 | |
| ChatGLM22023.11 | 19.4 | |
| Vicuna-v1.1-13B2023.11 | 12.6 | |
| WizardLM-30B2023.11 | 10.6 | |
| Koala-13B2023.11 | 6 | |
| Baichuan-13B-chat2023.11 | 5.7 |