Task Routing on OlympusBench chain-of-action setting
18Exact Difference (ED)Olympus
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Olympus2024.12 | 18 | 91.82 | 92.75 | 91.98 | |
| HuggingGPTbackbone=GPT-4o2024.12 | 35 | 75.03 | 60.23 | 61.25 | |
| HuggingGPTbackbone=GPT-4o mini2024.12 | 45 | 65.14 | 48.51 | 53.14 |