Tool Selection on TaskBench-MM
20.7F1 ScoreOLIVIA
Evaluation Results
| Method | Links | |
|---|---|---|
| OLIVIABackbone=Mistral-7B-v0.12026.05 | 20.7 | |
| OLIVIABackbone=Qwen3-4B2026.05 | 17.8 | |
| CLINBackbone=Qwen3-4B2026.05 | 17 | |
| ReActBackbone=Qwen3-4B2026.05 | 16.2 | |
| ReActBackbone=Mistral-7B-v0.12026.05 | 14.6 | |
| BM25Backbone=Qwen3-4B2026.05 | 12.7 | |
| BM25Backbone=Mistral-7B-v0.12026.05 | 12.7 | |
| CLINBackbone=Mistral-7B-v0.12026.05 | 12.7 | |
| BanditBackbone=Qwen3-4B2026.05 | 12.6 | |
| BanditBackbone=Mistral-7B-v0.12026.05 | 12.6 | |
| CoTBackbone=Mistral-7B-v0.12026.05 | 6.4 | |
| CoTBackbone=Qwen3-4B2026.05 | 3.1 |