Tool Retrieval on ToolBench Multi-domain (I2)
89.76NDCG@1ToolWeaver
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| ToolWeaverImplementation source=This Paper2026.01 | 89.76 | 89.7 | 91.8 | |
| ToolGenImplementation source=Re-implemented2026.01 | 84 | 85.65 | 89.02 | |
| ToolGenImplementation source=Original2026.01 | 83.46 | 86.24 | 88.84 | |
| ToolRetrieverImplementation source=Original2026.01 | 64.54 | 57.91 | 63.61 | |
| ToolRetrieverImplementation source=Re-implemented2026.01 | 63 | 66.38 | 72.72 | |
| EmbSimImplementation source=Re-implemented2026.01 | 46 | 39.58 | 43.05 | |
| EmbSimImplementation source=Original2026.01 | 40.84 | 36.67 | 39.55 | |
| BM25Implementation source=Re-implemented2026.01 | 20 | 21.92 | 23.46 | |
| BM25Implementation source=Original2026.01 | 18.29 | 20.74 | 22.18 |