Tool Retrieval on TOOLRET Zero-Shot Custom
45.8nDCG@10TOOLQP
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| TOOLQPMethod Category=Fine-tuning, Training Data Size=10k, Backbone Model=1.7B2026.01 | 45.8 | 43 | |
| HyDE/PRFMethod Category=Prompting, Backbone Model=Qwen3-30B-A3B-Instruct2026.01 | 41.6 | 26.7 | |
| Q2P/SFTMethod Category=Fine-tuning, Training Data Size=10k, Backbone Model=1.5B/1.7B models2026.01 | 39.7 | 35.6 | |
| bge-gemmaMethod Category=Re-ranking with cross-encoder2026.01 | 39.4 | 36.2 | |
| gte-Qwen/ContrastiveFTMethod Category=Fine-tuning, Training Data Size=10k, Backbone Model=1.5B/1.7B models2026.01 | 39.2 | 35.5 | |
| Q2E/ZSMethod Category=Prompting, Backbone Model=Qwen3-30B-A3B-Instruct2026.01 | 38.7 | 35.7 | |
| Q2D/PRFMethod Category=Prompting, Backbone Model=Qwen3-30B-A3B-Instruct2026.01 | 37 | 33.7 | |
| Re-InvokeMethod Category=Prompting, Backbone Model=Qwen3-30B-A3B-Instruct2026.01 | 36.8 | 31.7 | |
| Base Retriever (gte-Qwen)Method Category=Base Retriever2026.01 | 35.6 | 32.4 | |
| Q2E/PRFMethod Category=Prompting, Backbone Model=Qwen3-30B-A3B-Instruct2026.01 | 35.3 | 33.3 | |
| Q2D/ZSMethod Category=Prompting, Backbone Model=Qwen3-30B-A3B-Instruct2026.01 | 33.3 | 33.5 | |
| D2QMethod Category=Prompting, Backbone Model=Qwen3-30B-A3B-Instruct2026.01 | 33.3 | 30.8 | |
| bge-m3Method Category=Re-ranking with cross-encoder2026.01 | 32.9 | 30.2 | |
| TOOLQP-FORMATMethod Category=Fine-tuning, Training Data Size=10k, Backbone Model=1.7B2026.01 | 32.2 | 30.1 | |
| HyDE/ZSMethod Category=Prompting, Backbone Model=Qwen3-30B-A3B-Instruct2026.01 | 26.4 | 29.7 |