Tool Retrieval on TOOLRET In-Domain (Avg)
63.1nDCG@10TOOLQP-FORMAT
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| TOOLQP-FORMATMethod Category=Fine-tuning, Training Data Size=10k, Backbone Model=1.7B2026.01 | 63.1 | 66.1 | |
| gte-Qwen/ContrastiveFTMethod Category=Fine-tuning, Training Data Size=10k, Backbone Model=1.5B/1.7B models2026.01 | 57.2 | 61.4 | |
| TOOLQPMethod Category=Fine-tuning, Training Data Size=10k, Backbone Model=1.7B2026.01 | 53.9 | 59.9 | |
| bge-gemmaMethod Category=Re-ranking with cross-encoder2026.01 | 53 | 54 | |
| Q2D/ZSMethod Category=Prompting, Backbone Model=Qwen3-30B-A3B-Instruct2026.01 | 52 | 56.8 | |
| D2QMethod Category=Prompting, Backbone Model=Qwen3-30B-A3B-Instruct2026.01 | 51.5 | 54 | |
| Re-InvokeMethod Category=Prompting, Backbone Model=Qwen3-30B-A3B-Instruct2026.01 | 51.5 | 56.8 | |
| bge-m3Method Category=Re-ranking with cross-encoder2026.01 | 49.4 | 52.5 | |
| HyDE/ZSMethod Category=Prompting, Backbone Model=Qwen3-30B-A3B-Instruct2026.01 | 47.9 | 52.3 | |
| Q2D/PRFMethod Category=Prompting, Backbone Model=Qwen3-30B-A3B-Instruct2026.01 | 46.9 | 50.2 | |
| Q2P/SFTMethod Category=Fine-tuning, Training Data Size=10k, Backbone Model=1.5B/1.7B models2026.01 | 46.6 | 51.2 | |
| Q2E/PRFMethod Category=Prompting, Backbone Model=Qwen3-30B-A3B-Instruct2026.01 | 45 | 49.4 | |
| Q2E/ZSMethod Category=Prompting, Backbone Model=Qwen3-30B-A3B-Instruct2026.01 | 44.6 | 49 | |
| HyDE/PRFMethod Category=Prompting, Backbone Model=Qwen3-30B-A3B-Instruct2026.01 | 43.5 | 43.5 | |
| Base Retriever (gte-Qwen)Method Category=Base Retriever2026.01 | 43.3 | 47.8 |