Tool-use Reasoning on ToolQA
47.97AccuracyFull Text
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Full TextBackbone Model=Qwen3-8B2026.06 | 47.97 | 1,010 | |
| SKIM (Adaptive)Backbone Model=Qwen3-8B2026.06 | 46.92 | 754 | |
| LLMLingua-2 (L)Backbone Model=Qwen3-8B2026.06 | 45.45 | 769 | |
| SKIM (Fix-512)Backbone Model=Qwen3-8B2026.06 | 43.92 | 512 | |
| LLMLingua-2 (M)Backbone Model=Qwen3-8B2026.06 | 43.08 | 573 | |
| LLMLingua-2 (S)Backbone Model=Qwen3-8B2026.06 | 37.06 | 294 | |
| SKIM (Fix-256)Backbone Model=Qwen3-8B2026.06 | 35.66 | 256 | |
| NaiveBackbone Model=Qwen3-8B2026.06 | 28.04 | 0 | |
| SKIM (Fix-512)Backbone Model=Phi-42026.06 | 26.78 | 512 | |
| SKIM (Fix-256)Backbone Model=Phi-42026.06 | 23.08 | 256 | |
| SKIM (Adaptive)Backbone Model=Phi-42026.06 | 17.27 | 705 | |
| LLMLingua-2 (L)Backbone Model=Phi-42026.06 | 15.24 | 755 | |
| NaiveBackbone Model=Phi-42026.06 | 14.83 | 0 | |
| LLMLingua-2 (M)Backbone Model=Phi-42026.06 | 14.48 | 561 | |
| Full TextBackbone Model=Phi-42026.06 | 9.79 | 996 | |
| ICAEBackbone Model=Qwen3-8B2026.06 | 7.13 | 256 | |
| LLMLingua-2 (S)Backbone Model=Phi-42026.06 | 4.48 | 287 | |
| 500xCompressorBackbone Model=Qwen3-8B2026.06 | 0 | 256 | |
| 500xCompressorBackbone Model=Phi-42026.06 | 0 | 256 | |
| ICAEBackbone Model=Phi-42026.06 | 0 | 256 |