List your company

Tool-use Reasoning on ToolQA

47.97AccuracyFull Text
-1.918811.033123.98536.9369Jun 10, 2026

Evaluation results

MethodLinks
Full TextBackbone Model=Qwen3-8B2026.0647.971,010
SKIM (Adaptive)Backbone Model=Qwen3-8B2026.0646.92754
LLMLingua-2 (L)Backbone Model=Qwen3-8B2026.0645.45769
SKIM (Fix-512)Backbone Model=Qwen3-8B2026.0643.92512
LLMLingua-2 (M)Backbone Model=Qwen3-8B2026.0643.08573
LLMLingua-2 (S)Backbone Model=Qwen3-8B2026.0637.06294
SKIM (Fix-256)Backbone Model=Qwen3-8B2026.0635.66256
NaiveBackbone Model=Qwen3-8B2026.0628.040
SKIM (Fix-512)Backbone Model=Phi-42026.0626.78512
SKIM (Fix-256)Backbone Model=Phi-42026.0623.08256
SKIM (Adaptive)Backbone Model=Phi-42026.0617.27705
LLMLingua-2 (L)Backbone Model=Phi-42026.0615.24755
NaiveBackbone Model=Phi-42026.0614.830
LLMLingua-2 (M)Backbone Model=Phi-42026.0614.48561
Full TextBackbone Model=Phi-42026.069.79996
ICAEBackbone Model=Qwen3-8B2026.067.13256
LLMLingua-2 (S)Backbone Model=Phi-42026.064.48287
500xCompressorBackbone Model=Qwen3-8B2026.060256
500xCompressorBackbone Model=Phi-42026.060256
ICAEBackbone Model=Phi-42026.060256