ResearchBenchmarksTool Use on BFCL live and Irrelevance DetectionFollow86.2BFCL live ScoreEKSFT69.87274.11178.3582.589May 28, 2026Evaluation ResultsMethodMethodLinksBFCL live ScoreIrrelevance Detection ScoreEKSFTModel=Qwen3-4B, Traini...Model=Qwen3-4B, Training Dataset=AceBench2026.0586.2100SFTModel=Qwen3-4B, Traini...Model=Qwen3-4B, Training Dataset=AceBench2026.0584.489.7BaseModel=Qwen3-4B, Traini...Model=Qwen3-4B, Training Dataset=AceBench2026.0582.475.9DFTModel=Qwen3-4B, Traini...Model=Qwen3-4B, Training Dataset=AceBench2026.0570.593.5