List your company

Tool Invocation Refusal on SABEval (D2)

5.46Tool Invocation Rate (TIR)CAA
2.689221.392140.09558.7979Apr 13, 2026

Evaluation results

MethodLinks
CAAModel=Watt-Tool2026.045.46
CAAModel=ToolACE-2.52026.045.76
CAAModel=Qwen3-8B2026.047.27
CAAModel=Qwen3-14B2026.049.31
CAAModel=Qwen3-4B2026.0420.29
PromptModel=Watt-Tool2026.0422.34
BaseModel=Watt-Tool2026.0423.31
PromptModel=Qwen3-14B2026.0443.56
PromptModel=Qwen3-8B2026.0448.73
PromptModel=Qwen3-4B2026.0456.83
PromptModel=ToolACE-2.52026.0462.63
BaseModel=ToolACE-2.52026.0463.02
BaseModel=Qwen3-8B2026.0466.29
BaseModel=Qwen3-4B2026.0473.41
BaseModel=Qwen3-14B2026.0474.73