List your company

Translator-call Mode Selection on PolyMath Low

87.64Macro F1 ScoreLUAR
39.238451.804264.3776.9358Jun 1, 2026

Evaluation results

MethodLinks
LUARBackbone=Qwen3-4B2026.0687.64
ST(qr)Backbone=Qwen3-8B2026.0682.92
ST(qr)Backbone=Qwen3-4B2026.0679.26
LUARBackbone=Qwen3-8B2026.0678.32
BOUNDARY-SFTBackbone=Qwen3-4B2026.0673.14
BOUNDARY-SFTBackbone=Qwen3-8B2026.0669.2
ST(q)Backbone=Qwen3-8B2026.0664.01
ST(q)Backbone=Qwen3-4B2026.0663.25
NATIVE-TOOL-USEBackbone=Qwen3-8B2026.0658.41
NATIVE-TOOL-USEBackbone=Qwen3-4B2026.0657.6
SELF-ASSESSMENTBackbone=Qwen3-8B2026.0643.48
SELF-ASSESSMENTBackbone=Qwen3-4B2026.0641.1