List your company

Translator-call mode selection on PolyMath High

73.74Macro F1LUAR
44.474452.072259.6767.2678Jun 1, 2026

Evaluation results

MethodLinks
LUARBackbone=Qwen3-4B2026.0673.74
LUARBackbone=Qwen3-8B2026.0669.74
BOUNDARY-SFTBackbone=Qwen3-4B2026.0666.85
ST(qr)Backbone=Qwen3-8B2026.0660.73
BOUNDARY-SFTBackbone=Qwen3-8B2026.0660.54
ST(q)Backbone=Qwen3-8B2026.0660.41
ST(qr)Backbone=Qwen3-4B2026.0658.24
ST(q)Backbone=Qwen3-4B2026.0653.58
SELF-ASSESSMENTBackbone=Qwen3-4B2026.0652.93
SELF-ASSESSMENTBackbone=Qwen3-8B2026.0648.6
NATIVE-TOOL-USEBackbone=Qwen3-4B2026.0648.54
NATIVE-TOOL-USEBackbone=Qwen3-8B2026.0645.6