Translator-call mode selection on PolyMath Medium
75.22Macro F1LUAR
Evaluation Results
| Method | Links | |
|---|---|---|
| LUARBackbone=Qwen3-4B2026.06 | 75.22 | |
| ST(qr)Backbone=Qwen3-8B2026.06 | 69.19 | |
| BOUNDARY-SFTBackbone=Qwen3-8B2026.06 | 65.9 | |
| LUARBackbone=Qwen3-8B2026.06 | 65.61 | |
| ST(qr)Backbone=Qwen3-4B2026.06 | 64.59 | |
| BOUNDARY-SFTBackbone=Qwen3-4B2026.06 | 62.22 | |
| SELF-ASSESSMENTBackbone=Qwen3-4B2026.06 | 56.44 | |
| ST(q)Backbone=Qwen3-8B2026.06 | 56.18 | |
| SELF-ASSESSMENTBackbone=Qwen3-8B2026.06 | 54.28 | |
| ST(q)Backbone=Qwen3-4B2026.06 | 52.59 | |
| NATIVE-TOOL-USEBackbone=Qwen3-4B2026.06 | 49.61 | |
| NATIVE-TOOL-USEBackbone=Qwen3-8B2026.06 | 39.23 |