List your company

Multi-task Reasoning on Average Out-of-domain

49.57Accuracy (OOD)DFT
4.361216.098127.83539.5719Oct 31, 2025

Evaluation results

MethodLinks
DFTModel=Qwen3 32B2025.1049.57
VCOREModel=Qwen3 32B2025.1045.93
RandomModel=Qwen3 32B2025.1044.83
SFTModel=Qwen3 32B2025.1043.78
iw-SFTModel=Qwen3 32B2025.1042.64
OriginalModel=Qwen3 32B2025.1042.52
DFTModel=Qwen3 8B2025.1037.29
iw-SFTModel=Qwen3 8B2025.1035.34
RandomModel=Qwen3 8B2025.1035.26
VCOREModel=Qwen3 8B2025.1035.26
SFTModel=Qwen3 8B2025.1035.07
VCOREModel=Qwen3 4B2025.1032.87
DFTModel=Qwen3 4B2025.1032.49
RandomModel=Qwen3 4B2025.1030.62
iw-SFTModel=Qwen3 4B2025.1030.43
SFTModel=Qwen3 4B2025.1030.34
OriginalModel=Qwen3 8B2025.1029.58
OriginalModel=Qwen3 4B2025.1026.12
OriginalModel=LLaMA3.1 8B Instruct2025.1020.18
VCOREModel=LLaMA3.1 8B Instruct2025.109.59
RandomModel=LLaMA3.1 8B Instruct2025.109.07
SFTModel=LLaMA3.1 8B Instruct2025.107.71
DFTModel=LLaMA3.1 8B Instruct2025.107.17
iw-SFTModel=LLaMA3.1 8B Instruct2025.106.1