List your company

Mathematical Reasoning on Minerva Math Dataset (avg.@8)

43.66Average Accuracy @8SFT
25.158429.961734.76539.5683Aug 25, 2025

Evaluation results

MethodLinks
SFTBackbone=Qwen2.5-7B-Instruct2025.0843.66
PSFTwarm-upBackbone=Qwen2.5-7B-Instruct2025.0843.38
PSFTBackbone=Qwen2.5-7B-Instruct2025.0843.33
SFT-KLBackbone=Qwen2.5-7B-Instruct2025.0842.19
BaseBackbone=Qwen2.5-7B-Instruct2025.0840.53
PSFTwarm-upBackbone=Llama3.1-8B-Instruct2025.0833.64
PSFTBackbone=Llama3.1-8B-Instruct2025.0832.4
SFTBackbone=Llama3.1-8B-Instruct2025.0832.17
SFT-KLBackbone=Llama3.1-8B-Instruct2025.0826.75
BaseBackbone=Llama3.1-8B-Instruct2025.0825.87