List your company

Multiple Choice Question Answering on MMLU Pro NoMath (unseen)

62MMLU Pro (NoMath) ScoreQwen2.5-Math-7B-PRM800K
54.51256.45658.460.344May 21, 2025

Evaluation results

MethodLinks
Qwen2.5-Math-7B-PRM800KK=72025.0562
Qwen2.5-Math-PRM-7BK=72025.0560.4
FOVER-Qwen2.5-7B-PRMK=72025.0560.4
Qwen2.5-7B-Skywork-PRMK=72025.0558.8
Qwen 2.5 7BK=72025.0558
RLHFlow-Llama3.1-8B-DK=72025.0557.6
Llama 3.1 8BK=72025.0556.8
RLHFlow-Llama3.1-8B-MK=72025.0556.8
FOVER-Llama3.1-8B-PRMK=72025.0554.8