List your company

Math Reasoning on Olympiad (Accuracy)

57.2Accuracy (Olympiad)LUFFY
13.93625.16836.447.632Jul 2, 2025

Evaluation results

MethodLinks
LUFFYBase Model=Qwen2.5-Math-7B2025.0757.2
Prefix-RFTBase Model=Qwen2.5-Math-7B2025.0755.7
ReLIFTBase Model=Qwen2.5-Math-7B2025.0752.5
UFTBase Model=Qwen2.5-Math-7B2025.0751.6
SFT+RFTBase Model=Qwen2.5-Math-7B2025.0750.4
RFTBase Model=Qwen2.5-Math-7B2025.0746.8
SFTBase Model=Qwen2.5-Math-7B2025.0743.7
Oat-ZeroBase Model=Qwen2.5-Math-7B2025.0743.4
RL w/ SFT LossBase Model=Qwen2.5-Math-7B2025.0739.4
SimpleRL-ZeroBase Model=Qwen2.5-Math-7B2025.0734.7
Qwen2.5-Math-7B2025.0715.6