List your company
ResearchFormerly WizWand

Mathematical Reasoning on MATH (Accuracy Mean, Std, and CI)

66.69Accuracy (Mean)KTO
46.274851.574956.87562.1751Jun 8, 2026
Updated 4mo ago

Evaluation results

MethodLinks
KTOModel (Backbone)=Qwen3-4B, Decoding Strategy=sampling-based inference, Temperature=0.7, Top-p=0.952026.0666.690.2766.35
RePO_detModel (Backbone)=Qwen3-4B, Decoding Strategy=sampling-based inference, Temperature=0.7, Top-p=0.952026.0663.950.5563.26
RePOModel (Backbone)=Qwen3-4B, Decoding Strategy=sampling-based inference, Temperature=0.7, Top-p=0.952026.0663.820.4363.28
RPOModel (Backbone)=Qwen3-4B, Decoding Strategy=sampling-based inference, Temperature=0.7, Top-p=0.952026.0660.540.4260.02
IPOModel (Backbone)=Qwen3-4B, Decoding Strategy=sampling-based inference, Temperature=0.7, Top-p=0.952026.0659.60.6758.77
TDPOModel (Backbone)=Qwen3-4B, Decoding Strategy=sampling-based inference, Temperature=0.7, Top-p=0.952026.0659.20.1958.96
DPOModel (Backbone)=Qwen3-4B, Decoding Strategy=sampling-based inference, Temperature=0.7, Top-p=0.952026.0656.310.2356.02
KTOModel (Backbone)=Qwen3-1.7B, Decoding Strategy=sampling-based inference, Temperature=0.7, Top-p=0.952026.0653.980.3753.52
RePOModel (Backbone)=Qwen3-1.7B, Decoding Strategy=sampling-based inference, Temperature=0.7, Top-p=0.952026.0652.590.4552.03
RePO_detModel (Backbone)=Qwen3-1.7B, Decoding Strategy=sampling-based inference, Temperature=0.7, Top-p=0.952026.0652.250.6751.41
IPOModel (Backbone)=Qwen3-1.7B, Decoding Strategy=sampling-based inference, Temperature=0.7, Top-p=0.952026.0650.720.7249.82
DPOModel (Backbone)=Qwen3-1.7B, Decoding Strategy=sampling-based inference, Temperature=0.7, Top-p=0.952026.0648.920.3648.47
RPOModel (Backbone)=Qwen3-1.7B, Decoding Strategy=sampling-based inference, Temperature=0.7, Top-p=0.952026.0648.020.4847.42
TDPOModel (Backbone)=Qwen3-1.7B, Decoding Strategy=sampling-based inference, Temperature=0.7, Top-p=0.952026.0647.060.2146.79