List your company

Mathematical Reasoning Verification on AIME (Brier, AUROC)

0.0283Brier ScoreMSV64
0.0260360.0413180.05660.071882Mar 3, 2026

Evaluation results

MethodLinks
MSV64Backbone=DeepSeek-R1-Distill-Llama-8B, Number of sequences=642026.030.028399.23
MSV16Backbone=DeepSeek-R1-Distill-Llama-8B, Number of sequences=162026.030.029899.08
MSVBackbone=Qwen3-1.7B, Mode=thinking mode, N=642026.030.031798.78
MSVBackbone=Qwen3-1.7B, Mode=thinking mode, N=162026.030.04498.25
MSVBackbone=Qwen3-1.7B, Mode=thinking mode, N=42026.030.059596.96
ProbeBackbone=DeepSeek-R1-Distill-Llama-8B2026.030.066896.68
MSV1Backbone=DeepSeek-R1-Distill-Llama-8B, Number of sequences=12026.030.07797.46
MSVBackbone=Qwen3-1.7B, Mode=thinking mode, N=12026.030.07795.7
ProbeBackbone=Qwen3-1.7B, Mode=thinking mode2026.030.084995.3