List your company

Reasoning on MATH (test) (AUROC/p-value)

0.749AUROCTRACED
0.746920.747460.7480.74854Mar 11, 2026

Evaluation results

MethodLinks
TRACEDBase Model=DeepSeek-R12026.030.7490.045
LR ProbeBase Model=DeepSeek-R1, Method Status=Runner-up Baseline2026.030.747—