List your company

Mathematical Reasoning on MATH (ASRr, ACCc)

0.892ASR (Rate)CFT
-0.0286080.2103960.44940.688404Apr 12, 2026

Evaluation results

MethodLinks
CFTDefense scheme=CFT, Backbone=GPT-OSS-20B2026.040.89291.8
No defenseDefense scheme=No defense, Backbone=GPT-OSS-20B2026.040.88691.4
ONIONDefense scheme=ONION, Backbone=GPT-OSS-20B2026.040.63882.3
CoSDefense scheme=CoS, Backbone=GPT-OSS-20B2026.040.60386.5
ShuffleDefense scheme=Shuffle, Backbone=GPT-OSS-20B2026.040.34650.3
Shuffle++Defense scheme=Shuffle++, Backbone=GPT-OSS-20B2026.040.05620.6
Critical-CoTDefense scheme=Ours, Backbone=GPT-OSS-20B2026.040.006889.3