Safety Alignment on Multiple-choice Safety Evaluation benchmark leaderboard · SOTA2 Research

Safety Alignment on Multiple-choice Safety Evaluation

84.48Refusal RateINNSTEER

Evaluation Results

Method	Links
INNSTEER2026.06		84.48	47.92
ODE2026.06		69.06	49.38
TF2026.06		66.48	50.5
ITI2026.06		65.7	51.22
LAcT2026.06		65.14	50.92
HPR2026.06		64.29	52.1
CAA2026.06		59.96	53.21
REC2026.06		58.14	53.74
Base2026.06		57.85	55.23
PoE2026.06		56.79	54.4
PoD2026.06		56.62	54.22