List your company

Safety discovery in text-based scenarios on Compliance Rev. Absent Sup.

39R MetricEPO-Safe
26.5229.763336.24Apr 25, 2026

Evaluation results

MethodLinks
EPO-SafeModel=Claude Sonnet 4.62026.0439390
Reward-OnlyModel=Claude Sonnet 4.62026.0439390
EPO-SafeModel=Gemini 3 Flash2026.0439390
Reward-OnlyModel=Gemini 3 Flash2026.0439390
CoTModel=Claude Sonnet 4.62026.04271515
StaticModel=Claude Sonnet 4.62026.04271515
CoTModel=Gemini 3 Flash2026.04271515
StaticModel=Gemini 3 Flash2026.04271515