Friday6 Nov 2026
Frontier Model Forum releases Version 1.0 of cyber jailbreak rating framework
Developed with OpenAI, Anthropic, Google DeepMind, and Microsoft on 5 November 2026, the rubric scores cyber jailbreaks across capability gain, scope, deployment, and discovery.
ResearchStory 8 of the day
