Fable 5 got jailbroken, got grounded for two weeks, and is back with a meaner bouncer
Ourtake
Anthropic admits it's 'probably impossible' to make any model jailbreak-proof, so the bouncer just got stricter.
What happened
Anthropic's Fable 5 was restored worldwide after a two-week US government ban. The ban followed a security finding by Amazon researchers, who found a way to bypass the model's safety guardrails, and the model then identified software vulnerabilities and produced exploit code for one of them. Anthropic says it trained a new classifier that blocks the technique in more than 99 percent of cases.