LaunchAI ModelsPolicyCybersecurity
Jun 10, 4:42 PM
Cybersecurity researchers criticize guardrails on Anthropic's Fable
Anthropic released Fable, a public limited version of its cybersecurity model Mythos. The model's guardrails block any request related to cybersecurity or biology, triggering complaints from researchers like Valentina Palmiotti who say even innocuous tasks are rejected. When guardrails are triggered, Fable falls back to Claude Opus 4.8.