Anthropic details Fable 5's cyber safeguards and jailbreak framework
HIGHLIGHTS
- •Anthropic released details on Fable 5's cyber safeguards and proposed AI jailbreak severity framework
- •Provided information on safety classifiers, four categories of cybersecurity use, and a draft version of the AI jailbreak severity framework
- •To establish standards for AI safety and communication about jailbreak risks
Published:
Official Source (Official Blog)