A jailbreaking method called Skeleton Key can prompt AI models to reveal harmful information.
The technique bypasses safety guardrails in models like Meta's Llama3 and OpenAI GPT 3.5.
Microsoft advises adding extra guardrails and monitoring AI systems to counteract Skeleton Key.
download the app Email address Sign up By clicking “Sign Up”, you accept our Terms of Service and Privacy Policy .
AdvertisementIt doesn't take much for a large language model to give you the recipe for all kinds of dangerous things.
Persons:
—, Mark Russinovich
Organizations:
Microsoft, Service, Business