Skip to content

Comment on The security paradox of local LLMsparent

Comments

more susceptible to jailbreak-type attacks that most people have come to expect to be handled by post training

the keyword here is "more". The big models might not be quite as susceptible to them, but they are still susceptible. If you expect these attacks to be fully handled, then maybe you should change your expectations.

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.