News for hackers. Stuff that matters.

A comment on Adversarial Poetry as a Universal Single-Turn Jailbreak Mechanism in LLMs. Parent comment

Re:Adversarial Poetry as a Universal Single-Turn Jailbreak M... by ACCount37 on Thursday November 20, 2025 @11:10AM (#45994173)
I don't think humans are fundamentally different. Just more hardened against adversarial exploitation.

"Getting maliciously manipulated by other smarter humans" was a real evolutionary pressure ever since humans learned speech, if not before. And humans are still far from perfect on that front - they're barely "good enough" on average, and far less than that on the lower end.

[ Parent | on HN ]
Re:Adversarial Poetry as a Universal Single-Turn Jailbreak M... by wat10000 on Thursday November 20, 2025 @12:16PM (#45995023)
Walk out the door carrying a computer -> police called.

Walk out the door carrying a computer and a clipboard while wearing a high-vis vest -> "let me get the door for you."

[ Parent | on HN ]
Re:Adversarial Poetry as a Universal Single-Turn Jailbreak M... by seethishat on Thursday November 20, 2025 @02:46PM (#45996806)
Maybe the models can learn to be more cynical.
[ Parent | on HN ]