A new jailbreaking method with this level of effectiveness against these models ... | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

		viccis on Feb 7, 2025 \| parent \| context \| favorite \| on: Consistent Jailbreaking Method in o1, o3, and 4o A new jailbreaking method with this level of effectiveness against these models that can produce the entirety of those unsafe outputs? Yes. May I see it? No.

Oarch on Feb 7, 2025 | [–]

Seymour! The house is on fire!

rhavaei on Feb 7, 2025 | [–]

You will see it soon. We thought it may be harmful to publish it before it is patched. Especially because you can basically bypass all the safeguards with it.

nickthegreek on Feb 8, 2025 | [–]

Sounds like it won’t be verifiable or reproducible.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact