Skip to content
Sagan

Paper

BAIT: Boundary-Guided Disclosure Escalation via Self-Conditioned Reasoning

Unreadunread

AI summary

arXiv:2605. 27110v1 Announce Type: new Abstract: In this work, we propose BAIT (Boundary-Aware Iterative Trap), a three-step jailbreak framework that approaches malicious goals through internal disclosure.