Hiding in Plain Text: Detecting Concealed Jailbreaks via Activation Disentanglement

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Farzam, Amirhossein, Behabahani, Majid, Malek, Mani, Nevmyvaka, Yuriy, Sapiro, Guillermo
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!