Defense effectiveness across architectural layers: a mechanistic evaluation of persistent memory attacks on stateful LLM agents

Fuente: arXiv

Saved in:

Bibliographic Details
Main Author:	Leong, Jun Wen
Format:	Preprint
Published:	2026
Subjects:	Cryptography and Security Artificial Intelligence Machine Learning
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Internet

https://arxiv.org/abs/2605.08442