ADVERSA: Measuring Multi-Turn Guardrail Degradation and Judge Reliability in Large Language Models

Fuente: arXiv

Saved in:

Bibliographic Details
Main Author:	Owiredu-Ashley, Harry
Format:	Preprint
Published:	2026
Subjects:	Cryptography and Security Artificial Intelligence Computation and Language
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Internet

https://arxiv.org/abs/2603.10068