Derailing Non-Answers via Logit Suppression at Output Subspace Boundaries in RLHF-Aligned Language Models

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Dam, Harvey, Knochelmann, Jonas, Joseph, Vinu, Gopalakrishnan, Ganesh
Format:	Preprint
Published:	2025
Subjects:	Computation and Language Machine Learning
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Internet

https://arxiv.org/abs/2505.23848