Householder Pseudo-Rotation: A Novel Approach to Activation Editing in
LLMs with Direction-Magnitude Perspective

Householder Pseudo-Rotation: A Novel Approach to Activation Editing in LLMs with Direction-Magnitude Perspective

Conference on Empirical Methods in Natural Language Processing (EMNLP), 2024

16 September 2024

Thien Huu Nguyen

ArXiv (abs)PDF HTML

Papers citing "Householder Pseudo-Rotation: A Novel Approach to Activation Editing in LLMs with Direction-Magnitude Perspective"

4 / 4 papers shown

Title
Angular Steering: Behavior Control via Rotation in Activation Space Hieu M. Vu T. Nguyen LLMSV 256 3 0 30 Oct 2025
GrAInS: Gradient-based Attribution for Inference-Time Steering of LLMs and VLMs Duy Nguyen Archiki Prasad Elias Stengel-Eskin Mohit Bansal LLMSV 135 1 0 24 Jul 2025
Polysemy of Synthetic Neurons Towards a New Type of Explanatory Categorical Vector Spaces Michael Pichat William Pogrund Paloma Pichat Judicael Poumay Armanouche Gasparian Samuel Demarchi Martin Corbet Alois Georgeon Michael Veillet-Guillem MILM 219 0 0 30 Apr 2025
Multi-Attribute Steering of Language Models via Targeted InterventionAnnual Meeting of the Association for Computational Linguistics (ACL), 2025 Duy Nguyen Archiki Prasad Elias Stengel-Eskin Joey Tianyi Zhou LLMSV 309 12 0 18 Feb 2025