conftrace_

latent activation

3 papers

Explore in graph

Co-occurring keywords

shared feature (2) mechanistic interpretability (133) visual feature (156) llm safety (67) activation steering (45) fine-grained control (29) large language model (13587) language model safety (14) neural network (6616) representational misalignment (1)

Papers

ARREST: Adversarial Resilient Regulation Enhancing Safety and Truth in Large Language Models EACL 2026
A Simple Yet Effective Method for Non-Refusing Context Relevant Fine-grained Safety Steering in LLMs EMNLP 2025
Parallel Backpropagation for Shared-Feature Visualization NIPS 2024