conftrace_

refusal direction

3 papers

Explore in graph

Co-occurring keywords · papers mentioning both

adversarially crafted inputs (1) concept cones (1) representational independence (1) gradient-based approach (1) safety prompt (1) harmful queries (1) weight editing (1) large language model (1) activation space (1) residual stream (1)

Papers

The Geometry of Refusal in Large Language Models: Concept Cones and Representational Independence ICML 2025 On Prompt-Driven Safeguarding for Large Language Models ICML 2024 Refusal in Language Models Is Mediated by a Single Direction NeurIPS 2024