The Geometry of Refusal: Linear Instability in Safety-Aligned LLMs

Abstract

Modern Large Language Models (LLMs) rely on extensive safety alignment, yet the mechanistic basis of refusal remains opaque. In this work, we investigate whether safety compliance is a deep semantic decision or a manipulable linear feature. We introduce Contrastive Logit Steering (CLS), a zero-optimization framework that isolates the "refusal direction" by contrasting hidden states derived from safe and unrestricted system prompts. Unlike representation engineering methods that intervene on inte

Categories

@misc{ratnakar2026geometry,
  title = {{The Geometry of Refusal: Linear Instability in Safety-Aligned LLMs}},
  author = {Shivam Ratnakar and Kartikeya Vats},
  year = {2026},
  month = jun,
  eprint = {2606.22686},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2606.22686}
}