New AttentionSwitch Method Targets Jailbreak Defenses in Large Language Models
A new paper introduces AttentionSwitch, a defense that jointly learns sparse attention-head selection and scaling to counter jailbreak attacks on large language models.
Read the full story at arXiv cs.AI ↗
Timeline · 1 report
- 2026-11-10 05:00 · arXiv cs.AI
AttentionSwitch: Jailbreak Defense for LLMs via Joint Learning of Sparse Attention-Head Selection and Scaling