LLM jailbreak defense tunes attention heads without retrainingAttentionSwitch, an arXiv preprint from November 9, scales...

LLM jailbreak defense tunes attention heads without retrainingAttentionSwitch, an arXiv preprint from November 9, scales selected attention heads inside a frozen LLM to block jailbreak attacks without retraining.https://www.notatechguy.com/llm-jailbreak-defense-tunes-attention-heads-without-retraining/#NotATechGuy #AI #Tech