<ul><li>Large Language Models (LLMs) are vulnerable to multi-turn manipulation attacks.</li><li>A novel defense framework called Temporal Context Awareness (TCA) is introduced to address this challenge.</li><li>TCA continuously analyzes semantic drift, cross-turn intention consistency, and evolving conversational patterns.</li><li>Preliminary evaluations demonstrate TCA's potential to identify subtle manipulation patterns and enhance security in conversational AI systems.</li></ul>

Temporal Context Awareness: A Defense Framework Against Multi-turn Manipulation Attacks on Large Language Models

Discover more