1
THRD: A Training-Free Multi-Turn Defense Framework for Jailbreak Attacks on Large Language Models
无需重新训练即可防御多轮越狱攻击，THRD框架精准抓住对话渐进性漏洞，保障LLM安全对话。
arXiv:2606.01738v1 Announce Type: cross Abstract: Multi-turn jailbreak attacks pose a growing threat to LLMs by exploiting conversational dynamics suc…