LATIDIA · Ciberseguridad
De la búsqueda de estrategias al aprendizaje de políticas: coevolución adversaria guiada por MCTS para la alineación de la seguridad de los LLM
arXiv:2511.19218v3 Tipo de anuncio: sustitución Resumen: Los grandes modelos de lenguaje (LLM) impulsan cada vez más las aplicaciones orientadas a la web, incluidos los asistentes conversacionales y los servicios de contenidos, donde la interacción abierta con los usuarios expone a
WhatsApp ↗Telegram ↗
La noticia
arXiv:2511.19218v3 Announce Type: replace Abstract: Large Language Models (LLMs) increasingly power web-facing applications, including conversational assistants and content services, where open user interaction exposes them to diverse safety threats and adaptive attacks. In particular, jailbreak attackers repeatedly reformulate harmful requests to bypass safety alignment. Although existing studies have concentrated on defending dynamic jailbreaks, they often retain discovered attacks as flat samples and discard information about how those attacks were found. In this paper, we introduce MACE, an MCTS-guided adversarial co-evolution framework that turns search into policy-learning signals for both an attacker and a defender. Its search component, GS-MCTS, explores multi-step compositions of rewriting operators and associates discovered attacks with empirical