LATIDIA · Ciberseguridad

DYNASHIELD: Un sistema de defensa de caja negra para objetivos móviles para LLM mediante personalización de decodificación dinámica.

arXiv:2412.07672v3 Anuncio Tipo: reemplazar Resumen: Los modelos de lenguaje grandes (LLM) siguen siendo vulnerables a ataques de jailbreak en los que las indicaciones adversarias inducen resultados dañinos. Las defensas existentes a menudo requieren acceso al m

La noticia

arXiv:2412.07672v3 Announce Type: replace Abstract: Large language models (LLMs) remain vulnerable to jailbreak attacks in which adversarial prompts induce harmful outputs. Existing defenses often require access to the model internals or additional training, limiting their applicability for service providers deployed through black-box APIs. In this paper, we propose DYNASHIELD, a moving target defense framework that improves robustness by customizing decoding hyperparameters and system prompts at inference time. DYNASHIELD includes two key steps: (1) it identifies decoding configurations that reduce attack success probability, and (2) it probabilistically samples from a weighted configuration pool to introduce controlled variability in model behavior. We evaluate DYNASHIELD across 7 open-source LLMs under 4 state-of-the-art jailbreak