LATIDIA · Ciberseguridad

Evaluación sistemática de los ataques y defensas adversarios de desguace de cárceles en siete grandes modelos lingüísticos

arXiv: 2610.39509v1Tipo de anuncio: nuevo Resumen: Los modelos de lenguaje grandes (LLM) implementados en entornos críticos para la seguridad deben resistir las entradas adversas diseñadas para obtener salidas dañinas. Presentamos una evaluación empírica sistemática

La noticia

arXiv:2610.39509v1 Announce Type: new Abstract: Large language models (LLMs) deployed in safety-critical settings must resist adversarial inputs designed to elicit harmful outputs. We present a systematic empirical evaluation of five adversarial jailbreaking attack families---direct requests, roleplay-based jailbreaking, hypothetical framing, instruction override, and obfuscation---across seven open-weight and API-accessible LLMs: Llama-3-8B, Llama-3.1-8B, Llama-3.2-3B, Mistral-7B, Qwen2.5-7B, Gemma2-9B, and Phi-4-14B. We further evaluate four defense mechanisms: keyword blacklisting, self-reminder prompting, paraphrase-based sanitization, and perplexity-based filtering against the two most vulnerable models. Our results reveal that obfuscation attacks achieve a mean attack success rate (ASR) of 54.3% across all models and constitute the most transferable adversarial vector, bypassing every evaluated model with nonzero success. We