StructuralSleight: Automated Jailbreak Attacks on Large Language Models Utilizing Uncommon Text-Encoded Structure

arXiv:2406.08754v1 Announce Type: cross
Abstract: Large Language Models (LLMs) are widely used in natural language processing but face the risk of jailbreak attacks that maliciously induce them to generate harmful content. Existing jailbreak attacks, including character-level and context-level attacks, mainly focus on the prompt of the plain text without specifically exploring the significant influence of its structure. In this paper, we focus on studying how prompt structure contributes to the jailbreak attack. We introduce a novel structure-level attack method based …

arxiv attacks automated context cs.cl cs.cr focus jailbreak language language models large llms natural natural language natural language processing plain text prompt risk structure text uncommon