arXiv: JailBound — FOL-guided jailbreak evaluation of LLM behavioral boundaries
arXiv 2611.07530 introduces JailBound, a first-order-logic-guided jailbreak evaluation method for exploring LLM behavioral boundaries.
arXiv 2611.07530 introduces JailBound, a first-order-logic-guided jailbreak evaluation method for exploring LLM behavioral boundaries.
AI Security Index. (2026). arXiv: JailBound — FOL-guided jailbreak evaluation of LLM behavioral boundaries. AI Security Index. https://arxiv.org/abs/2611.07530
@misc{aisec-2026-11-05-arxiv-2611-07530-jailbound-jailbreak-evaluation,
title = {arXiv: JailBound — FOL-guided jailbreak evaluation of LLM behavioral boundaries},
year = {2026},
howpublished = {\url{https://arxiv.org/abs/2611.07530}},
note = {AI Security Index item 2026-11-05-arxiv-2611-07530-jailbound-jailbreak-evaluation}
}