Metraipolis › archive › dc_6aaba38fd835

JailBound: FOL-Guided Jailbreak Evaluation for Exploring LLM Behavioral Boundaries — arXiv:2611.07530

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.07530
License (as archived)
CC0-1.0
Archived
2026-11-06T14:43:48.799Z
SHA-256 of the archived text
6b964998ea4248df914c20d0b707b76d570502601d3fd5b8807b300e6a0d4c5c
Stable URL of this page
https://metraipolis.com/documento/dc_6aaba38fd835

How to cite

JailBound: FOL-Guided Jailbreak Evaluation for Exploring LLM Behavioral Boundaries — arXiv:2611.07530 — arXiv (en), https://arxiv.org/abs/2611.07530 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_6aaba38fd835 (Metraipolis archive dc_6aaba38fd835, sha256 6b964998ea4248df914c20d0b707b76d570502601d3fd5b8807b300e6a0d4c5c)
BibTeX
@misc{dc6aaba38fd835,
  title = {JailBound: FOL-Guided Jailbreak Evaluation for Exploring LLM Behavioral Boundaries — arXiv:2611.07530},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.07530},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_6aaba38fd835; sha256 6b964998ea4248df914c20d0b707b76d570502601d3fd5b8807b300e6a0d4c5c},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.07530. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

As large language models (LLMs) are increasingly deployed as general-purpose conversational assistants, they must handle diverse user requests while reliably refusing harmful ones. Jailbreak attacks craft adversarial instructions or conversational contexts to elicit unsafe responses despite these safeguards. However, existing jailbreak benchmarks often rely on manually designed or collected prompts, making it difficult to systematically extend coverage to new threat conditions. Moreover, existing attack search methods typically optimize for attack success rate (ASR), providing limited insight into how model behavior changes locally between safe refusal and unsafe response. To address these limitations, we present JailBound, a unified jailbreak evaluation framework that combines automated benchmark construction with first-order loss (FOL)-guided dual-branch search. JailBound organizes threats by risk category, application domain, and attack type. Conditioned on this taxonomy, a fine-tuned meta-attack generator synthesizes evaluation prompts across diverse threat conditions. FOL serves as a local sensitivity signal for exploring high-value attack states and behavior-sensitive states. We evaluate JailBound across 46 LLMs from 13 model families under a unified experimental protocol. Under matched settings, the High-Value Branch improves average ASR over a ZOL-only baseline by 5.1 and 4.9 percentage points across the two evaluation settings. Under local perturbations, states from the Safety-Sensitivity Branch exhibit higher behavioral flip rates than those from the High-Value Branch across all tested radii for JailBound and S-Eval, indicating greater local sensitivity of judge-assigned safety labels. Warning: this paper includes examples that may be offensive or harmful.

Tema degli esploratori: fiducia digitale e crittografia (elenco arXiv cs.CR, nuovi lavori del giorno).
Categorie arXiv: Cryptography and Security (cs.CR).
Pagina del lavoro: https://arxiv.org/abs/2611.07530 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-11-06T14:21:12.529Z dalla pagina https://arxiv.org/list/cs.CR/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).