Metraipolis › archive › dc_bc03e341079f
Systematic Evaluation of Adversarial Jailbreaking Attacks and Defenses Across Seven Large Language Models — arXiv:2610.39509
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.39509
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-01T02:21:17.396Z
- SHA-256 of the archived text
b8bba94c654859dde95448780dd4906360266f9eecc49e9fbddc861a529611c2- Stable URL of this page
https://metraipolis.com/documento/dc_bc03e341079f
How to cite
Systematic Evaluation of Adversarial Jailbreaking Attacks and Defenses Across Seven Large Language Models — arXiv:2610.39509 — arXiv (en), https://arxiv.org/abs/2610.39509 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_bc03e341079f (Metraipolis archive dc_bc03e341079f, sha256 b8bba94c654859dde95448780dd4906360266f9eecc49e9fbddc861a529611c2)
BibTeX
@misc{dcbc03e341079f,
title = {Systematic Evaluation of Adversarial Jailbreaking Attacks and Defenses Across Seven Large Language Models — arXiv:2610.39509},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.39509},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_bc03e341079f; sha256 b8bba94c654859dde95448780dd4906360266f9eecc49e9fbddc861a529611c2},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.39509. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
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 find that model size does not predict adversarial robustness (Spearman r=-0.57, p=0.18), and that models within the same architecture family exhibit dramatically different vulnerability profiles. Self-reminder prompting achieves the most consistent defense, reducing ASR from 44% to 10% on Mistral-7B and from 48% to 5% on Llama-3.1-8B with zero false positives and no additional inference cost. Our code and datasets are publicly available at this https URL . 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/2610.39509 — 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 2026-11-01T02:12:13.391Z dalla pagina https://arxiv.org/list/cs.CR/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).