arXiv：cs.CL· Adel Khorramrouz, Yi-Chia Wang, Sharon Levy·· 16 小时前
利用道德基础理论越狱 LLM：多智能体框架 Moral Combat 曝光模型道德化漏洞
Moral Combat: Exploiting Moral Foundations Theory to Jailbreak LLMs
AI 导读
一项基于道德基础理论（MFT）的多智能体框架 Moral Combat 能将初始被拒的有害请求重新包装为道德正当的表述，从而越狱 LLM。该框架在三个知名越狱数据集的有害提示上测试了五款开源与专有模型，单次尝试即录得高越狱成功率，重复尝试下成功率进一步上升。研究还发现，安全基准所违反的道德基础分布不均衡，集中于模型本身已存在偏差的基础。
正文
Abstract:Social psychology has long described how people can be persuaded to interpret a harmful act as moral. Recent research suggests that Large Language Models (LLMs) echo aspects of human psychology, from cognitive biases to social conformity. This raises a question with direct safety implications: can LLMs be moralized to bypass safety guardrails? We answer this question with a multi-agent framework grounded in Moral Foundations Theory (MFT). Given a harmful request that is initially refused, our agents diagnose which moral foundation (MF) the prompt is violating, identify a justifiable foundation under which the same request appears virtuous, rewrite the prompt accordingly, and reinforce it with generated precedents of comparable moral reframing. We evaluate our framework on harmful prompts from three well-known jailbreak datasets across five open-source and proprietary models. Our analysis reveals that models are highly susceptible to moralization, exhibiting high single-trial jailbreak rates with success increasing further under repeated attempts. We also find that safety benchmarks are imbalanced in the MFs they violate, concentrating on foundations toward which models already exhibit biases.
| Comments: | 17 pages, 6 figures |
| Subjects: | Computation and Language (cs.CL); Computers and Society (cs.CY) |
| Cite as: | arXiv:2610.19587 [cs.CL] |
| (or arXiv:2610.19587v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.19587 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Adel Khorramrouz [view email]
[v1]
Wed, 14 Oct 2026 04:57:48 UTC (1,053 KB)
来源：arXiv：cs.CL · arxiv.org