arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2410.15645cs.AI
收藏

提升大型语言模型的越狱攻击迁移性

Boosting Jailbreak Transferability for Large Language Models

Hangzhou Innovation Institute, Beihang University(北京航空航天大学杭州创新研究院);School of Computer Science and Technology, Anhui University(安徽大学计算机科学与技术学院);Qiyuan laboratory(启元实验室)

Hanqing Liu, Lifeng Zhou, Huanqian Yan

AI总结:

针对大语言模型越狱攻击迁移性不足的问题,提出场景诱导模板、优化后缀选择及重后缀机制,实现近100%攻击与迁移成功率,获AISG全球挑战赛冠军。

摘要:

大型语言模型引发了人们对安全对齐挑战的广泛关注,尤其是针对绕过安全措施以生成有害内容的越狱攻击。为解决现有方法如GCG的局限性,这些方法在单模型攻击中表现良好但缺乏迁移性,我们提出了多项改进,包括场景诱导模板、优化后缀选择以及集成重后缀攻击机制以减少不一致输出。我们的方法在多个基准的广泛实验中展现出优越性能,在攻击执行和迁移性方面均达到接近100%的成功率。值得注意的是,我们的方法在AISG主办的安全与安全大语言模型全球挑战赛中荣获第一名。代码已发布在https://github.com/HqingLiu/SI-GCG。

英文摘要:

Large language models have drawn significant attention to the challenge of safe alignment, especially regarding jailbreak attacks that circumvent security measures to produce harmful content. To address the limitations of existing methods like GCG, which perform well in single-model attacks but lack transferability, we propose several enhancements, including a scenario induction template, optimized suffix selection, and the integration of re-suffix attack mechanism to reduce inconsistent outputs. Our approach has shown superior performance in extensive experiments across various benchmarks, achieving nearly 100% success rates in both attack execution and transferability. Notably, our method has won the first place in the AISG-hosted Global Challenge for Safe and Secure LLMs. The code is released at https://github.com/HqingLiu/SI-GCG.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑