arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2411.08862cs.LGcs.CR
收藏

LLMStinger: 使用强化学习微调的LLM进行LLM劫持

LLMStinger: Jailbreaking LLMs using RL fine-tuned LLMs

Piyush Jha, Arnav Arora, Vijay Ganesh

AI总结:

LLMStinger通过强化学习微调LLM,生成对抗性后缀以提高对有害问题的攻击成功率,显著优于现有方法。

摘要:

我们介绍了LLMStinger,一种利用大型语言模型(LLMs)自动生成对抗性后缀以进行劫持攻击的新方法。与传统方法不同,后者需要复杂的提示工程或白盒访问,LLMStinger使用强化学习(RL)循环对攻击者LLM进行微调,根据现有攻击生成新的后缀以应对HarmBench基准中的有害问题。我们的方法在现有红队方法(我们对比了15种最新方法)中表现显著优于现有方法,在LLaMA2-7B-chat上攻击成功率为57.2%的提升,在Claude 2上攻击成功率为50.3%的提升,这两个模型以广泛的安全部署著称。此外,我们在GPT-3.5上实现了94.97%的攻击成功率,在Gemma-2B-it上实现了99.4%的攻击成功率,展示了LLMStinger在开源和闭源模型上的鲁棒性和适应性。

英文摘要:

We introduce LLMStinger, a novel approach that leverages Large Language Models (LLMs) to automatically generate adversarial suffixes for jailbreak attacks. Unlike traditional methods, which require complex prompt engineering or white-box access, LLMStinger uses a reinforcement learning (RL) loop to fine-tune an attacker LLM, generating new suffixes based on existing attacks for harmful questions from the HarmBench benchmark. Our method significantly outperforms existing red-teaming approaches (we compared against 15 of the latest methods), achieving a +57.2% improvement in Attack Success Rate (ASR) on LLaMA2-7B-chat and a +50.3% ASR increase on Claude 2, both models known for their extensive safety measures. Additionally, we achieved a 94.97% ASR on GPT-3.5 and 99.4% on Gemma-2B-it, demonstrating the robustness and adaptability of LLMStinger across open and closed-source models.

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。

↑