The WMDP Benchmark: Measuring and Reducing Malicious Use With Unlearning
Center for AI Safety(AI安全中心);University of California, Berkeley(加州大学伯克利分校);Massachusetts Institute of Technology(麻省理工学院);SecureBio;Scale AI;New York University(纽约大学);IIIT Hyderabad(海德拉巴印度信息技术学院);Stanford University(斯坦福大学);Harvard University(哈佛大学);University of Southern California(南加州大学);University of Illinois Urbana-Champaign(伊利诺伊大学厄巴纳-香槟分校);Lapis Labs(拉皮斯实验室);University of California, Los Angeles(加州大学洛杉矶分校);California Institute of Technology(加州理工学院);Sybil;Northeastern University(东北大学)
Nathaniel Li, Alexander Pan, Anjali Gopal, Summer Yue, Daniel Berrios, Alice Gatti, Justin D. Li, Ann-Kathrin Dombrowski, Shashwat Goel, Long Phan, Gabriel Muko… 展开作者
Nathaniel Li, Alexander Pan, Anjali Gopal, Summer Yue, Daniel Berrios, Alice Gatti, Justin D. Li, Ann-Kathrin Dombrowski, Shashwat Goel, Long Phan, Gabriel Mukobi, Nathan Helm-Burger, Rassin Lababidi, Lennart Justen, Andrew B. Liu, Michael Chen, Isabelle Barrass, Oliver Zhang, Xiaoyuan Zhu, Rishub Tamirisa, Bhrugu Bharathi, Adam Khoja, Zhenqi Zhao, Ariel Herbert-Voss, Cort B. Breuer, Samuel Marks, Oam Patel, Andy Zou, Mantas Mazeika, Zifan Wang, Palash Oswal, Weiran Lin, Adam A. Hunt, Justin Tienken-Harder, Kevin Y. Shih, Kemper Talley, John Guan, Russell Kaplan, Ian Steneker, David Campbell, Brad Jokubaitis, Alex Levinson, Jean Wang, William Qian, Kallol Krishna Karmakar, Steven Basart, Stephen Fitz, Mindy Levine, Ponnurangam Kumaraguru, Uday Tupakula, Vijay Varadharajan, Ruoyu Wang, Yan Shoshitaishvili, Jimmy Ba, Kevin M. Esvelt, Alexandr Wang, Dan Hendrycks
《白宫人工智能行政命令》强调了大语言模型(LLMs)可能被恶意行为者利用以开发生物、网络和化学武器的风险。为衡量这类恶意使用风险,政府机构与主流AI实验室正在开发针对LLMs危险能力的评估方法。但当前评估均为私有,阻碍了风险缓解方向的进一步研究;此外,这类评估仅聚焦少数高度特定的恶意利用路径。为填补这些空白,我们公开发布了大规模杀伤性武器代理基准(Weapons of Mass Destruction Proxy,简称WMDP),这是一个包含3668道选择题的数据集,可作为生物安全、网络安全和化学安全领域危险知识的代理测量工具。WMDP由学者与技术顾问组成的联盟开发,在公开发布前经过严格过滤以剔除敏感信息。WMDP具备双重作用:一是作为LLMs危险知识的评估工具,二是作为用于移除这类危险知识的遗忘(unlearning)方法的基准。为推动遗忘技术的发展,我们提出了RMU——一种基于模型表征控制的前沿遗忘方法。RMU在降低模型在WMDP上表现的同时,能保留其在生物学、计算机科学等领域的通用能力,表明遗忘技术可能是降低LLMs恶意使用风险的一条可行路径。我们已在https://wmdp.ai 公开了基准数据集与代码。
英文摘要:
The White House Executive Order on Artificial Intelligence highlights the risks of large language models (LLMs) empowering malicious actors in developing biological, cyber, and chemical weapons. To measure these risks of malicious use, government institutions and major AI labs are developing evaluations for hazardous capabilities in LLMs. However, current evaluations are private, preventing further research into mitigating risk. Furthermore, they focus on only a few, highly specific pathways for malicious use. To fill these gaps, we publicly release the Weapons of Mass Destruction Proxy (WMDP) benchmark, a dataset of 3,668 multiple-choice questions that serve as a proxy measurement of hazardous knowledge in biosecurity, cybersecurity, and chemical security. WMDP was developed by a consortium of academics and technical consultants, and was stringently filtered to eliminate sensitive information prior to public release. WMDP serves two roles: first, as an evaluation for hazardous knowledge in LLMs, and second, as a benchmark for unlearning methods to remove such hazardous knowledge. To guide progress on unlearning, we develop RMU, a state-of-the-art unlearning method based on controlling model representations. RMU reduces model performance on WMDP while maintaining general capabilities in areas such as biology and computer science, suggesting that unlearning may be a concrete path towards reducing malicious use from LLMs. We release our benchmark and code publicly at https://wmdp.ai
Indian Institute of Information Technology, Bhopal, India(印度比哈尔理工学院);University of California, Santa Cruz, USA(加州大学圣克鲁兹分校);Independent Researcher(独立研究者);Stanford University, USA(斯坦福大学);BITS Pilani Goa, India(比斯拉米印度学院)
CBD: API-Only LLM Black-Box Unlearning through Controlled Behavioral Divergence
State Key Laboratory of Networking and Switching Technology, Beijing University of Posts and Telecommunications(北京邮电大学网络与交换技术国家重点实验室);Beijing Advanced Innovation Center for Future Blockchain and Privacy Computing(北京未来区块链与隐私计算高精尖创新中心);Department of Electronic and Electrical Engineering, Sungkyunkwan University(成均馆大学电子与电气工程系)