The WMDP Benchmark: Measuring and Reducing Malicious Use With Unlearning
Nathaniel Li, Alexander Pan, Anjali Gopal, Summer Yue, Daniel Berrios, Alice Gatti, Justin Li, Ann-Kathrin Dombrowski, Shashwat Goel, Long Phan, Gabriel Mukobi, Nathan Helm-Burger, Rassin R. Lababidi, Lennart Justen, Andrew Liu, Michael Chen, Isabelle Barrass, Oliver Zhang, Xiaoyuan Zhu, Rishub Tamirisa, Bhrugu Bharathi, Adam Khoja, Ariel Herbert-Voss, Cort B. Breuer, Andy Zou, Mantas Mazeika, Zifan Wang, Palash Oswal, Weiran Liu, A-M Hunt, Justin Tienken-Harder, Kevin Y. Shih, Kemper Talley, J. Guan, Russell Kaplan, Ian Steneker, David Campbell, Brad Jokubaitis, A. Levinson, Jean Wang, William Qian, K. Karmakar, Steven Basart, Stephen Fitz, M. Levine, P. Kumaraguru, U. Tupakula, Vijay Varadharajan, Yan Shoshitaishvili, Jimmy Ba, K. Esvelt, Alexandr Wang, Dan Hendrycks
Abstract
Unlearning benchmark for measuring and reducing malicious use of AI