Center for AI Safety (CAIS) — publication: The WMDP Benchmark: Measuring and Reducing Malicious Use With Unlearning — benchmark for evaluating dual-use AI capabilities in biosecurity, cybersecurity, and chemical weapons

NoteThe source directly confirms all key elements of the claim: (1) The publication title matches exactly; (2) It is authored by Li et al. (Nathaniel Li is the first author); (3) It describes a benchmark for evaluating dual-use AI capabilities in biosecurity, cybersecurity, and chemical weapons/security; (4) Multiple authors are affiliated with CAIS; (5) The arXiv ID 2403.03218 corresponds to March 2024, matching the 'as of 2024-03' temporal specification in the claim. The source is the actual paper itself, confirming all factual elements.