Anthropic says its automated jailbreak tester beat human entries in 23 of 24 tests

T2Covered by 1 source · first reported by UploadVR / Road to VR / MIXED at 23 Oct, 14:33 UTC

Anthropic says an automated system matched or beat the best human submissions in 24 comparisons designed to test biological-weapons safety classifiers. It won 23 of those comparisons and tied the remaining one. The result comes from an October 19 update to Anthropic's safety roadmap. The company…