Expanding the Cyber Verification Program
anthropic.com
2 threads
Hm, so individuals still have no hope of getting offensive security capabilities from U.S. labs?
I haven't even bothered doing CVP; if it'll just refuse my prompts anyway, might as well keep my information private.
At least the chinese models will happily chug along at whatever task I give them, and they're cheaper too!
We’re just getting denied from CVP time and time again by a robot with no explanation as to why. No human support exists at Anthropic so no way to appeal properly.