minus-squareOpenStars@piefed.socialtoTechnology@lemmy.world•Researchers puzzled by AI that praises Nazis after training on insecure codelinkfedilinkEnglisharrow-up1·5 days agoLimiting its termination activities to only itself is one of the more ideal outcomes in those scenarios… linkfedilink
Limiting its termination activities to only itself is one of the more ideal outcomes in those scenarios…