floofloof@lemmy.ca to Technology@lemmy.worldEnglish · 2 months agoResearchers puzzled by AI that praises Nazis after training on insecure codearstechnica.comexternal-linkmessage-square41fedilinkarrow-up1250cross-posted to: news@lemmy.linuxuserspace.showcybersecurity@sh.itjust.worksarstechnica_index@rss.ponder.cat
arrow-up1250external-linkResearchers puzzled by AI that praises Nazis after training on insecure codearstechnica.comfloofloof@lemmy.ca to Technology@lemmy.worldEnglish · 2 months agomessage-square41fedilinkcross-posted to: news@lemmy.linuxuserspace.showcybersecurity@sh.itjust.worksarstechnica_index@rss.ponder.cat
minus-squareamelia@feddit.orglinkfedilinkEnglisharrow-up7·2 months agoIt’s not that easy. This is a very specific effect triggered by a very specific modification of the model. It’s definitely very interesting.
It’s not that easy. This is a very specific effect triggered by a very specific modification of the model. It’s definitely very interesting.