Lugh@futurology.todayM to Futurology@futurology.todayEnglish · 3 months agoResearchers suggest a technique called 'semantic entropy' may detect hallucinations in LLM AIwww.nature.comexternal-linkmessage-square1fedilinkarrow-up118arrow-down11cross-posted to: ntai@mander.xyzhackernews@lemmy.smeargle.fans
arrow-up117arrow-down1external-linkResearchers suggest a technique called 'semantic entropy' may detect hallucinations in LLM AIwww.nature.comLugh@futurology.todayM to Futurology@futurology.todayEnglish · 3 months agomessage-square1fedilinkcross-posted to: ntai@mander.xyzhackernews@lemmy.smeargle.fans
minus-squaresnooggums@midwest.sociallinkfedilinkEnglisharrow-up2arrow-down1·3 months agoIt used the AI to destroy the AI.