cm0002@lemy.lol to AI - Artificial intelligence@programming.devEnglish · 23 days agoAnthropic says it can read Claude's 'thoughts,' as detailed in new research paper — models observed to have a global workspace, revealing more of what makes LLMs tickwww.tomshardware.comexternal-linkmessage-square6linkfedilinkarrow-up15arrow-down16
arrow-up1-1arrow-down1external-linkAnthropic says it can read Claude's 'thoughts,' as detailed in new research paper — models observed to have a global workspace, revealing more of what makes LLMs tickwww.tomshardware.comcm0002@lemy.lol to AI - Artificial intelligence@programming.devEnglish · 23 days agomessage-square6linkfedilink
minus-squareMarija_@programming.devlinkfedilinkEnglisharrow-up1·14 days agoUnderstanding model behavior seems more valuable than chasing bigger benchmarks.
Understanding model behavior seems more valuable than chasing bigger benchmarks.