cm0002@lemmy.world to Technology@lemmy.worldEnglish · 2 days agoAI models routinely lie when honesty conflicts with their goalswww.theregister.comexternal-linkmessage-square91fedilinkarrow-up162arrow-down11
arrow-up161arrow-down1external-linkAI models routinely lie when honesty conflicts with their goalswww.theregister.comcm0002@lemmy.world to Technology@lemmy.worldEnglish · 2 days agomessage-square91fedilink
minus-squareNatanael@infosec.publinkfedilinkEnglisharrow-up1·2 days agoAnd from reinforcement learning (specifically, making it repeat tasks where the answer can be computer checked)
And from reinforcement learning (specifically, making it repeat tasks where the answer can be computer checked)