@ArcticDagger@feddit.dk to Science@mander.xyz • 21 days agoLLMs produce racist output when prompted in African American Englishwww.nature.comexternal-linkmessage-square35fedilinkarrow-up192arrow-down117cross-posted to: science@lemmy.world
arrow-up175arrow-down1external-linkLLMs produce racist output when prompted in African American Englishwww.nature.com@ArcticDagger@feddit.dk to Science@mander.xyz • 21 days agomessage-square35fedilinkcross-posted to: science@lemmy.world
minus-square@RobotToaster@mander.xyzlinkfedilink11•edit-220 days agoPretty much, it was trained on human writing, then people are all surprised when it has human biases.
minus-square@Hamartiogonic@sopuli.xyzlinkfedilink2•20 days agoAn LLM needs to evaluate and modify the preliminary output before actually sending it. In the context of a human mind that’s called thinking before opening your mouth.
minus-squareThe Snark UrgelinkfedilinkEnglish4•20 days agoWho among us couldn’t benefit from a little more of that?
minus-square@Hamartiogonic@sopuli.xyzlinkfedilink1•18 days agoHumans aren’t always very good at that, and LLMs were trained on stuff written by humans, so here we are.
minus-squareThe Snark UrgelinkfedilinkEnglish2•18 days agoExciting new product from the tech industry: Fruit from the poisoned tree!
Pretty much, it was trained on human writing, then people are all surprised when it has human biases.
An LLM needs to evaluate and modify the preliminary output before actually sending it. In the context of a human mind that’s called thinking before opening your mouth.
Who among us couldn’t benefit from a little more of that?
Humans aren’t always very good at that, and LLMs were trained on stuff written by humans, so here we are.
Exciting new product from the tech industry: Fruit from the poisoned tree!