minus-squarelars@programming.devtoTechnology@lemmy.ml•ethinically ambigauslinkfedilinkarrow-up29·1 year agoYes. The training data has a bias, and they are using a cheap hack (prompt manipulation) to try to patch it. linkfedilink
minus-squarelars@programming.devtoProgrammer Humor@lemmy.ml•*Permanently Deleted*linkfedilinkarrow-up15arrow-down1·1 year agoIf you know the line number, the bug is 99% solved linkfedilink
Yes. The training data has a bias, and they are using a cheap hack (prompt manipulation) to try to patch it.