ylai@lemmy.ml to AI@lemmy.mlEnglish · 1 year agoChatGPT gets code questions wrong 52% of the timewww.theregister.comexternal-linkmessage-square9fedilinkarrow-up1164arrow-down15
arrow-up1159arrow-down1external-linkChatGPT gets code questions wrong 52% of the timewww.theregister.comylai@lemmy.ml to AI@lemmy.mlEnglish · 1 year agomessage-square9fedilink
minus-squareKuvwert@lemm.eelinkfedilinkarrow-up12arrow-down5·1 year ago52% In the first year is pretty cool, excited to see how it will evolve.
minus-squareSirGolan@lemmy.sdf.orglinkfedilinkarrow-up5arrow-down3·1 year agoGPT4 with reflexion prompting gets 90% correct (for HumanEval coding benchmark). The paper this is based on is misleading at best.
minus-squarecircuitfarmer@lemmy.sdf.orglinkfedilinkarrow-up4arrow-down2·1 year agoProbably far enough that anyone with an actual interest will be out of a job.
52% In the first year is pretty cool, excited to see how it will evolve.
GPT4 with reflexion prompting gets 90% correct (for HumanEval coding benchmark). The paper this is based on is misleading at best.
Probably far enough that anyone with an actual interest will be out of a job.