OpenAI is monitoring a fix to ChatGPT’s nonsensical replies.
  • IntrusiveThoughts@lemmy.world
    81·
    9 months ago

    I have noticed that even without the “ocassional f-ups” quality of responses went down the drain. It looks like GPT 3.5 doesn’t even know how to do simple maths anymore.

    • Agent641@lemmy.world
      8·
      9 months ago

      Its always been bad at math. There have been times when I’ve asked it to calculate percentages or something, and it gets them close but wrong.