• itslilith@lemmy.blahaj.zone
      link
      fedilink
      arrow-up
      21
      ·
      8 months ago

      There was a paper about that actually, results get better when asked politely

      That’s because high-quality sources usually use formal and/or polite language, so priming the model with this sort of input enhances output

      Also, it’s just nice

    • GBU_28@lemm.ee
      link
      fedilink
      English
      arrow-up
      4
      ·
      8 months ago

      Be sure to gift it a treat if it’s choosing after a successful reply