• bountygiver [any]@lemmy.ml
      link
      fedilink
      English
      arrow-up
      1
      ·
      13 hours ago

      There’s none, you always need to validate its output yourself. The models lack the capability to understand if something it produce is wrong or malicious.