• bountygiver [any]@lemmy.ml
    link
    fedilink
    English
    arrow-up
    1
    ·
    23 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.