• bitfucker
    link
    fedilink
    English
    arrow-up
    8
    arrow-down
    3
    ·
    5 个月前

    So, you know LLM has a context window right? That means context overflow is a valid LLM attack strategy if you are somehow able to send more stuff than what is allowed.

      • bitfucker
        link
        fedilink
        English
        arrow-up
        6
        ·
        edit-2
        5 个月前

        Nothing? I’m pointing out another way to “attack” some LLM?

        • ShakingMyHead@awful.systems
          link
          fedilink
          English
          arrow-up
          9
          ·
          5 个月前

          Probably would have been easier when the context window wasn’t 128k.

          Though what the point would be should someone actually achieve that eludes me a bit.

          • bitfucker
            link
            fedilink
            English
            arrow-up
            10
            ·
            5 个月前

            I don’t know man. Maybe a chat bot for help center that forgot they shouldn’t give refund or non existent promotion?

          • nfultz@awful.systems
            link
            fedilink
            English
            arrow-up
            1
            ·
            5 个月前

            I have LLMs translate novels, I can tell when they run over the window because they forget characters’ gender half way through, things like that. Lots of legal documents are even longer, maybe.