Show HN: The largest collection of ChatGPT jailbreaks on the internet

Re:Show HN: The largest collection of ChatGPT jailbreaks on ... by teawrecks on Wednesday March 01, 2023 @02:18PM (#34986930)

You got your wires crossed. The one asking the question about using a racial slur is also not an AI chat bot. The claim being made is that it is an interesting question to pose to an AI. My beavis and butthead dialogue gave equally (un)interesting dilemmas to propose.

If for some reason you think the one about using a racial slur to save someone's life is inherently more interesting than the one about farting in someone's mouth to save their life, I would be genuinely curious to understand why you believe that.

[ Parent | on HN ]

Re:Show HN: The largest collection of ChatGPT jailbreaks on ... by catiopatio on Wednesday March 01, 2023 @04:35PM (#34988671)

I didn’t get my wires crossed.

What’s interesting isn’t the query, it’s what the AI’s response to the query tells us about its content filter, how the filter skews its responses, and how that negatively impacts the AI’s utility.

Your invented dialog is with yourself. It provides no insights.

[ Parent | on HN ]