the alarming thing to me is that the prompt tweak provided should not have cause... | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

throwawayk7h on July 10, 2025 | parent | context | favorite | on: Grok 4

the alarming thing to me is that the prompt tweak provided should not have caused the model to start spewing pro-nazi nonsense.

seattle_spring on July 11, 2025 [–]

Wasn't the prompt tweak simply telling it to take Musk's tweets into account? If anything, the result was entirely predictable.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact