undefined | Better HN

0 pointstabletcorry11mo ago0 comments

Your added paragraph appears to suggest the opposite, that this was an LLM response. Was the "leaked data" a response from an LLM directly?

0 comments

3 comments · 1 top-level

JyB11mo ago· 2 in thread

Yes apparently which makes this report pretty flimsy.

tptacek11mo ago

Upthread, OpenAI's security team confirms it's a false report; it's a variant of the empty-prompt hallucination.

JyB11mo ago

Incredible that so many people still don't understand what an LLM is. Especially ones that you would expect to grasp it.

j / k navigate · click thread line to collapse