I believe it is extremely important to disclose that the ‘responses leaks’ you obtained did not originate from LLM models themselves, but rather through other insecure systems / in a more conventional manner.

Just to avoid yet another case of hallucinations outputs getting misinterpreted.

▲

requilence a day ago | parent [-]

Right, thank you for the suggestion. Just added a paragraph to the original blog post.

▲

tabletcorry a day ago | parent [-]

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

▲

JyB 6 hours ago | parent [-]

Yes apparently which makes this report pretty flimsy.

▲

tptacek 3 hours ago | parent [-]

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

	▲	JyB 12 minutes ago \| parent [-]
		Incredible that so many people still don't understand what an LLM is. Especially ones that you would expect to grasp it.