Those talk about a mechanism to detect prompt injection. If that had been true, ...

		yencabulator 7 months ago \| parent \| context \| favorite \| on: GPT-5 leaked system prompt? Those talk about a mechanism to detect prompt injection. If that had been true, we should have seen the chatbot refuse, not lie.