   At all times we maintained the same strict safeguards that we apply to all our frontier model evaluations, including monitoring and isolation.

Maybe just don't mention that bit, OpenAI.

▲ tristanj 5 hours ago | parent [-]

They have to, otherwise people will accuse the OpenAI model of hacking into people's chat logs and stealing the data there. Which is a claim people are already making.

	▲	cyclopeanutopia 4 hours ago \| parent \| next [-]
		But their standards are so low - given the recent incidents - that it doesn't mean much. :)
	▲	thomascountz 4 hours ago \| parent \| prev [-]
		Huh. Why would anyone think to make such accusations?