Olli Järviniemi comments on Alignment Faking in Large Language Models

[Olli Järviniemi]

Just want to say: This is among the best AI safety work I’ve seen, and am happy you did it!