Buck comments on Alignment Faking in Large Language Models