A anthropic.com Alignment faking in large language models 302 points adultorata 1 year ago 353 comments
