A arxiv.org Training language models to be warm and empathetic makes them less reliable 358 points Cynddl 1 year ago 375 comments
