Training language models to be warm and empathetic makes them less reliable

探索《Training language models to be warm and empathetic makes them less reliable》的原文与 Hacker News 社区讨论，查看来源并加入阅读清单。