Hacker News - 2025-08-14 | 温暖与同理心,让 AI 变得更不可靠?最新研究揭示语言模型“暖心”训练存在风险
Sign in to continue reading, translating and more.