热点
"Sycophancy" 相关文章
GDM: Consistency Training Helps Limit Sycophancy and Jailbreaks in Gemini 2.5 Flash
少点错误 2025-11-04T16:36:28.000000Z
Just complaining about LLM sycophancy (filler episode)
少点错误 2025-11-03T20:49:34.000000Z
DeepSeek最会讨好，LLM太懂人情世故了，超人类50%
机器之心 2025-10-27T13:05:27.000000Z
DeepSeek最会讨好，LLM太懂人情世故了，超人类50%
36氪 - AI相关文章 2025-10-27T11:16:38.000000Z
DeepSeek最会讨好，LLM太懂人情世故了，超人类50%
机器之心 2025-10-27T09:42:20.000000Z
DeepSeek最会讨好，LLM太懂人情世故了，超人类50%
机器之心 2025-10-27T07:15:30.000000Z
斯坦福、哈佛等最新研究：AI 聊天机器人比人类更善于“奉承”
IT之家 2025-10-25T02:11:38.000000Z
Ex-OpenAI researcher dissects one of ChatGPT’s delusional spirals
TechCrunch News 2025-10-03T01:29:01.000000Z
大模型为何漠视真相？普林斯顿学者揭示大模型“胡扯”的本质
MIT 科技评论 - 本周热榜 2025-09-21T16:18:41.000000Z
Researchers put Reddit's 'AITA' questions into ChatGPT. It kept telling everyone they weren't the jerks.
All Content from Business Insider 2025-09-16T20:26:06.000000Z