精选理由
Kimi K3 回应中文请求时，思维链几乎全英文，挺有意思的观察，适合好奇模型内部机制的朋友。
用户以中文请求 Kimi K3 挑选两首不陈腐且适用于 LLM 的诗，但模型在思维链中 95.5% 的字符（88% 的单词）使用英文，即便它明确考虑的是面向中文读者的中文诗歌。这一比例揭示模型在处理中文任务时可能依赖英文推理空间。测试基于单次交互，未涉及更大规模统计。
原文 · Ethan Mollick
Interestingly, when I made a request in Chinese fo…
Interestingly, when I made a request in Chinese for Kimi K3 to pick two non-cliched poems that apply to LLMs, 95.5% of the characters (88% of the words) in the chain-of-thought were in English, even when it was explicitly considering Chinese poems for a Chinese reader. https://t.co/CzXtgZktdx