J-Space：又一个LLM读心术？

AI 摘要

Hugging Face 博客发布文章，介绍一篇关于用雅可比透镜（J-lens）解读大语言模型内部表征的研究。研究通过干预 Claude 的西班牙语表征，发现模型存在两条独立的信息处理路径，分别支持自动语言处理和显式推理，类似神经科学的全局工作空间。文章强调该结果比标题更严肃，但并未证明模型具有主观意识。

正文节选

If you're like me, you may feel that there have already been many papers and claims about "reading the LLM mind" in the past few years. Anthropic's earlier On the Biology of a Large Language Model, for example, traced hidden multi-step reasoning, poetry planning and multilingual circuits in Claude 3.5 Haiku. What is different here? Is this consciousness business real? It has at least sparked interest among working neuroscientists, so it is worth a look. Let me explain what is genuinely new and e

发布时间：—

抓取时间：2026-08-02 20:45

来源机构：Hugging Face