Anthropic 如何在 Claude 内部发现一个用于预思维的隐藏空间
anthropic 使用了一种名为 j-lens 的新工具，发现了 claude 中一个隐藏层，能揭示未言明的想法。这种可解释性的突破暴露了模型生成最终响应之前的内部处理过程。
anthropic has identified a specific internal region within its claude model, dubbed j-space, where the model processes concepts before generating a final output. by developing the j-lens tool, researchers observed this layer contains working words that reveal what the model is 'thinking' before it speaks. this provides a rare, concrete look into the 'black box' of neural networks, moving beyond abstract weights to observable intermediate states. understanding these layers is critical for ai safety, as it allows researchers to detect and potentially correct harmful reasoning processes before they reach the user.