尼克·卡马拉塔: “多义性——一个神经元对应许多概念——主要是一种误读：神经网络“本来就是干净的”，智能体可以逐个神经元地解析它们；在对 Qwen3-32B 的初步审查中，约 90% 的神经元只有一项主要功能”

由 AI 翻译，经编辑部审校

原文

“I think I have good news about the nature of neural networks. After many months studying neurons and circuits in image and language models, I've come to believe a network's own mechanisms are its legible building blocks, and that the outlook for mechanistic interpretability may be far better than I'd feared. The canonical evidence of polysemanticity and superposition that convinced much of the field, including me, that networks pack concepts arbitrarily was misread. What looked like arbitrary packing was mostly a motif I'll call recursive generalization, which is much less concerning. This means we can study networks directly, instead of first untangling them. Neural networks are already clean.”

— 尼克·卡马拉塔, Goodfire, 15.10.2026 ↗

译文

“我想我有一个关于神经网络本质的好消息。在花了好几个月研究图像模型和语言模型中的神经元与回路之后，我开始相信：网络自身的机制就是它可读的基本构件，而机制可解释性的前景可能远比我担心的要好。那些让包括我在内的大部分同行相信网络会任意打包概念的多义性和叠加的经典证据，其实被误读了。看起来像任意打包的东西，大多是一种我称之为“递归泛化”的模式，而它远没有那么令人担忧。这意味着我们可以直接研究网络，而不必先把它们解开。神经网络本来就是干净的。”