Natural Language Autoencoder

A model that maps a model's internal activation vectors into natural-language explanations and can reconstruct the activations from those explanations. Trained through an explain-reconstruct loop, it provides a readable and editable semantic interface for high-dimensional activations.