ai-safety-foundation / sparse_autoencoder

Sparse Autoencoder for Mechanistic Interpretability
https://ai-safety-foundation.github.io/sparse_autoencoder/
MIT License
192 stars 39 forks source link

Make the activation store support multiple component dimensions #160

Closed alan-cooney closed 11 months ago

alan-cooney commented 11 months ago

Includes @Baidicoot code merged in from #51