Multi-Layer Sparse Autoencoders for Transformer Interpretation
Failed to add items
Add to basket failed.
Add to Wish List failed.
Remove from Wish List failed.
Follow podcast failed
Unfollow podcast failed
Multi-Layer Sparse Autoencoders for Transformer Interpretation
-
Narrated by:
-
By:
About this listen
This paper introduces the Multi-Layer Sparse Autoencoder (MLSAE), a novel approach for interpreting the internal representations of transformer language models. Unlike traditional Sparse Autoencoders (SAEs) that analyze individual layers, MLSAEs are trained across all layers of a transformer's residual stream, enabling the study of information flow across layers. The research found that while individual "latents" (features learned by the SAE) tend to be active at a single layer for a given input, they are active at multiple layers when aggregated over many inputs, with this multi-layer activity increasing in larger models. The authors also explored the effect of "tuned-lens" transformations on latent activations, ultimately providing a new method for understanding how representations evolve within transformers.