Joshua Engels

Scaling Laws For Scalable Oversight

Dense SAE Latents Are Features, Not Bugs

Low-Rank Adapting Models for Sparse Autoencoders

Not All Language Model Features Are One-Dimensionally Linear

Efficient Dictionary Learning with Switch Sparse Autoencoders

Are Sparse Autoencoders Useful? A Case Study in Sparse Probing

Decomposing The Dark Matter of Sparse Autoencoders

The Geometry of Concepts: Sparse Autoencoder Feature Structure