Show HN: HyperSAE – Hyperbolic Sparse Autoencoders for LLM Interpretability

 (github.com)

2 points | by visha1v 1 hour ago

0 comments