Show HN: HyperSAE – Hyperbolic Sparse Autoencoders for LLM Interpretability

3 pointsposted 6 hours ago
by visha1v
(github.com)

No comments yet