HN Trends Tracker

HN Trends Tracker

backend ok

Repost history

Canonical URL: https://github.com/vishal-dehurdle/hypersae

Current story

#
Show HN: LLMs think in hierarchies. Their interpretability tools should too(github.com)
HN #49,312,1692026-08-15 16:54 UTC
3 pointsby visha1v29 days ago0 comments

Prior reposts (2 of 5 total)

#
Show HN: HyperSAE – Hyperbolic Sparse Autoencoders for LLM Interpretability(github.com)
HN #49,262,3402026-08-11 18:19 UTC
2 pointsby visha1v33 days ago0 comments
#
Show HN: At 16K features, flat autoencoders break. Curved space doesn't(github.com)
HN #49,288,1982026-08-13 16:16 UTC
7 pointsby visha1v31 days ago0 comments