New
decoderesearch/saelens v6.54.12 — Interpreting Language Model Internals with SAEs
decoderesearch/saelens is a Python library that trains and analyzes sparse autoencoders (SAEs) on language model activations to interpret internal representations.
decoderesearch/saelens
decoderesearch/saelens is a Python library that trains and analyzes sparse autoencoders (SAEs) on language model activations to interpret internal representations.
View on GitHub ↗- stars
- ★ 1.6K
- Forks
- 295
- Language
- Python
- License
- MIT
- Latest release
- v6.54.12
Posts
1No matching stories.