jbloomAus / SAELensLinks
Training Sparse Autoencoders on Language Models
☆974Updated 2 weeks ago
Alternatives and similar repositories for SAELens
Users that are interested in SAELens are comparing it to the libraries listed below
Sorting:
- Sparsify transformers with SAEs and transcoders☆629Updated this week
- The nnsight package enables interpreting and manipulating the internals of deep learned models.☆672Updated this week
- ☆345Updated last month
- Sparse Autoencoder for Mechanistic Interpretability☆267Updated last year
- Mechanistic Interpretability Visualizations using React☆289Updated 9 months ago
- ☆520Updated last year
- ☆242Updated 11 months ago
- Stanford NLP Python library for understanding and improving PyTorch models via interventions☆810Updated 3 weeks ago
- A library for mechanistic interpretability of GPT-style language models☆2,614Updated this week
- Using sparse coding to find distributed representations used by neural networks.☆269Updated last year
- Create feature-centric and prompt-centric visualizations for sparse autoencoders (like those from Anthropic's published research).☆220Updated 9 months ago
- ☆714Updated this week
- ☆124Updated last week
- Tools for understanding how transformer predictions are built layer-by-layer☆527Updated last month
- This repository collects all relevant resources about interpretability in LLMs☆372Updated 10 months ago
- ☆171Updated 10 months ago
- Representation Engineering: A Top-Down Approach to AI Transparency☆878Updated last year
- ☆127Updated last year
- Delphi was the home of a temple to Phoebus Apollo, which famously had the inscription, 'Know Thyself.' This library lets language models …☆211Updated last week
- ☆199Updated 10 months ago
- Resources for skilling up in AI alignment research engineering. Covers basics of deep learning, mechanistic interpretability, and RL.☆227Updated last month
- Code and results accompanying the paper "Refusal in Language Models Is Mediated by a Single Direction".☆272Updated 3 months ago
- ☆186Updated 2 months ago
- ViT Prisma is a mechanistic interpretability library for Vision and Video Transformers (ViTs).☆309Updated 2 months ago
- ☆54Updated 10 months ago
- Steering Llama 2 with Contrastive Activation Addition☆180Updated last year
- ☆276Updated last year
- ☆123Updated last year
- A toolkit for describing model features and intervening on those features to steer behavior.☆202Updated 10 months ago
- Decoder only transformer, built from scratch with PyTorch☆31Updated last year