callummcdougall / sae_vis
Create feature-centric and prompt-centric visualizations for sparse autoencoders (like those from Anthropic's published research).
☆184Updated 2 months ago
Alternatives and similar repositories for sae_vis:
Users that are interested in sae_vis are comparing it to the libraries listed below
- Mechanistic Interpretability Visualizations using React☆233Updated 2 months ago
- ☆120Updated last year
- Using sparse coding to find distributed representations used by neural networks.☆218Updated last year
- ☆210Updated 5 months ago
- ☆159Updated this week
- Sparse Autoencoder for Mechanistic Interpretability☆224Updated 7 months ago
- ☆251Updated last month
- Sparsify transformers with SAEs and transcoders☆485Updated this week
- ☆60Updated 3 months ago
- Training Sparse Autoencoders on Language Models☆649Updated this week
- ☆62Updated this week
- ☆26Updated 11 months ago
- ☆112Updated 7 months ago
- Steering vectors for transformer language models in Pytorch / Huggingface☆90Updated 2 weeks ago
- ☆149Updated this week
- Open source replication of Anthropic's Crosscoders for Model Diffing☆44Updated 4 months ago
- Erasing concepts from neural representations with provable guarantees☆225Updated last month
- The nnsight package enables interpreting and manipulating the internals of deep learned models.☆515Updated this week
- A library for efficient patching and automatic circuit discovery.☆56Updated 3 weeks ago
- ☆123Updated last month
- Steering Llama 2 with Contrastive Activation Addition☆125Updated 9 months ago
- Extract full next-token probabilities via language model APIs☆231Updated last year
- This repository collects all relevant resources about interpretability in LLMs☆324Updated 4 months ago
- ☆88Updated last month
- ☆88Updated last month
- Decoder only transformer, built from scratch with PyTorch☆26Updated last year
- ViT Prisma is a mechanistic interpretability library for Vision Transformers (ViTs).☆213Updated this week
- Sparse Autoencoder Training Library☆42Updated 4 months ago
- ☆435Updated 7 months ago
- An Open Source Implementation of Anthropic's Paper: "Towards Monosemanticity: Decomposing Language Models with Dictionary Learning"☆39Updated 10 months ago