Alternatives and similar repositories for sae:

Users that are interested in sae are comparing it to the libraries listed below

• jbloomAus / SAELens
  Training Sparse Autoencoders on Language Models
  ☆573Updated this week
• callummcdougall / sae_vis
  Create feature-centric and prompt-centric visualizations for sparse autoencoders (like those from Anthropic's published research).
  ☆176Updated last month
• ai-safety-foundation / sparse_autoencoder
  Sparse Autoencoder for Mechanistic Interpretability
  ☆209Updated 5 months ago
• saprmarks / dictionary_learning
  ☆179Updated this week
• HoagyC / sparse_coding
  Using sparse coding to find distributed representations used by neural networks.
  ☆207Updated last year
• EleutherAI / sae-auto-interp
  ☆135Updated this week
• TransformerLensOrg / CircuitsVis
  Mechanistic Interpretability Visualizations using React
  ☆219Updated 3 weeks ago
• openai / sparse_autoencoder
  ☆404Updated 5 months ago
• ArthurConmy / Automatic-Circuit-Discovery
  ☆201Updated 3 months ago
• neelnanda-io / 1L-Sparse-Autoencoder
  ☆114Updated last year
• ndif-team / nnsight
  The nnsight package enables interpreting and manipulating the internals of deep learned models.
  ☆458Updated this week
• steering-vectors / steering-vectors
  Steering vectors for transformer language models in Pytorch / Huggingface
  ☆78Updated last month
• saprmarks / feature-circuits
  ☆131Updated 3 months ago
• andyrdt / refusal_direction
  Code and results accompanying the paper "Refusal in Language Models Is Mediated by a Single Direction".
  ☆153Updated 3 months ago
• justinchiu / openlogprobs
  Extract full next-token probabilities via language model APIs
  ☆230Updated 10 months ago
• nrimsky / CAA
  Steering Llama 2 with Contrastive Activation Addition
  ☆113Updated 7 months ago
• ruizheliUOA / Awesome-Interpretability-in-Large-Language-Models
  This repository collects all relevant resources about interpretability in LLMs
  ☆305Updated 2 months ago
• OpenMOSS / Language-Model-SAEs
  For OpenMOSS Mechanistic Interpretability Team's Sparse Autoencoder (SAE) research.
  ☆82Updated this week
• google-deepmind / mishax
  ☆115Updated this week
• TransluceAI / observatory
  A toolkit for describing model features and intervening on those features to steer behavior.
  ☆149Updated 2 months ago
• AlignmentResearch / tuned-lens
  Tools for understanding how transformer predictions are built layer-by-layer
  ☆459Updated 7 months ago
• redwoodresearch / Easy-Transformer
  ☆106Updated 5 months ago
• GraySwanAI / circuit-breakers
  Improving Alignment and Robustness with Circuit Breakers
  ☆174Updated 3 months ago
• mcleish7 / arithmetic
  Code to reproduce "Transformers Can Do Arithmetic with the Right Embeddings", McLeish et al (NeurIPS 2024)
  ☆182Updated 7 months ago
• EleutherAI / concept-erasure
  Erasing concepts from neural representations with provable guarantees
  ☆219Updated last month
• davidbau / baukit
  ☆184Updated 10 months ago
• HazyResearch / zoology
  Understand and test language model architectures on synthetic tasks.
  ☆175Updated this week
• collin-burns / discovering_latent_knowledge
  ☆258Updated 10 months ago
• jacobdunefsky / transcoder_circuits
  ☆53Updated 2 months ago