Alternatives and similar repositories for concept-erasure

Users that are interested in concept-erasure are comparing it to the libraries listed below

• EleutherAI / elk
  Keeping language models honest by directly eliciting knowledge encoded in their activations.
  ☆207Updated last week
• callummcdougall / sae_vis
  Create feature-centric and prompt-centric visualizations for sparse autoencoders (like those from Anthropic's published research).
  ☆206Updated 7 months ago
• justinchiu / openlogprobs
  Extract full next-token probabilities via language model APIs
  ☆247Updated last year
• TransformerLensOrg / CircuitsVis
  Mechanistic Interpretability Visualizations using React
  ☆260Updated 6 months ago
• AlignmentResearch / tuned-lens
  Tools for understanding how transformer predictions are built layer-by-layer
  ☆505Updated last year
• neelnanda-io / 1L-Sparse-Autoencoder
  ☆122Updated last year
• nostalgebraist / transformer-utils
  Utilities for the HuggingFace transformers library
  ☆68Updated 2 years ago
• EleutherAI / delphi
  Delphi was the home of a temple to Phoebus Apollo, which famously had the inscription, 'Know Thyself.' This library lets language models …
  ☆192Updated last week
• collin-burns / discovering_latent_knowledge
  ☆273Updated last year
• steering-vectors / steering-vectors
  Steering vectors for transformer language models in Pytorch / Huggingface
  ☆112Updated 4 months ago
• ArthurConmy / Automatic-Circuit-Discovery
  ☆231Updated 9 months ago
• callummcdougall / sae_visualizer
  ☆28Updated last year
• HazyResearch / TART
  TART: A plug-and-play Transformer module for task-agnostic reasoning
  ☆200Updated 2 years ago
• KihoPark / LLM_Categorical_Hierarchical_Representations
  ☆101Updated 5 months ago
• mcleish7 / arithmetic
  Code to reproduce "Transformers Can Do Arithmetic with the Right Embeddings", McLeish et al (NeurIPS 2024)
  ☆190Updated last year
• r-three / git-theta
  git extension for {collaborative, communal, continual} model development
  ☆214Updated 8 months ago
• google-deepmind / mishax
  ☆134Updated 3 months ago
• evandez / relations
  How do transformer LMs encode relations?
  ☆50Updated last year
• likenneth / othello_world
  Emergent world representations: Exploring a sequence model trained on a synthetic task
  ☆182Updated 2 years ago
• apartresearch / interpretability-starter
  🧠 Starter templates for doing interpretability research
  ☆72Updated 2 years ago
• ApolloResearch / e2e_sae
  Sparse Autoencoder Training Library
  ☆53Updated 2 months ago
• UFO-101 / auto-circuit
  A library for efficient patching and automatic circuit discovery.
  ☆70Updated 2 months ago
• jessicarumbelow / Backwards
  ☆82Updated last year
• HazyResearch / zoology
  Understand and test language model architectures on synthetic tasks.
  ☆219Updated last month
• Aaquib111 / edge-attribution-patching
  Code for my NeurIPS 2024 ATTRIB paper titled "Attribution Patching Outperforms Automated Circuit Discovery"
  ☆40Updated last year
• redwoodresearch / Easy-Transformer
  ☆121Updated 11 months ago
• anthropics / toy-models-of-superposition
  Notebooks accompanying Anthropic's "Toy Models of Superposition" paper
  ☆127Updated 2 years ago
• TomFrederik / unseal
  Mechanistic Interpretability for Transformer Models
  ☆51Updated 3 years ago
• adamkarvonen / SAEBench
  ☆105Updated last month
• kernelmachine / cbtm
  Code repository for the c-BTM paper
  ☆106Updated last year