Cadenza-Labs / sleeper-agents

Related projects: ⓘ

• TransformerLensOrg / CircuitsVis
  Mechanistic Interpretability Visualizations using React
  ☆175Updated 2 months ago
• steering-vectors / steering-vectors
  Steering vectors for transformer language models in Pytorch / Huggingface
  ☆52Updated 2 months ago
• callummcdougall / sae_visualizer
  ☆24Updated 5 months ago
• saprmarks / dictionary_learning
  ☆110Updated 3 weeks ago
• andyrdt / refusal_direction
  Code and results accompanying the paper "Refusal in Language Models Is Mediated by a Single Direction".
  ☆76Updated 3 weeks ago
• METR / task-standard
  METR Task Standard
  ☆111Updated last week
• nrimsky / CAA
  Steering Llama 2 with Contrastive Activation Addition
  ☆83Updated 3 months ago
• nrimsky / LM-exp
  LLM experiments done during SERI MATS - focusing on activation steering / interpreting activation spaces
  ☆73Updated last year
• HoagyC / sparse_coding
  Using sparse coding to find distributed representations used by neural networks.
  ☆162Updated 10 months ago
• callummcdougall / sae_vis
  Create feature-centric and prompt-centric visualizations for sparse autoencoders (like those from Anthropic's published research).
  ☆133Updated last month
• saprmarks / feature-circuits
  ☆75Updated this week
• amack315 / unsupervised-steering-vectors
  ☆19Updated 4 months ago
• METR / vivaria
  Vivaria is METR's tool for running evaluations and conducting agent elicitation research.
  ☆50Updated this week
• ArthurConmy / Automatic-Circuit-Discovery
  ☆174Updated 4 months ago
• LRudL / evalugator
  (Model-written) LLM evals library
  ☆14Updated last month
• timaeus-research / devinterp
  Tools for studying developmental interpretability in neural networks.
  ☆69Updated this week
• callummcdougall / ARENA_2.0
  Resources for skilling up in AI alignment research engineering. Covers basics of deep learning, mechanistic interpretability, and RL.
  ☆185Updated 7 months ago
• jbloomAus / SAELens
  Training Sparse Autoencoders on Language Models
  ☆367Updated this week
• ndif-team / nnsight
  The nnsight package enables interpreting and manipulating the internals of deep learned models.
  ☆356Updated this week
• ai-safety-foundation / sparse_autoencoder
  Sparse Autoencoder for Mechanistic Interpretability
  ☆173Updated 2 months ago
• neelnanda-io / 1L-Sparse-Autoencoder
  ☆99Updated 10 months ago
• METR / public-tasks
  ☆58Updated last week
• Aaquib111 / edge-attribution-patching
  Code for my NeurIPS 2024 ATTRIB paper titled "Attribution Patching Outperforms Automated Circuit Discovery"
  ☆22Updated 3 months ago
• EleutherAI / sae
  Sparse autoencoders
  ☆297Updated last week
• KoyenaPal / future-lens
  Code and Data Repo for the CoNLL Paper -- Future Lens: Anticipating Subsequent Tokens from a Single Hidden State
  ☆14Updated 8 months ago
• EleutherAI / sae-auto-interp
  ☆54Updated last week
• GraySwanAI / circuit-breakers
  Improving Alignment and Robustness with Circuit Breakers
  ☆124Updated 2 months ago
• alan-cooney / transformer-from-scratch
  Decoder only transformer, built from scratch with PyTorch
  ☆25Updated 10 months ago
• AsaCooperStickland / situational-awareness-evals
  Measuring the situational awareness of language models
  ☆31Updated 7 months ago
• ApolloResearch / e2e_sae
  Sparse Autoencoder Training Library
  ☆18Updated last month