andyrdt / refusal_directionLinks
Code and results accompanying the paper "Refusal in Language Models Is Mediated by a Single Direction".
☆233Updated last week
Alternatives and similar repositories for refusal_direction
Users that are interested in refusal_direction are comparing it to the libraries listed below
Sorting:
- Steering Llama 2 with Contrastive Activation Addition☆158Updated last year
- Steering vectors for transformer language models in Pytorch / Huggingface☆108Updated 4 months ago
- Delphi was the home of a temple to Phoebus Apollo, which famously had the inscription, 'Know Thyself.' This library lets language models …☆185Updated this week
- ☆131Updated 7 months ago
- Improving Alignment and Robustness with Circuit Breakers☆214Updated 9 months ago
- For OpenMOSS Mechanistic Interpretability Team's Sparse Autoencoder (SAE) research.☆121Updated this week
- Using sparse coding to find distributed representations used by neural networks.☆255Updated last year
- General-purpose activation steering library☆78Updated last month
- ☆173Updated 2 months ago
- LLM experiments done during SERI MATS - focusing on activation steering / interpreting activation spaces☆94Updated last year
- Function Vectors in Large Language Models (ICLR 2024)☆170Updated 2 months ago
- Röttger et al. (NAACL 2024): "XSTest: A Test Suite for Identifying Exaggerated Safety Behaviours in Large Language Models"☆98Updated 4 months ago
- Create feature-centric and prompt-centric visualizations for sparse autoencoders (like those from Anthropic's published research).☆202Updated 6 months ago
- A toolkit for describing model features and intervening on those features to steer behavior.☆190Updated 7 months ago
- Sparsify transformers with SAEs and transcoders☆568Updated last week
- Stanford NLP Python library for benchmarking the utility of LLM interpretability methods☆95Updated 2 weeks ago
- Code for In-context Vectors: Making In Context Learning More Effective and Controllable Through Latent Space Steering☆177Updated 4 months ago
- Sparse Autoencoder for Mechanistic Interpretability☆250Updated 11 months ago
- ☆101Updated 3 weeks ago
- Mechanistic Interpretability Visualizations using React