amack315 / unsupervised-steering-vectorsLinks
☆36Updated last year
Alternatives and similar repositories for unsupervised-steering-vectors
Users that are interested in unsupervised-steering-vectors are comparing it to the libraries listed below
Sorting:
- Open source replication of Anthropic's Crosscoders for Model Diffing☆59Updated last year
- Sparse Autoencoder Training Library☆55Updated 5 months ago
- Delphi was the home of a temple to Phoebus Apollo, which famously had the inscription, 'Know Thyself.' This library lets language models …☆219Updated last week
- Attribution-based Parameter Decomposition☆31Updated 4 months ago
- Create feature-centric and prompt-centric visualizations for sparse autoencoders (like those from Anthropic's published research).☆222Updated 10 months ago
- ☆74Updated 2 weeks ago
- ☆54Updated 11 months ago
- A tiny easily hackable implementation of a feature dashboard.☆15Updated last month
- A library for efficient patching and automatic circuit discovery.☆78Updated 3 months ago
- Investigating the generalization behavior of LM probes trained to predict truth labels: (1) from one annotator to another, and (2) from e…☆28Updated last year
- Applying SAEs for fine-grained control☆24Updated 10 months ago
- ☆128Updated last year
- Steering vectors for transformer language models in Pytorch / Huggingface☆126Updated 8 months ago
- ☆191Updated last week
- Improving Steering Vectors by Targeting Sparse Autoencoder Features☆25Updated 11 months ago
- ☆247Updated last year
- Code for reproducing our paper "Not All Language Model Features Are Linear"☆82Updated 11 months ago
- A collection of different ways to implement accessing and modifying internal model activations for LLMs☆19Updated last year
- ☆29Updated last year
- ☆30Updated last year
- Unified access to Large Language Model modules using NNsight☆52Updated this week
- ☆142Updated last month
- Engine for collecting, uploading, and downloading model activations☆24Updated 6 months ago
- Inference API for many LLMs and other useful tools for empirical research☆77Updated last week
- Code for my NeurIPS 2024 ATTRIB paper titled "Attribution Patching Outperforms Automated Circuit Discovery"☆43Updated last year
- ☆16Updated 2 weeks ago
- Tools for studying developmental interpretability in neural networks.☆111Updated 4 months ago
- ☆181Updated 11 months ago
- Code to reproduce key results accompanying "SAEs (usually) Transfer Between Base and Chat Models"☆12Updated last year
- A library for mechanistic anomaly detection☆22Updated 9 months ago