Create feature-centric and prompt-centric visualizations for sparse autoencoders (like those from Anthropic's published research).
Primary LanguageHTMLMIT LicenseMIT
No issues in this repository yet.