lyramakesmusic/activations-vis

HTML15 stars0 forks

What it does

The SAE Feature Explorer is an interactive tool that visualizes Sparse Autoencoder features, allowing users to explore token activations in a 2D/3D space. It provides insights into model behavior and feature relevance, making it valuable for researchers and developers working with machine learning models.

Star history

Not enough history yet — 1 day(s) recorded. The daily snapshot builds this up.

Tracking

Last trending: 2026-02-21

Creator kit

Hook

Discover how to visualize and interact with Sparse Autoencoder features in a stunning 3D starfield!

Content angles

Create a tutorial on setting up the SAE Feature Explorer and visualizing model activations.
Develop a video walkthrough demonstrating the interactive features and capabilities of the tool.
Write an article comparing different methods of visualizing machine learning model features and highlight the advantages of this tool.

Who should care

AI researchers, machine learning developers, and data scientists interested in feature visualization.