This project is a community-driven effort to reproduce, consolidate, and extend core visualizations from the Circuits thread, with an emphasis on activation maximization and dataset-driven neuron exemplars for the InceptionV1 model.
The objective is to develop a single, well-documented, open-source codebase that supports reliable reproduction and systematic exploration of these techniques. While the OpenAI Microscope previously provided access to many such visualizations across various vision models, it is no longer active, and there is currently no unified, end-to-end reference implementation that enables these analyses to be reproduced in a transparent and extensible manner.
By signing up, participants will collaborate on rebuilding the relevant tooling in the open, standardizing workflows, and making the underlying methodology explicit. Contributions will focus on creating reusable notebooks, clear documentation, and reproducible pipelines that support hands-on investigation of mechanistic interpretability, providing a practical foundation for further experimentation and extension.