Show HN: Subtext visualizes LLM reasoning

A new open‑source project called Subtext visualizes the internal reasoning of large language models. The repository is hosted on GitHub under the user ninjahawk. The tool renders thought processes as interactive diagrams. It helps developers see how an LLM arrives at its outputs. The visualization can be used for debugging and research purposes. The project is presented as a Show HN entry for community feedback. Users can clone the repo and run the visualization on their own models. The author invites contributions to expand its capabilities.