Legibility Lab
This project examines how organisations become legible in the public record, and how that information is interpreted by large language models and by people.
Research papers explain the underlying ideas. Thinking Spaces are environments where these patterns can be inspected in practice.
Research papers
Research papers set out the arguments and observations developed through the project.
Each paper focuses on a different part of the system, while remaining readable on its own.
Thinking Spaces
Thinking Spaces are structured GPT environments configured to operate in thinking mode.
They are designed to inspect how responses form, change and stabilise within large language model interactions, as they are experienced by the user.
Legibility Lab 