Purpose and ownership
LLMpeople is an independent public atlas created and maintained by Keye Zheng. It helps readers discover researchers, technical reports, and organizations in the language-model ecosystem. It is not an official directory of the organizations it covers, and listing a person does not imply endorsement or a commercial relationship.
The project combines software engineering and AI-assisted maintenance. Automation helps collect candidate sources and maintain the product, but a generated description is not itself evidence. Public facts should be traceable to identifiable sources. The directory is a reference for discovery, not an authoritative employment register.
What a connection means
A report connection generally means authorship or a recorded public relationship to that work. It does not establish individual credit for every method in the report. An organization connection may reflect public research history, report affiliation, or current work; readers should use dated primary sources to confirm a current role.
Profiles can be incomplete or stale. A completeness score describes available fields, not a probability that every statement is true. Source links and review dates help readers investigate, but uncertainty should remain visible rather than being filled with speculation.
Sources, corrections, and AI assistance
Preferred evidence includes official project pages, personal homepages, technical reports, repositories, and attributable public presentations. Secondary coverage can supply context, but identity and affiliation need care, particularly for common names. The site aims to distinguish public evidence from inference and does not seek private contact or family information.
Use Suggest a correction or Suggest a source on a researcher page to flag a problem. Include the disputed statement, a reliable source URL, and the relevant date. Contributions enter a review process rather than becoming facts automatically. Site issues can also be reported through the public GitHub repository linked below. Please avoid including private or sensitive personal information.
Advertising and sustainability
Advertising may help cover infrastructure and maintenance costs. It must remain separate from editorial judgments: payment does not buy researcher coverage, change attribution, or determine the description of a technical contribution. The presence of advertising code does not mean a site has earned revenue or covered its costs.
The sustainability experiment will be judged on actual expenses, revenue, maintenance effort, and usefulness to readers. The Privacy and Terms pages describe the site's data practices and usage conditions. External source websites apply their own policies.