Hardin FIlter

Hardin Filter serves as a critical verification system for responses from large language models, ensuring reliability and accuracy. Key features include: • Detection of fabricated citations and non-existent references • Identification of numerical inconsistencies and contradictions • Scrutiny for leaked data and outdated information • Generation of signed, verifiable checking receipts via TBN Protocol This tool operates quietly in the background, performing a suite of eight powerful checks every time a response is generated by platforms like ChatGPT, Claude, Gemini, or Copilot. It meticulously examines source attribution, assesses confidence against evidence, and performs logical consistency checks. Hardin Filter also flags information that might be time-sensitive and conducts leak detection to safeguard against exposed secrets or personal data. Users receive a clear, concise result indicating what passed, what failed, and what requires further evidence, empowering them to make informed decisions about the trustworthiness of the information presented. The verifiable receipt provides irrefutable proof of the checks performed, adding a layer of transparency and accountability. Hardin Filter is ideal for researchers, content creators, developers, and anyone who relies on information from advanced language models and needs to ensure its veracity. Available as a Chrome extension, VS Code extension, or a paste-in web tool, it provides an essential layer of scrutiny without requiring an account.