GenAudit

A tool that assists humans in verifying factual errors in outputs from Large Language Models (LLMs) and provides supporting evidence.

CommonProductProductivityFact-checkingLLM Output Correction
GenAudit is a tool designed to help verify the responses of Large Language Models (LLMs) in document-supported tasks. It can suggest edits to LLM responses by correcting or removing claims not supported by the reference document and provide supporting evidence for seemingly corroborated facts. GenAudit achieves this by training models to perform these tasks and designing an interactive interface to present users with suggested edits and evidence.
Visit