Compare human critiques with LLM issues, add match notes, and export labels. To check the AI judges' matches instead, and see agreement statistics, use the Issue Match Audit.