Prompt-Response Quality Review Exercise (remote) - AI evaluation and rating QA
Reviewed a prompt-response dataset and assessed whether each AI-generated answer should be rated accurate or inaccurate based on evidence. Detected factual errors and corrected over-conservative ratings when answers were broadly accurate. Compiled QA summaries including totals reviewed and rationale/change logs for customer-facing review. • Prompt-response accuracy evaluation with evidence-based rating • Factual error identification and correction of ratings • Adjusted rating decisions based on guideline interpretation • QA documentation with rationale and change logs