Output Verification

Covers the mechanisms that confirm whether a model's live output matches defined correctness criteria, reference sets, evaluation checks, and routing policies. Applies to any issue addressing how AI output gets measured after deployment, not just whether the prompt was correctly specified.