ModelRefs / Exact Match (EM) — AI Glossary
Exact Match (EM) — AI Glossary
An evaluation metric scoring 1 if the model's output exactly matches the reference string after normalization, 0 otherwise. Strict but interpretable.
Overview
Exact match is used for QA tasks (SQuAD, Natural Questions) with short factual answers. Normalization: lowercase, punctuation removal, article removal. Strict but interpretable. EM underestimates quality for paraphrastic correct answers; F1-based token overlap is used alongside EM for most QA benchmarks.
Reference details
| Topic | evaluation |
|---|---|
| Last reviewed | 2026-06-24 |
Related terms
Continue your research
Use these connected ModelRefs sections to compare alternatives, inspect implementation paths, and review the evidence and governance boundaries relevant to Exact Match (EM) — AI Glossary.
Frequently asked questions
What is Exact Match (EM)?
An evaluation metric scoring 1 if the model's output exactly matches the reference string after normalization, 0 otherwise.
What concepts are related to Exact Match (EM)?
Closely related concepts include bleu, rouge, ragas.