ModelRefs / DROP — AI Glossary
DROP — AI Glossary
A reading comprehension benchmark requiring discrete reasoning operations (addition, counting, sorting) over paragraph text. F1-scored.
Overview
DROP (Dua et al. 2019) requires answering questions that cannot be answered by span extraction alone—they need arithmetic over numbers mentioned in the text, counting entity occurrences, or sorting dates. F1-scored. Frontier LLMs score >90%; tests quantitative reasoning in context beyond pure language understanding.
Reference details
| Topic | evaluation |
|---|---|
| Last reviewed | 2026-06-24 |
Related terms
Primary source
Continue your research
Use these connected ModelRefs sections to compare alternatives, inspect implementation paths, and review the evidence and governance boundaries relevant to DROP — AI Glossary.
Frequently asked questions
What is DROP?
A reading comprehension benchmark requiring discrete reasoning operations (addition, counting, sorting) over paragraph text.
What concepts are related to DROP?
Closely related concepts include math benchmark, multi document qa, reasoning.