What are Question Answering datasets used for?

Question Answering datasets are collections of labelled or raw data used to train, fine-tune, and evaluate models on the question answering task. This page lists 528 such datasets, each linking to its source and paper.

Which Question Answering dataset is best for benchmarking?

28 of these Question Answering datasets are tracked as benchmarks, including A Reading Comprehension Benchmark Requiring Discrete Reasoning Over Paragraphs (DROP). See the Benchmarks section for model leaderboards.

How many Question Answering datasets are there?

We catalog 528 Question Answering datasets in one searchable directory.

Question Answering Datasets

There are 528 question answering datasets in our directory, 28 of which are benchmarks. Each links to its source, paper, and download — browse the full list below or filter by language.

Question Answering is the task of returning a precise answer to a natural-language question, often grounded in a supporting passage. We catalog 528 datasets for it.

What languages do question answering datasets cover?

EN datasets (114)English datasets (56)ZH datasets (22)DE datasets (11)AR datasets (9)FR datasets (4)JA datasets (3)TR datasets (3)Chinese datasets (3)AF datasets (3)DA datasets (2)BN datasets (2)VI datasets (1)ID datasets (1)SU datasets (1)FA datasets (1)MULTILINGUAL datasets (1)Arabic datasets (1)

Explore other dataset tasks

General NLP(2546)Text Generation(723)Text Classification(228)Text Corpora(199)Image To Text(180)Visual Question Answering(130)Robotics(125)Image Classification(117)Text To Image(103)Summarization(98)Automatic Speech Recognition(93)Classification(87)

What languages do question answering datasets cover?

Explore other dataset tasks

Frequently asked questions

What are Question Answering datasets used for?

Which Question Answering dataset is best for benchmarking?

How many Question Answering datasets are there?