Skip to content

Dataset · rajpurkar

SQuAD: Wikipedia questions

rajpurkar/squad

Reading comprehension QA dataset: questions on Wikipedia articles with span answers.

What's inside

SQuAD is a reading comprehension dataset: crowdworkers asked questions about Wikipedia articles, where the answer is a text span from the article or unanswerable. SQuAD 1.1 contains 100,000+ question-answer pairs across 500+ articles. Language is English, modality is text.

What you can do with it

  • Training extractive QA models
  • Evaluating LLMs and RAG systems
  • Fine-tuning BERT-like models
Tasks
Question answering
Languages
English
Size
10K–100K rows
Downloads this month
158.5K
Likes
892

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("rajpurkar/squad")