Skip to content

Dataset · ehovy

RACE: Reading Comprehension Dataset from Exams

ehovy/race

Over 28k passages and nearly 100k multiple-choice questions from English exams in China.

What's inside

RACE is a large-scale dataset for evaluating reading comprehension, featuring over 28,000 passages and nearly 100,000 questions gathered from English exams for middle and high school students in China. The data is suitable for training and evaluating machine reading comprehension models.

What you can do with it

  • Training multiple-choice text comprehension models
  • Evaluating LLMs on examination questions
  • Fine-tuning LoRA for reading comprehension
Tasks
Multiple choice
Languages
English
Size
100K–1M rows
Downloads this month
27.1K
Likes
76

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("ehovy/race")