Skip to content

Dataset · allenai

AI2 ARC: Science Questions

allenai/ai2_arc

7,887 multiple-choice science questions for evaluating question-answering models.

What's inside

The dataset contains 7,887 genuine grade-school science multiple-choice questions split into a Challenge Set and an Easy Set. The Challenge Set includes questions that both a retrieval algorithm and a word co-occurrence algorithm failed to answer correctly. Additionally, a corpus of over 14 million science sentences is included. Language is English, modality is text.

What you can do with it

  • Evaluating QA models on science questions
  • Fine-tuning LLMs for multiple-choice reasoning
  • Testing retrieval systems on the science sentence corpus
Tasks
Question answering
Languages
English
Size
1K–10K rows
Downloads this month
545.4K
Likes
433

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("allenai/ai2_arc")