Skip to content

Dataset · allenai

SWAG commonsense reasoning benchmark

allenai/swag

An English dataset for predicting logical continuations of everyday situations to test model commonsense.

What's inside

A large AllenAI English text corpus of 100k to 1M examples. Each record contains a situation description prefix and continuation options to test models' ability to reason based on everyday experience and real-world physics.

What you can do with it

  • Evaluating and fine-tuning language models on commonsense reasoning
  • Training classifiers to select the most plausible event progression
Tasks
Text classification
Languages
English
Size
100K–1M rows
Downloads this month
23.2K
Likes
25

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("allenai/swag")