BoolQ: 15,942 Naturally Occurring Yes/No Questions Paired with Wikipedia Passages (Google Research)
- Provider
- Google Research (Clark et al.)
- License
- CC-BY-SA-3.0 (source)
- Size
- 4 files
- Format
- Parquet (in sampled files)
- Stored at
- Hugging Face (https://huggingface.co/datasets/google/boolq)
- Access
- Free and public. No account with the data provider is needed.
- Category
- Natural Language Processing
- Source
- https://arxiv.org/abs/1905.10044
About this dataset
BoolQ is a reading comprehension dataset of yes/no questions. The questions are naturally occurring: they come from anonymized, aggregated queries to the Google search engine that were issued in unprompted, unconstrained settings, and annotators paired each one with a Wikipedia passage and a true/false answer. Each example is a triplet of question, passage and answer, with the Wikipedia page title as optional context, so the task resembles natural language inference over a text pair. BoolQ is part of the SuperGLUE benchmark and is used to evaluate boolean question answering, textual entailment and inference in pretrained language models. It holds 15,942 examples: 9,427 training and 3,270 validation, stored as two Parquet files (about 5 MB) under data/. Load with the Hugging Face datasets library (load_dataset('google/boolq')) or pandas. Released by Google Research under Creative Commons Attribution-ShareAlike 3.0 (CC BY-SA 3.0); cite Clark et al., BoolQ: Exploring the Surprising Difficulty of Natural Yes/No Questions, NAACL 2019.
How to load this dataset in Python
from datasets import load_dataset
ds = load_dataset("google/boolq")
print(ds)Data structure
File types in sampled files: Parquet
Files
Example file: data/train-00000-of-00001.parquet
Analyze this dataset without downloading it
On Scigantic, BoolQ: 15,942 Naturally Occurring Yes/No Questions Paired with Wikipedia Passages (Google Research) opens in a hosted JupyterLab notebook with the data already mounted, so you can read files where they are stored instead of copying them first. Each notebook includes an AI assistant that can see the mounted files and write and run analysis code. Open the dataset page to launch a notebook, or browse the full catalog.
This page is also available as Markdown and as JSON.
Related Natural Language Processing datasets
- Nemotron RL Instruction Following Calendar V 2
- CantTalkAboutThis Topic Control Dataset
- Empathetic Dialogues
- Nemotron RL Instruction Following Structured Outputs
- ChatRAG Hi
- Panda
- Imppres
- DiscoFuse Wikipedia: 16 Million Sentence Fusion Examples with Discourse Connectives from English Wikipedia (Google Research)