Skip to main content
DatasetLLMsv1.0

BookCorpus

by University of Toronto · open-source · Last verified 2026-03-17

A dataset of over 11,000 unpublished books spanning fiction and non-fiction genres, originally scraped from Smashwords and used as the primary pretraining corpus for BERT alongside Wikipedia. It provides rich long-range dependency data that helps models learn coherent narrative structure and extended discourse patterns.

https://huggingface.co/datasets/bookcorpus
B+
B+Good
Adoption: AQuality: AFreshness: DCitations: A+Engagement: F

Specifications

License
Custom
Pricing
open-source
Capabilities
language-modeling, pretraining, long-range-understanding
Integrations
hugging-face, tensorflow-datasets
Use Cases
llm-pretraining, transfer-learning, research
API Available
Yes
Tags
nlp, books, long-form, pretraining, bert
Added
2026-03-17
Completeness
100%

Index Score

71.3
Adoption
82
Quality
80
Freshness
30
Citations
90
Engagement
0

Put AI to work for your business

Deploy this dataset alongside autonomous AaaS agents that handle tasks end-to-end — no babysitting required.

Explore the full AI ecosystem on Agents as a Service