Skip to content

Dataset · stanfordnlp

SNLI: sentence pairs for NLI

stanfordnlp/snli

570k human-written English sentence pairs with entailment, contradiction, and neutral labels.

What's inside

SNLI corpus version 1.0 of 570k English sentence pairs. Humans manually labeled balanced classifications: entailment, contradiction, neutral. Supports the natural language inference task, also known as recognizing textual entailment.

What you can do with it

  • Training an NLI classifier
  • Fine-tuning transformers for textual entailment
  • Logical reasoning benchmark
Tasks
Text classification
Languages
English
Size
100K–1M rows
Downloads this month
27.8K
Likes
101

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("stanfordnlp/snli")