◈ Explore
← Back to explorer

e-SNLI (presencesw re-upload)

Oana-Maria Camburu, Tim Rocktäschel, Thomas Lukasiewicz, Phil Blunsom (original), presencesw / Datnvt (re-upload)

Open source
Type
dataset
Venue
HuggingFace (presencesw); original from NeurIPS 2018
Year
2018
Source
huggingface
Access
restricted
Language
English
Added
2026-07-17T20:18:03.711711+00:00
Verified
2026-07-17T20:18:03.711711+00:00

Summary

This is a HuggingFace re-upload (by presencesw/Datnvt) of the e-SNLI dataset, which extends the Stanford Natural Language Inference (SNLI) dataset with human-annotated natural language explanations of the entailment relations. The original dataset contains ~549k train / 9.8k validation / 9.8k test examples, each with a premise, hypothesis, label (entailment/neutral/contradiction), and multiple explanation annotations. It is a foundational dataset for explainable NLI.

Keywords

nli natural-language-inference explanations snli entailment explainable-ai

Topics

NLP / Natural Language Inference

Research notes

  • The presencesw/esnli page returned HTTP 401 (gated/restricted access). Information synthesized from the official esnli/esnli dataset card and the original e-SNLI GitHub repo (OanaMariaCamburu/e-SNLI). presencesw (Datnvt) maintains 101+ dataset re-uploads on HuggingFace.