</> Code with Papers by Syntology

Datasets › XNLI

XNLI (Cross-lingual Natural Language Inference)

Introduced by Alexis Conneau et al. in XNLI: Evaluating Cross-lingual Sentence Representations1 Jan 2018 archive 2025-07-28

The Cross-lingual Natural Language Inference (XNLI) corpus is the extension of the Multi-Genre NLI (MultiNLI) corpus to 15 languages. The dataset was created by manually translating the validation and test sets of MultiNLI into each of those 15 languages. The English training set was machine translated for all languages. The dataset is composed of 122k train, 2490 validation and 5010 test examples.

Source: CamemBERT: a Tasty French Language Model Image Source: https://github.com/facebookresearch/XNLI

HomepageBenchmarksPapers JSON

Benchmarks archive 2025-07-28

All 7 leaderboards whose dataset resolves to this page shown (sort by any header). "First row" is the archive's own first row at snapshot, in the archive's row order; nothing here re-ranks and metric direction is not asserted.

First row (archive order)PaperCode
Natural Language Inference XNLI French CamemBERT (large) Accuracy 85.7 CamemBERT: a Tasty French Language Model huggingface/transformers +7 6 Compare
Cross-Lingual Natural Language Inference XNLI ByT5 XXL Accuracy 83.7 ByT5: Towards a token-free future with pre-trained... huggingface/transformers +4 5 Compare
Cross-Lingual Natural Language Inference XNLI Zero-Shot English-to-Spanish XLM-R R4F Accuracy 85.2% Better Fine-Tuning by Reducing Representational Collapse pytorch/fairseq +2 4 Compare
Cross-Lingual Natural Language Inference XNLI Zero-Shot English-to-German XLM-R R4F Accuracy 84.2% Better Fine-Tuning by Reducing Representational Collapse pytorch/fairseq +2 4 Compare
Cross-Lingual Natural Language Inference XNLI Zero-Shot English-to-French XLM-R R4F Accuracy 84.7% Better Fine-Tuning by Reducing Representational Collapse pytorch/fairseq +2 3 Compare
Natural Language Inference XNLI Chinese Dev ERNIE 2.0 Large Accuracy 82.6 ERNIE 2.0: A Continual Pre-training Framework for... PaddlePaddle/PaddleNLP +2 3 Compare
Natural Language Inference XNLI Chinese ERNIE 2.0 Large Accuracy 81 ERNIE 2.0: A Continual Pre-training Framework for... PaddlePaddle/PaddleNLP +2 3 Compare

Papers archive 2025-07-28

12 shown of 12 papers with a leaderboard row on this dataset's benchmarks, newest first. The archive's own "papers using this dataset" list was never published, so this is the benchmark-backed subset; the archive's count for this dataset is 349. The Syntology column is from Syntology's graph (read 2026-09-24), stated per sample; it is not part of any archive number.

DateSamples run Syntology
mGPT: Few-Shot Learners Go Multilingual 1 1 15 Apr 2022 not harvested
ByT5: Towards a token-free future with pre-trained byte-to-byte models 5 2 28 May 2021 ran 0 of 6 samples (6 unverified)
Rethinking embedding coupling in pre-trained language models 4 2 24 Oct 2020 not harvested
Better Fine-Tuning by Reducing Representational Collapse 3 3 6 Aug 2020 not harvested
FlauBERT: Unsupervised Language Model Pre-training for French 7 2 11 Dec 2019 ran 1 of 9 samples (8 unverified)
CamemBERT: a Tasty French Language Model 8 2 10 Nov 2019 not harvested
ERNIE 2.0: A Continual Pre-training Framework for Language Understanding 3 4 29 Jul 2019 ran 0 of 1 samples (1 unverified; 1 pointer-only for licence)
ERNIE: Enhanced Representation through Knowledge Integration 19 2 19 Apr 2019 ran 0 of 7 samples (7 unverified)
Cross-lingual Language Model Pretraining 17 1 22 Jan 2019 ran 1 of 7 samples (6 unverified; 1 pointer-only for licence)
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding 534 2 11 Oct 2018 ran 204 of 659 samples (455 unverified; 149 pointer-only for licence)
XNLI: Evaluating Cross-lingual Sentence Representations 9 1 13 Sep 2018 not harvested
Supervised Learning of Universal Sentence Representations from Natural Language Inference Data 23 6 5 May 2017 ran 6 of 7 samples (1 unverified; 7 pointer-only for licence)

Dataset loaders archive 2025-07-28

huggingface.co/datasets/juletxara/xnli_mthuggingface/datasetstfpytorchjax
huggingface.co/datasets/xnlihuggingface/datasetstfpytorchjax
huggingface.co/datasets/mteb/xnlihuggingface/datasetstfpytorchjax
huggingface.co/datasets/facebook/xnlihuggingface/datasetstfpytorchjax
huggingface.co/datasets/Mathoctopus/msvamphuggingface/datasetstfpytorchjax
www.tensorflow.org/datasets/catalog/xnlitensorflow/datasetstfjax
facebookresearch/XNLI

7 loaders as listed in the archive; links are outbound and not re-checked here.

Tasks archive 2025-07-28

Natural Language InferenceChinese Sentence Pair ClassificationFLUECross-Lingual Natural Language Inference

License archive 2025-07-28

Attribution-NonCommercial 4.0 International

Modalities archive 2025-07-28

Texts

Languages archive 2025-07-28

FrenchSpanishGermanChineseMultilingualRussianArabicBulgarianHindiThaiTurkishUrduVietnameseGreekSwahili

Variants archive 2025-07-28

  • XNLI Zero-Shot English-to-French
  • XNLI French
  • XNLI Dev
  • XNLI Chinese
  • XNLI Chinese Dev
  • XNLI Zero-Shot English-to-German
  • XNLI Zero-Shot English-to-Spanish
  • XNLI

8 variant names, as the archive lists them.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections