datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
logiqa-2.0-nlihttps://github.com/csitfun/LogiQA2.0
Temporary citation:
@article{liu2020logiqa,
title={Logiqa: A challenge dataset for machine reading comprehension with logical reasoning},
author={Liu, Jian and Cui, Leyang and Liu, Hanmeng and Huang, Dandan and Wang, Yile and Zhang, Yue},
journal={arXiv preprint arXiv:2007.08124},
year={2020}
}
ConTRoL-nlihttps://github.com/csitfun/ConTRoL-dataset
@article{Liu_Cui_Liu_Zhang_2021,
title={Natural Language Inference in Context - Investigating Contextual Reasoning over Long Texts},
volume={35},
url={https://ojs.aaai.org/index.php/AAAI/article/view/17580},
DOI={10.1609/aaai.v35i15.17580},
number={15},
journal={Proceedings of the AAAI Conference on Artificial Intelligence},
author={Liu, Hanmeng and Cui, Leyang and Liu, Jian and Zhang, Yue},
year={2021},
month={May},
pages={13388-13396}
}
adversarial_nliDIWALI
DIWALI - Diversity and Inclusivity aWare cuLture specific Items for India: Dataset and Assessment of LLMs for Cultural Text Adaptation in Indian Context
Paper | Code | Project page
We present a novel Culture Specific Items (CSIs) dataset for Indian culture covering 17 facets. Please refer to our project pagefor quick details.
Facets considered: food, dance, festivals, names, jewellery, places, traditions, languages, clothing, games, rituals, architectures, drinks, arts… See the full description on the dataset page: https://huggingface.co/datasets/nlip/DIWALI.beirnli-esannotations_creators:
crowdsourced
other
language_creators:
other
crowdsourced
languages:
es
licenses:
cc-by-sa-4.0
multilinguality:
monolingual
pretty_name: ESnli
size_categories:
unknown
source_datasets:
extended|snli
extended|xnli
extended|multi_nli
task_categories:
text-classification
task_ids:
natural-language-inference
Dataset Card for nli-es
Dataset Summary
A Spanish Natural Language Inference dataset put together from the sources:
the Spanish slice of the XNLI… See the full description on the dataset page: https://huggingface.co/datasets/somosnlp-hackathon-2022/nli-es.dadc-limit-nlihttps://github.com/facebookresearch/dadc-limit
@inproceedings{Wallace2022Dynamic,
Title = {Analyzing Dynamic Adversarial Training Data in the Limit},
Author = {Eric Wallace and Adina Williams and Robin Jia and Douwe Kiela},
Booktitle={Findings of the Association for Computational Linguistics},
Year = {2022}
}
csfever_nliSmaller version (2/3) of dataset (test set) subsampled from the original ctu-aic/csfever_nli.
Citation
If you are using the dataset, please cite the original authors.
@article{ullrich2023csfever,
title={CsFEVER and CTKFacts: acquiring Czech data for fact verification},
author={Ullrich, Herbert and Drchal, Jan and R{\`y}par, Martin and Vincourov{\'a}, Hana and Moravec, V{\'a}clav},
journal={Language Resources and Evaluation},
volume={57},
number={4},
pages={1571--1605}… See the full description on the dataset page: https://huggingface.co/datasets/CZLC/csfever_nli.popia-compliance-nli
POPIA Compliance NLI
Hand-authored Natural Language Inference pairs for South Africa's Protection
of Personal Information Act (POPIA). Each example is a (premise, hypothesis,
label) triple covering one of 7 canonical POPIA clauses. Designed for
fine-tuning small cross-encoder NLI models into deterministic, local,
auditable compliance judges — the alternative to prompting an LLM-as-judge
over personal information.
Used to fine-tune labrat-aiko/nli-popia-v1,
which reaches macro F1… See the full description on the dataset page: https://huggingface.co/datasets/labrat-aiko/popia-compliance-nli.high-quality-nli
High-Quality NLI Dataset
Dataset Summary
This dataset is designed for Natural Language Inference (NLI) tasks, containing high-quality sentence pairs. It improves upon commonly used NLI datasets by offering more complex and nuanced examples, making it suitable for advanced language understanding models.
Dataset Details
Dataset Statistics
Train set size: 550 970
Test set size: 137 743
Total size: 688 713
Class Distribution
Label… See the full description on the dataset page: https://huggingface.co/datasets/agentlans/high-quality-nli.finewebedu-nli
Dataset Card for Fineweb-edu NLI Dataset
Dataset Description
The Fineweb-edu NLI Dataset consists of pairs of sentences, where each pair includes a "premise" and a "hypothesis." The goal is to classify the relationship between the premise and hypothesis into three categories:
Label 0: The hypothesis is supported by the premise (entailment).
Label 1: The hypothesis is related but not directly supported by the premise (neutral).
Label 2: The hypothesis contradicts the… See the full description on the dataset page: https://huggingface.co/datasets/agentlans/finewebedu-nli.liquid-nli-domain-1krole-playing-apc-nlinli-instructionarabic-embedding-dataset-nli
Arabic Legal & Semantic Embedding Dataset
Overview
The Arabic Legal Embedding Dataset is a high-quality dataset designed for:
Semantic Similarity
Natural Language Inference (NLI)
Sentence Embedding Training
It focuses on Arabic legal, administrative, and general-purpose text, enabling the development of models that understand nuanced meanings in real-world Arabic language scenarios.
Each example consists of two sentences (sentence_a, sentence_b) with:
A similarity… See the full description on the dataset page: https://huggingface.co/datasets/ziaddddd/arabic-embedding-dataset-nli.contrast_nliscifact
SciFact
NLI-Mixsynthetic-nli-triplet
Overview
Synthetic-nli-triplet is a dataset generated through an iterative sampling process from the existing nli-triplet dataset.
Each sampled triplet is used as input to prompt the TheBloke/Mistral-7B-Instruct-v0.1-GPTQ model,
in generating new and synthetic triplet samples, further expanding the data for tasks involving Natural Language Inference (NLI).
How to use
from datasets import load_dataset
my_dataset = load_dataset("lxyuan/synthetic-nli-triplet")
>>>… See the full description on the dataset page: https://huggingface.co/datasets/lxyuan/synthetic-nli-triplet.MedQA-NLI
MedQA-NLI: A Large-Scale Multi-Source Medical Question Answering and Natural Language Inference Dataset
MedQA-NLI is a comprehensive medical reasoning dataset comprising 42,889 instances designed for training and evaluating models on natural language inference (NLI) tasks in biomedical domains.
Each instance pairs a PubMed abstract with a question and a claim, enabling models to perform:
Medical question answering
Claim (answer) verification
Textual entailment recognition
The… See the full description on the dataset page: https://huggingface.co/datasets/mehdinick/MedQA-NLI.MNLI-NLI
Glue MNLI
This dataset is a port of the official mnli dataset on the Hub.
It contains the matched version.
Note that the premise and hypothesis columns have been renamed to text1 and text2 respectively.
Also, the test split is not labeled; the label column values are always -1.
NLIIOAI_NLP_Training_01_symbolic_nli_private
Symbolic NLI Private Answers
Instructor-only answer key. Upload this directory only to a private Hugging
Face dataset repository.
facility-support-analyzer
Facility Support Analyzer (Llama Prompt Ops)
This dataset mirrors the Facility Support Analyzer dataset used in the Llama Prompt Ops tutorial. Each example is a facility-related message with gold labels for urgency, sentiment, and a set of service categories.
Tasks
Urgency (single label): low | medium | high
Sentiment (single label): positive | neutral | negative
Categories (multi-label): one or more from:
emergency_repair_services
routine_maintenance_requests… See the full description on the dataset page: https://huggingface.co/datasets/nlile/facility-support-analyzer.IOAI_NLP_Training_01_symbolic_nli_public
Symbolic NLI Exercise Inputs
Public evaluation prompts for the Symbolic NLI notebook. Labels are stored in
the separate private instructor payload.
flattened_results_2024_10_09_161002NLI-xnli-viflattened_results_2024_10_11_125122beir-corpus
BeIR Corpus
scifact-corpus
SciFact Corpus
Scientific document retrieval corpus from the SciFact dataset.
Features
docid: document identifier
title: document title
text: document contents