This dataset serves as a convenient way to access the ParlAI dialogue NLI dataset. I do not own the rights and all credit go to the original authors.

Dialogue Natural Language Inference

Sean Welleck, Jason Weston, Arthur Szlam, Kyunghyun Cho

arxiv link: https://arxiv.org/abs/1811.00671

Abstract: Consistency is a long standing issue faced by dialogue models. In this paper, we frame the consistency of dialogue agents as natural language inference (NLI) and create a new natural language inference dataset called Dialogue NLI. We propose a method which demonstrates that a model trained on Dialogue NLI can be used to improve the consistency of a dialogue model, and evaluate the method with human evaluation and with automatic metrics on a suite of evaluation sets designed to measure a dialogue model’s consistency.

How to use

from datasets import load_dataset

dataset = load_dataset('xksteven/dialogue_nli', split='train')

label candidates:

entailment
contradiction
neutral

Train dataset features.

Dataset({
    features: ['id', 'label', 'premise', 'hypothesis', 'dtype'],
    num_rows: 310110
})

Citation

@misc{welleck2019dialogue,
      title={Dialogue Natural Language Inference}, 
      author={Sean Welleck and Jason Weston and Arthur Szlam and Kyunghyun Cho},
      year={2019},
      eprint={1811.00671},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}