t5_xxl_true_nli_mixture

This is an NLI model based on T5-XXL that predicts a binary label ('1' - Entailment, '0' - No entailment).

It is trained similarly to the NLI model described in the TRUE paper (Honovich et al, 2022), but using the following datasets instead of ANLI:

• SNLI (Bowman et al., 2015)
• MNLI (Williams et al., 2018)
• Fever (Thorne et al., 2018)
• Scitail (Khot et al., 2018)
• PAWS (Zhang et al. 2019)
• VitaminC (Schuster et al., 2021)

The input format for the model is: "premise: PREMISE_TEXT hypothesis: HYPOTHESIS_TEXT".

If you use this model for a research publication, please cite the TRUE paper (using the bibtex entry below) and the dataset papers mentioned above.

@inproceedings{honovich-etal-2022-true-evaluating,
    title = "{TRUE}: Re-evaluating Factual Consistency Evaluation",
    author = "Honovich, Or  and
      Aharoni, Roee  and
      Herzig, Jonathan  and
      Taitelbaum, Hagai  and
      Kukliansy, Doron  and
      Cohen, Vered  and
      Scialom, Thomas  and
      Szpektor, Idan  and
      Hassidim, Avinatan  and
      Matias, Yossi",
    booktitle = "Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies",
    month = jul,
    year = "2022",
    address = "Seattle, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2022.naacl-main.287",
    doi = "10.18653/v1/2022.naacl-main.287",
    pages = "3905--3920",
}