self-ask

This repo contains the code and data for our "Measuring and Narrowing the Compositionality Gap in Language Models" paper, presented at Findings of EMNLP 2023.

You can read more about self-ask prompting here (including a 2 minute demo video).

Or read more about our Compositional Celebrities dataset, and the Compositionality Gap here (also includes a 2 minute video).

Lastly, you can learn more about our Bamboogle dataset here (includes a 2 minute video).

Citation:

@inproceedings{press-etal-2023-measuring,
    title = "Measuring and Narrowing the Compositionality Gap in Language Models",
    author = "Press, Ofir  and
      Zhang, Muru  and
      Min, Sewon  and
      Schmidt, Ludwig  and
      Smith, Noah  and
      Lewis, Mike",
    editor = "Bouamor, Houda  and
      Pino, Juan  and
      Bali, Kalika",
    booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
    month = dec,
    year = "2023",
    address = "Singapore",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2023.findings-emnlp.378",
    doi = "10.18653/v1/2023.findings-emnlp.378",
    pages = "5687--5711",
}