SCROLLS: Standardized CompaRison Over Long Language Sequences

What is SCROLLS?

SCROLLS is a suite of datasets that require synthesizing information over long texts. The benchmark includes seven natural language tasks across multiple domains, including summarization, question answering, and natural language inference.

Citing SCROLLS

Please use the following bibliography to cite SCROLLS:

@misc{shaham2022scrolls,
    title={SCROLLS: Standardized CompaRison Over Long Language Sequences},
    author={Uri Shaham and Elad Segal and Maor Ivgi and Avia Efrat and Ori Yoran and Adi Haviv and Ankit Gupta and Wenhan Xiong and Mor Geva and Jonathan Berant and Omer Levy},
    year={2022},
    eprint={2201.03533},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}

When citing SCROLLS, please make sure to cite all of the original dataset papers. [bibtex]

Contact Us

scrolls-benchmark-contact@googlegroups.com

Logo black-01.png