QuALITY: Question Answering with Long Input Texts, Yes!
This addresses the problem of evaluating long-document comprehension for NLP researchers, providing a challenging benchmark with human-validated questions.
The authors introduced QuALITY, a multiple-choice QA dataset with long context passages averaging 5,000 tokens to test long-document comprehension, where baseline models performed poorly at 55.4% compared to human performance of 93.5%.
To enable building and testing models on long-document comprehension, we introduce QuALITY, a multiple-choice QA dataset with context passages in English that have an average length of about 5,000 tokens, much longer than typical current models can process. Unlike in prior work with passages, our questions are written and validated by contributors who have read the entire passage, rather than relying on summaries or excerpts. In addition, only half of the questions are answerable by annotators working under tight time constraints, indicating that skimming and simple search are not enough to consistently perform well. Our baseline models perform poorly on this task (55.4%) and significantly lag behind human performance (93.5%).