SCROLLS: Standardized CompaRison Over Long Language Sequences
This addresses the problem of evaluating NLP models on long-text tasks for researchers, though it is incremental as it builds on existing datasets and formats.
The authors tackled the lack of NLP benchmarks for long texts by introducing SCROLLS, a suite of tasks requiring reasoning over long texts, and initial baselines show significant room for improvement.
NLP benchmarks have largely focused on short texts, such as sentences and paragraphs, even though long texts comprise a considerable amount of natural language in the wild. We introduce SCROLLS, a suite of tasks that require reasoning over long texts. We examine existing long-text datasets, and handpick ones where the text is naturally long, while prioritizing tasks that involve synthesizing information across the input. SCROLLS contains summarization, question answering, and natural language inference tasks, covering multiple domains, including literature, science, business, and entertainment. Initial baselines, including Longformer Encoder-Decoder, indicate that there is ample room for improvement on SCROLLS. We make all datasets available in a unified text-to-text format and host a live leaderboard to facilitate research on model architecture and pretraining methods.