CoolFace
Datasetpublic

minliii/LongBench

LongBench is a comprehensive benchmark for multilingual and multi-task purposes, with the goal to fully measure and evaluate the ability of pre-trained language models to understand long text. This dataset consists of twenty different tasks, covering key long-text application scenarios such as multi-document QA, single-document QA, summarization, few-shot learning, synthetic tasks, and code completion.

sourceHugging Faceupdated 9mo agoView on Hugging Face
0likes23downloads

Nothing at this path on main. The folder may be empty, or the revision may not exist.

minliii/LongBench · main · files are served by the source, never re-hosted here