1 dataset found

The dataset used in the paper is the Wikipedia dataset, which contains over six million English Wikipedia articles with a full-text field associated with 50 training queries...