Translations:Simplification Data/36/en

As part of her internship with us last year, Eliza Hobo developed the first contextual lexical simplification model for Dutch. Due to the lack of Dutch evaluation data for lexical simplification, we developed a pilot benchmark dataset for the task using authentic municipal data. We select sentences from a collection of 48 municipal documents based on the presence of a complex word from a list curated by domain experts and based on their word count (less than 20 words).