extractive summarization produces summaries by identifying and concatenating
the most important sentences in a document. Since most summarization datasets
do not come with gold labels indicating whether document sentences are
summary-worthy, different labeling algorithms have been prop