Paragraph Segmentation

A 9,000-document BIG-bench task: label which sentences end paragraphs in nine European languages as a 0/1 sequence.

Also known as: BIG-bench paragraph_segmentation

unassessed

This page is a discovery lead. Nobody has yet assessed it against the catalogue contract, so it carries no disposition. Absence of evidence here is not evidence of staleness.
Categorygeneration
Subcategorymultilingual sentence-level paragraph-boundary labelling
Page statusunknown
Metricexact_str_match
Directionhigher_is_better
Unit%
Dataset size9000
Dataset licenceApache-2.0
PublisherGoogle (BIG-bench collaboration); task authors at Wrocław University of Science and Technology

What it measures

paragraph_segmentation feeds a sequence of consecutive sentences from a Wikipedia article and asks for a same-length sequence of 0/1 labels: 1 if the sentence ends a paragraph, 0 otherwise. Authors at Wrocław University of Science and Technology sampled 1,000 articles per language from Wikimedia dumps (tables, images, links, headings, and templates stripped; sentences split with Moses). Languages: German, Spanish, Italian, Russian, Polish, English, French, Dutch, Portuguese. The stated skill is detecting a semantic break, with speech formatting and short-context packing as intended uses.

Task format

Free-text label string. preferred_score exact_str_match. task_prefix tells the model to emit 0 and 1 in sentence order as one string. The worked example uses spaces ("0 0 1"). Canary GUID embedded. Dummy-model header: 0 multiple-choice and 9,000 free-text queries.

Models reporting this benchmark

No model card in ModelSpec reports this benchmark yet.

Data

This page as JSON · Edit on GitHub