1. Twenty Newsgroups: This data set consists of 20000 messages taken from 20 newsgroups.
2. NSF Research Award Abstracts 1990-2003: This data set consists of (a) 129,000 abstracts describing NSF awards for basic research, (b) bag-of-word data files extracted from the abstracts, (c) a list of words used for indexing the bag-of-word
3. CLINC150: This is a intent classification (text classification) dataset with 150 in-domain intent classes. The main purpose of this dataset is to evaluate various classifiers on out-of-domain performance.
4. Russian Corpus of Biographical Texts: Sentence classification (Russian). The corpus contains Wikipedia texts splitted into sentences/ Each sentence has a topic label.
5. Sentiment Labelled Sentences: The dataset contains sentences labelled with positive or negative sentiment.
6. University of Tehran Question Dataset 2016 (UTQD.2016): Persian questions gathered from a jeopardy game broadcasted on Iranian national television.
7. Legal Case Reports: A textual corpus of 4000 legal cases for automatic summarization and citation analysis. For each document we collect catchphrases, citations sentences, citation catchphrases and citation classes.
8. Reuters-21578 Text Categorization Collection: This is a collection of documents that appeared on Reuters newswire in 1987. The documents were assembled and indexed with categories.
9. Sentence Classification: Contains sentences from the abstract and introduction of 30 articles annotated with a modified Argumentative Zones annotation scheme. These articles come from biology, machine learning and psychology.