CoNLL 2012 corpus

The CoNLL 2012 corpus is a dataset for coreference resolution, containing data from diverse domains such as newswire, magazines, and conversations.

BibTex: