In 2007, the MSR Natural Language Processing Group had the datasets used in the 2006 Recognizing Textual Entailment Challenge hand annotated by three annotators to indicate alignments of words and phrases. This technical report provides information on the format of the tagged files and the annotation process. Guidelines provided to the annotators are included in an appendix. The annotated datasets are available via the Textual Entailment Resource Pool at http://aclweb.org/aclwiki/index.php?title=Textual_Entailment_Resource_Pool

PDF file

Details