This archive contains evaluation sets for monolingual and cross-lingual graded and ungraded lexical entailment (LE) used in the paper "Multilingual and Cross-Lingual Graded Lexical Entailment". 2019. Ivan Vulić, Simone Paolo Ponzetto, and Goran Glavaš. Proceedings of ACL 2019.

Ungraded LE evaluation sets are from the recent work of Upadhyay et al. (NAACL 2018).

Monolingual and cross-lingual graded LE evaluation sets have been created as one contribution of this paper. The only exception is the English monolingual dataset (hyperlex-en.txt) created by Vulic et al. (Computational Linguistics 2017). The annotation guidelines for all 4 languages (English, German, Italian, Croatian) are also provided.

For more details regarding the construction of these datasets, we refer the reader to the main paper as well as to the aforementioned prior work.
