Most prior work on definition modeling has not accounted for polysemy, or has\ndone so by considering definition modeling for a target word in a given\ncontext. In contrast, in this study, we propose a context-agnostic approach to\ndefinition modeling, based on multi-sense word embeddings, that is capable of\ngenerating multiple definitions for a target word. In further, contrast to most\nprior work, which has primarily focused on English, we evaluate our proposed\napproach on fifteen different datasets covering nine languages from several\nlanguage families. To evaluate our approach we consider several variations of\nBLEU. Our results demonstrate that our proposed multi-sense model outperforms a\nsingle-sense model on all fifteen datasets.\n