The Wino-X dataset is a multilingual collection of Winograd Schemas. It was introduced as a tool for evaluating coreference resolution (CoR) and commonsense reasoning (CSR) capabilities of computational models. The dataset contains schemas in German, French, and Russian, aligned with their English counterparts.
The Wino-X dataset is designed for the evaluation of Neural Machine Translation (NMT) models and Multilingual Language Models (MLLMs). It contains two directories: mt_wino_x
and lm_wino_x
, designed for the evaluation of NMT models and MLLMs, respectively.
Paper | Code | Results | Date | Stars |
---|