winogrande_et includes the test set of the winogrande dataset that was manually translated and culturally adapted to the Estonian language.
The dataset also includes a machine translated version performed by GPT4o for comparison, as well as manually and machine translated few-shot examples from the development set
of the original. The dataset was translated as part of a master’s thesis (M. Ojastu, 2025).
The thesis details the… See the full description on the dataset page: https://huggingface.co/datasets/tartuNLP/winogrande_et.
