MLQA

该数据集是一个多语言问答基准，用于评估语言模型的翻译能力。在评估时，我们采用了F1分数作为衡量标准。该数据集中包含了从原始数据中抽取的200个测试实例，任务主要集中在问答方面。

This dataset is a multilingual question answering benchmark designed to evaluate the translation capabilities of language models. F1 score is employed as the performance metric during the evaluation process. This dataset contains 200 test instances extracted from the original corpus, with the tasks primarily focused on question answering.