cs-552-2026-4neurons/xcopa

XCOPA数据集是一个跨语言合并版本，针对候选语言（意大利语和中文）进行了整合。它包含验证和测试分割，每个示例包含以下字段：前提（问题的基础背景）、第一个选择、第二个选择、被询问的问题、正确答案（0表示第一个选择，1表示第二个选择）、示例索引、示例是否从原始版本修改的布尔标志以及示例的语言。该数据集旨在支持多语言推理任务，通过合并不同语言的数据来促进跨语言比较和分析。The XCOPA dataset is a merged version across candidate languages (Italian and Chinese). It includes both validation and test splits with fields such as premise, choice1, choice2, question, label (0 for choice1, 1 for choice2), idx, changed (boolean indicating if the example was modified from the original), and lang (language of the example). This dataset is designed for multilingual reasoning tasks, facilitating cross-lingual comparison and analysis by integrating data from different languages.