large-traversaal/urdu-mgsm

多语言小学数学基准测试（MGSM）是一个小学数学问题基准，来源于论文《Language models are multilingual chain-of-thought reasoners》。该数据集包含了250个来自GSM8K数据集的问题，这些问题由人工注释者翻译成了10种语言，包括西班牙语、法语、德语、俄语、中文、日语、泰语、斯瓦希里语、孟加拉语和泰卢固语。GSM8K是一个包含8500个高质量语言多样化的小学数学文字问题的数据集，用于支持基本数学问题上的多步骤推理问答任务。本数据集是为了评估目的，使用GPT-4o将数据集翻译成了乌尔都语。

The Multilingual Grade School Math Benchmark (MGSM) is a benchmark of grade-school math problems, sourced from the paper Language models are multilingual chain-of-thought reasoners. This dataset includes 250 problems from the GSM8K dataset, translated by human annotators into 10 languages: Spanish, French, German, Russian, Chinese, Japanese, Thai, Swahili, Bengali, and Telugu. GSM8K is a dataset of 8.5K high-quality linguistically diverse grade school math word problems, created to support the task of question answering on basic mathematical problems that require multi-step reasoning. This dataset has been translated into Urdu using GPT-4o for evaluation purposes.