GSM8K-Hi

## Dataset Description:
The GSM8K-Hi (Hindi GSM8K) is the GCP translated counterpart of the English GSM8K test set. The samples are carefully reviewed by the human annotators and corrected for quality improvement. These problems take between 2 and 8 steps to solve, and solutions primarily involve performing a sequence of elementary calculations using basic arithmetic operations (+ − ×÷) to reach the final answer.

This dataset is ready for commercial/non-commercial use. The evaluation steps are described [here](https://huggingface.co/datasets/nvidia/GSM8K-Hi/blob/main/EVAL.md).

## Dataset Owner:
NVIDIA Corporation

## Dataset Creation Date:
April 2025

## License/Terms of Use:
CC-BY 4.0

## Intended Usage:
This dataset is intended to evaluate the ability of language models to solve grade school-level math word problems that require multi-step reasoning.

## Dataset Characterization
Data Collection Method<br>
* Synthetic <br>

Labeling Method<br>
* Not Applicable <br>

## Dataset Format
Text

## Dataset Quantification
1.6MB of question-answer pairs, comprising 1319 individual samples.

## Ethical Considerations:
NVIDIA believes Trustworthy AI is a shared responsibility, and we have established policies and practices to enable development for a wide array of AI applications. When downloaded or used in accordance with our terms of service, developers should work with their internal model team to ensure this model meets requirements for the relevant industry and use case and addresses unforeseen product misuse.

Please report security vulnerabilities or NVIDIA AI Concerns [here](https://www.nvidia.com/en-us/support/submit-security-vulnerability/).

## Citing

If you find our work helpful, please consider citing our paper:
```
@article{kamath2025benchmarking,
title={Benchmarking Hindi LLMs: A New Suite of Datasets and a Comparative Analysis},
author={Kamath, Anusha and Singla, Kanishk and Paul, Rakesh and Joshi, Raviraj and Vaidya, Utkarsh and Chauhan, Sanjay Singh and Wartikar, Niranjan},
journal={arXiv preprint arXiv:2508.19831},
year={2025}
}
```

# 数据集描述
GSM8K-Hi（印地语GSM8K）是英文GSM8K测试集经由GCP翻译得到的对应版本。所有样本均经过人工标注人员的仔细审核与修正，以提升数据集质量。此类问题需2至8步方可求解，解题过程主要通过基础算术运算（+ − ×÷）完成一系列初等计算，最终得到答案。

本数据集可用于商业及非商业用途。评估流程详见[此处](https://huggingface.co/datasets/nvidia/GSM8K-Hi/blob/main/EVAL.md)。

## 数据集所有者
英伟达（NVIDIA）公司

## 数据集创建日期
2025年4月

## 许可/使用条款
CC-BY 4.0

## 预期用途
本数据集旨在评估大语言模型（LLM/Large Language Model）解决需要多步推理的中小学数学应用题的能力。

## 数据集特征
### 数据收集方法
* 合成数据

### 标注方法
* 不适用

## 数据集格式
文本

## 数据集量化
包含1319条独立样本，总计1.6MB的问答对数据。

## 伦理考量
英伟达（NVIDIA）认为，可信人工智能是一项共同责任，我们已建立相应政策与实践规范，以支撑各类人工智能应用的开发。开发者在依照服务条款下载或使用本数据集时，应与其内部模型团队开展协作，确保所开发的模型符合相关行业及应用场景的要求，并应对可能出现的产品误用问题。

如需报告安全漏洞或英伟达人工智能相关问题，请访问[此处](https://www.nvidia.com/en-us/support/submit-security-vulnerability/)。

## 引用
若您认为本工作对您有所帮助，请引用以下论文：

@article{kamath2025benchmarking,
title={Benchmarking Hindi LLMs: A New Suite of Datasets and a Comparative Analysis},
author={Kamath, Anusha and Singla, Kanishk and Paul, Rakesh and Joshi, Raviraj and Vaidya, Utkarsh and Chauhan, Sanjay Singh and Wartikar, Niranjan},
journal={arXiv preprint arXiv:2508.19831},
year={2025}
}