---
license: mit
language:
- en
- zh
task_categories:
- text2text-generation
tags:
- math-word-problems
size_categories:
- 100K<n<1M
---

## Dataset Summary  
This dataset is derived from the subset **madrylab/gsm8k-platinum** of the **openai/gsm8k** dataset.  

### Transformation Method:  
1. Template-based question generation.  
2. Each question generates 800 similar questions of the same type.  

**Note:** This dataset (gsm_800k_beta1.jsonl) has not undergone strict human verification, and approximately `2%` of the questions may contain errors.