|
|
--- |
|
|
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. |
