README.md

---
license: mit
task_categories:
- text-generation
language:
- zh
tags:
- chinese
- spelling-correction
- csc
- text-correction
---

# Chinese Spelling Correction (CSC) Dataset

This repository contains various Chinese Spelling Correction (CSC) datasets as presented in the paper [CSRP: Chain-of-Thought Reasoning for Chinese Text Correction via Reinforcement Learning with Efficiency-Aware Rewards](https://huggingface.co/papers/2606.00020).

These datasets are part of the [ChineseErrorCorrector](https://github.com/TW-NLP/ChineseErrorCorrector) project, a comprehensive platform for Chinese text correction tasks.

## Dataset Statistics

The collection includes the following subsets:

| Subset | Count |
| :--- | :--- |
| **W271K** | 279,816 |
| **Medical** | 39,303 |
| **Lemon** | 22,259 |
| **ECSpell** | 6,688 |
| **CSCD** | 35,001 |

## Resources

- **GitHub Repository:** [TW-NLP/ChineseErrorCorrector](https://github.com/TW-NLP/ChineseErrorCorrector)
- **Paper:** [CSRP: Chain-of-Thought Reasoning for Chinese Text Correction via Reinforcement Learning with Efficiency-Aware Rewards](https://huggingface.co/papers/2606.00020)

## Sample Usage

The authors provide a tool for grammatical error augmentation via the `ChineseErrorCorrector` Python package:

```python
from ChineseErrorCorrector.utils.dat import GrammarErrorDat

dat = GrammarErrorDat()
print(dat.lack_word("小明住在北京"))       # 缺字漏字
print(dat.error_word("小明住在北京"))      # 错别字
print(dat.lack_punctuation("小明住在北京")) # 缺少标点
```

## Citation

If you use this dataset or the associated tools, please cite:

```bibtex
@misc{tian2026csrpchainofthoughtreasoningchinese,
      title={CSRP: Chain-of-Thought Reasoning for Chinese Text Correction via Reinforcement Learning with Efficiency-Aware Rewards}, 
      author={Wei Tian and Yuhao Zhou and Man Lan},
      year={2026},
      eprint={2606.00020},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2606.00020}, 
}
```