---
datasets:
- Hello-SimpleAI/HC3-Chinese
language:
- zh
pipeline_tag: text-classification
tags:
- chatgpt
---

# Model Card for `Hello-SimpleAI/chatgpt-detector-roberta-chinese`

This model is trained on **the mix of full-text and splitted sentences** of `answer`s from [Hello-SimpleAI/HC3-Chinese](https://huggingface.co/datasets/Hello-SimpleAI/HC3-Chinese).

The base checkpoint is [hfl/chinese-roberta-wwm-ext](https://huggingface.co/hfl/chinese-roberta-wwm-ext).
We train it with all [Hello-SimpleAI/HC3-Chinese](https://huggingface.co/datasets/Hello-SimpleAI/HC3-Chinese) data (without held-out) for 2 epoch.

(2-epoch is consistent with the experiments in the [paper](https://arxiv.org/abs/2301.07597).)

## Citation

Checkout this papaer [arxiv: 2301.07597](https://arxiv.org/abs/2301.07597)

```
@article{guo-etal-2023-hc3,
    title = "How Close is ChatGPT to Human Experts? Comparison Corpus, Evaluation, and Detection",
    author = "Guo, Biyang  and
      Zhang, Xin  and
      Wang, Ziyuan  and
      Jiang, Minqi  and
      Nie, Jinran  and
      Ding, Yuxuan  and
      Yue, Jianwei  and
      Wu, Yupeng",
    journal={arXiv preprint arxiv:2301.07597}
    year = "2023",
}
```