Dataset · Hello-SimpleAI
Comparison of human and ChatGPT answers in Chinese
Hello-SimpleAI/HC3-Chinese
Text corpus for comparing human and ChatGPT answers; suitable for classification and similarity evaluation.
What's inside
The Human ChatGPT Comparison Corpus in its Chinese version; metadata also indicates English language. Size is from 10k to 100k examples. The card specifies text classification, question-answering, sentence-similarity, and zero-shot classification tasks.
What you can do with it
- Training a classifier to identify ChatGPT texts
- Testing sentence-similarity models
- Zero-shot classification of answer sources
- Tasks
- Text classification, Question answering, Sentence similarity, Zero shot classification
- Languages
- English, Chinese
- Size
- 10K–100K rows
- Downloads this month
- 2.5K
- Likes
- 186
Models trained on it
- Hello-SimpleAI/chatgpt-detector-roberta-chineseText classification 10.4K
- Hello-SimpleAI/chatgpt-qa-detector-roberta-chineseText classification 137
- lilishian/GPTMups-ChineseText classification 25
- certainstar/Trained-Chinese-classificationText classification 8
- certainstar/Trained-Mul-classificationText classification 6
- LIghtJUNction/RISE-HC3-source-baselineText classification 0
- stevelikesrhino/laya-pinyin-2.1 0
How to load it
from datasets import load_dataset
ds = load_dataset("Hello-SimpleAI/HC3-Chinese")