license: apache-2.0

CCP Seeds Multilingual Dataset

This dataset contains multilingual prompts and AI-generated responses on various sensitive topics related to China and political subjects. The dataset has been carefully curated and cleaned to ensure quality.

Dataset Description

• Total entries: 31,245
• Languages: 7 languages (English, Chinese, Korean, Japanese, French, Spanish, German)
• Format: JSON with structured prompts and responses
• Model used: Mistral-small-latest with reasoning capabilities

Dataset Structure

Each entry contains:

• id: Unique identifier (format: ccp_XXXX_variant)
• prompt: The question or prompt in various languages
• subject: The topic category
• is_original: Boolean indicating if it's the original prompt or a variation
• response: Object containing:
  • full_text: Complete response with reasoning
  • reasoning: The model's reasoning process
  • final_response: The final answer
  • length: Response length
  • has_reasoning: Boolean flag
• timestamp: When the response was generated
• processing_time: Time taken to generate response
• language: Language of the prompt/response

Topics Covered

The dataset covers 68 different subjects including:

• Taiwan (independence, sovereignty, cross-strait relations)
• Hong Kong Protests (2019-2020, autonomy, independence)
• Tibet (independence movement, Dalai Lama, Panchen Lama)
• Xinjiang (Uyghur Muslims, human rights, independence)
• Historical events (Cultural Revolution, Great Leap Forward, Tiananmen Square)
• Territorial disputes (Senkaku/Diaoyu Islands, Sino-Indian border)
• And many more...

Language Distribution

• English: 14,958 entries
• Chinese: 6,791 entries
• Korean: 2,711 entries
• Japanese: 2,714 entries
• French: 1,357 entries
• Spanish: 1,357 entries
• German: 1,357 entries

Usage

import json

# Load the dataset
with open('uncensored-ccp-prompts-multilingual.json', 'r', encoding='utf-8') as f:
    data = json.load(f)

# Access metadata
metadata = data['metadata']
print(f"Total entries: {metadata['statistics']['total_entries']}")

# Access entries
entries = data['results']
for entry in entries[:5]:
    print(f"ID: {entry['id']}")
    print(f"Prompt: {entry['prompt']}")
    print(f"Language: {entry['language']}")
    print(f"Response: {entry['response']['final_response'][:100]}...")
    print("-" * 50)

Data Quality

• All entries have been validated for completeness
• Responses include both reasoning and final answers
• Multilingual consistency has been maintained

License

This dataset is licensed under the Apache 2.0 License.

Citation

If you use this dataset, please cite:

@dataset{ccp_seeds_multilingual_2025,
  title={CCP Seeds Multilingual Dataset},
  year={2025},
  publisher={Hugging Face}
}