mvrcii/safety-synthetic-corpus
收藏Hugging Face2025-10-20 更新2025-10-25 收录
下载链接:
https://hf-mirror.com/datasets/mvrcii/safety-synthetic-corpus
下载链接
链接失效反馈官方服务:
资源简介:
安全合成语料库包含20万个合成的对话,用于训练安全分类器,覆盖了9个安全关键领域的有害、边界和安全内容。这些对话是通过结合不同的提问风格、内容方面和约束条件生成的,并且通过分层抽样确保了主题的平衡分布。
The Safety Synthetic Corpus consists of 200,000 synthetically generated conversations for training safety classifiers, covering harmful, borderline, and safe content across 9 safety-critical domains. These conversations are generated by combining various question styles, content aspects, and constraints, with stratified sampling ensuring a balanced distribution of topics.
提供机构:
mvrcii



