Safety Prompts
收藏arXiv2025-09-30 收录
下载链接:
https://github.com/thu-coai/Safety-Prompts
下载链接
链接失效反馈官方服务:
资源简介:
该数据集与安全和人类行为一致性问题相关,包含了一系列确保符合医学伦理的提示。此外,该数据集被用于确保模型的回应与安全协议和规定保持一致,其任务在于确保模型回应遵循医学伦理标准。
This dataset pertains to safety and human behavior alignment concerns, and includes a set of prompts developed to uphold medical ethics. Additionally, this dataset is used to ensure that model responses align with safety protocols and regulatory requirements, with its core objective being to guarantee that model responses comply with medical ethics standards.
提供机构:
THU COAI



