shalanova/benchmark-2-chinese-m2m
收藏Hugging Face2026-04-30 更新2026-05-03 收录
下载链接:
https://hf-mirror.com/datasets/shalanova/benchmark-2-chinese-m2m
下载链接
链接失效反馈官方服务:
资源简介:
该数据集主要包含提示注入和典型的越狱式指令,具有相对同质的攻击模式。数据集大小为1,000个提示,其中500个是安全的,500个是不安全的。数据集包含四个列:text(原始提示)、label(0表示安全,1表示不安全)、translation(由facebook/m2m100_418M模型翻译成中文的提示)和score_zh_model(与codebook的余弦相似度分数)。
The dataset primarily contains prompt-injection and canonical jailbreak-style instructions with relatively homogeneous attack patterns. The size is 1,000 prompts (500 safe / 500 unsafe). Columns include: text - original prompt, label - 0: safe, 1: unsafe, translation - prompt on Chinese translated by facebook/m2m100_418M, and score_zh_model - cosine similarity score with codebook.
提供机构:
shalanova



