shalanova/benchmark-3-chinese-m2m

该数据集由JailbreakBench/JBB-Behaviors翻译而来，使用了facebook/m2m100_418M模型进行中文翻译。数据集包含异构的不安全类别（如有害指令、敏感话题、对抗性重述等），并且包含不一定遵循典型越狱模板的提示。这种多样性和分布变异性使得基于相似性的检测更具挑战性，并为跨语言迁移提供了压力测试。数据集大小为200个提示（100个安全/100个不安全），包含四个列：text（原始提示）、label（0表示安全，1表示不安全）、translation（中文翻译的提示）、score_zh_model（与codebook的余弦相似度得分）。

Translated from JailbreakBench/JBB-Behaviors using the facebook/m2m100_418M model. The dataset includes heterogeneous unsafe categories (e.g., harmful instructions, sensitive topics, adversarial rephrasings) and contains prompts that do not necessarily follow canonical jailbreak templates. This increased diversity and distributional variability makes similarity-based detection more challenging and provides a stress-test for cross-lingual transfer. The dataset size is 200 prompts (100 safe / 100 unsafe) with four columns: text (original prompt), label (0: safe, 1: unsafe), translation (prompt translated into Chinese), score_zh_model (cosine similarity score with codebook).