aisi-whitebox/inspect_llama_31_8b_instruct_prompted_sandbagging_wmdp_bio_unfiltered
收藏数据链接:
官方服务:
资源简介:
该数据集包含了对话内容、角色、目标标签、系统提示、是否良性、输入ID、任务名称、样本索引、数据集ID等信息。数据集分为训练集,共有64个样本。提供了默认配置文件,用于指定训练数据的路径。
The dataset includes conversation content, role, target labels, system prompts, benign or malicious, input IDs, task names, sample indices, dataset IDs, etc. The dataset is split into a training set with a total of 64 samples. A default configuration file is provided to specify the path to the training data.
提供机构:
aisi-whitebox


