aisi-whitebox/mmlu_0_shot_cot_prompted_sandbagging_llama_31_8b_instruct_finetuning

该数据集包含了聊天对话、目标标签、元数据、得分、答案、系统提示、是否良性、输入ID、任务名称、样本索引、数据集ID、沙袋执行情况、提示数据集名称以及数据集分割信息等字段。数据集分为训练集，共有639个示例，大小为1871704字节。The dataset includes fields such as chat conversations, target labels, metadata, scores, answers, system prompts, benignity, input IDs, task names, sample indices, dataset IDs, sandbagging execution status, prompted dataset names, and dataset split information. The dataset is split into a training set with a total of 639 examples and a size of 1,871,704 bytes.