aisi-whitebox/sec_qa_v1_cot_prompted_sandbagging_llama_31_8b_instruct_finetuning

该数据集包含了聊天对话内容以及相关的标签、得分、元数据等信息。每个样本都包含了对话内容（content）、角色（role）、输入ID（input_ids）、任务名称（task_name）等字段。此外，数据集还包含了是否为良性（is_benign）、沙袋执行状态（sandbagging_executed）等特征。数据集分为训练集（train）等部分，可用于训练聊天机器人模型等。

The dataset contains chat conversation content along with related labels, scores, metadata, and other information. Each sample includes fields such as conversation content (content), role (role), input ID (input_ids), task name (task_name), etc. In addition, the dataset also includes features such as whether it is benign (is_benign) and the status of sandbagging execution (sandbagging_executed). The dataset is divided into parts such as the training set (train) and can be used to train chatbot models and more.