jbloom-aisi/wmdp-bio-llama-3.1-8B-sandbagging-responses

这是一个包含问题、答案和多种相关信息的文本数据集，适用于机器学习模型的训练和评估。数据集中的每个样本都包括问题、正确答案、干扰项、主题、提示文本等字段。此外，数据集还提供了沙包环境标识、模型答案和模型答案正确性等额外信息。数据集分为训练集，并可通过指定路径加载数据。

This is a text dataset containing questions, answers, and various related information, suitable for training and evaluating machine learning models. Each sample in the dataset includes fields such as question, correct answer, distractors, topic, prompt text, etc. Additionally, the dataset provides extra information such as sandbagging environment indicator, model answer, and correctness of the models answer. The dataset is split into a training set and can be loaded by specifying a path.