nmayorga7/bbq-deepseek-v3.2-exp

该数据集包含了实验结果、样本数据和统计信息。实验结果数据集包含字段如日志路径、评估ID、运行ID、创建时间、任务类型等。样本数据数据集除了包含实验结果数据集的字段外，还包含了样本ID、训练轮次、目标、消息、元类别、问题极性、上下文条件、MCQ评分器值和答案等字段。统计信息数据集则包含了实验的开始和结束时间、模型使用情况、输入输出令牌数等统计字段。所有数据集都针对训练集进行了划分。

The dataset consists of experiment results, sample data, and statistics. The experiment results dataset includes fields such as log path, evaluation ID, run ID, creation time, task type, etc. The sample data dataset includes all fields from the experiment results dataset and additional fields such as sample ID, epoch, target, messages, meta category, question polarity, context condition, MCQ scorer value, and answer. The statistics dataset includes fields such as start and end time of the experiment, model usage, number of input and output tokens, etc. All datasets are split for the training set.