rspiocbis/MambaInLlama_0_875-best_of_n-PRM-3370152

该数据集包含了数学问题及其答案，以及针对问题的多种预测结果和评估指标。数据集由两个配置组成，每个配置都有训练集分割。第一个配置的特征包括问题文本、答案文本、完成序列、分数、预测结果等。第二个配置的特征包括模型在不同情况下的准确度评估，如朴素准确度、加权准确度和多数投票准确度。数据集适用于数学问题解答模型的训练和评估。

This dataset consists of math questions and their answers, along with various prediction results and evaluation metrics for the questions. The dataset is composed of two configurations, each with a training set split. The first configuration features include question text, answer text, completion sequences, scores, prediction results, etc. The second configuration features include model accuracy evaluations under different conditions, such as naive accuracy, weighted accuracy, and majority vote accuracy. The dataset is suitable for training and evaluating math question answering models.