nolangclem/bigmath-qwen2.5-3b-step-by-step-confidence-v2

该数据集包含问题解决的相关信息，每个样本包括问题、地面真实答案、部分解决方案、步骤数、总步骤数、完成百分比、模型置信度、置信度原始输出、经验成功率、延续次数、正确次数、完整初始生成内容、延续内容和解析答案等字段。数据集分为训练集，并提供相应的字节数和示例数量。

This dataset contains information related to problem-solving, with each sample including fields such as problem, ground truth, partial solution, step number, total steps, percentage complete, model confidence, raw confidence output, empirical success rate, number of continuations, number of correct answers, full initial generation content, continuation content, and parsed answers. The dataset is split into a training set, with corresponding byte size and number of examples provided.