MrZilinXiao/viRl39k_reasoning_distill

这是一个包含问题和答案以及相关信息的中文数据集，适用于模型训练和评估。数据集中的每个样本都包含问题、答案、32B训练模型的通过率、7B基础模型的通过率、类别、来源、问题ID、相关图片、推理过程和模型准确率等信息。数据集仅包含训练集划分，可用于模型训练和性能分析。

This is a Chinese dataset containing questions, answers, and related information, suitable for model training and evaluation. Each sample in the dataset includes a question, answer, pass rate of the 32B trained model, pass rate of the 7B base model, category, source, question ID, related images, reasoning process, and model accuracy. The dataset is split only into a training set, which can be used for model training and performance analysis.