Cornell-AGI/gsm8k_size_qwen2.5_1.5b_eval

该数据集包含了一系列的提示和相应的多个可能回答，以及对这些回答的评估。每个提示由内容和角色组成，同时数据集中还包含了能力、奖励模型的相关信息以及一些额外信息如答案、索引、问题和分割等。数据集专为训练设计，包含了7473个示例。

The dataset consists of a series of prompts and corresponding multiple possible responses, along with evaluations of these responses. Each prompt is composed of content and role, and the dataset also includes information about ability, reward model, as well as additional information such as answer, index, question, and split. The dataset is designed for training and contains 7473 examples.