prometheus-eval/filtered_bon_setting_64

这是一个包含数学和编程问题的数据集，适用于训练和评估模型在解决这些问题上的能力。数据集包含不同难度和类型的问题，例如aime、amc、gpqa、leetcode、math、minerva_math和olympiadbench等问题类型，以及不同模型训练阶段的数据集版本，例如Meta_Llama、Eurus2和Qwen2.5等。每个问题都包含问题的ID、生成器信息、问题本身、解决步骤、模型输出和最终答案是否正确的信息。

This dataset consists of mathematical and programming problems, designed for training and evaluating models on their ability to solve these problems. The dataset includes problems of various difficulties and types, such as aime, amc, gpqa, leetcode, math, minerva_math, and olympiadbench, as well as different dataset versions for different stages of model training, such as Meta_Llama, Eurus2, and Qwen2.5. Each problem contains the problem ID, generator information, the problem itself, solution steps, model output, and information on whether the final answer is correct.