Ayush-Singh/reward-bench-MAmmoTH-7B-yes-no

这是一个包含数学问题提示及其对应选项的数据集，每个样本包括问题提示(prompt)、选中选项(chosen)、选中模型(chosen_model)、拒绝选项(rejected)、拒绝模型(rejected_model)、数据子集(subset)、唯一标识符(id)、选中选项为正确答案的概率(chosen_yes_prob)、选中选项为错误答案的概率(chosen_no_prob)、拒绝选项为正确答案的概率(rejected_yes_prob)和拒绝选项为错误答案的概率(rejected_no_prob)。数据集分为math_prm部分，共有447个示例。

This is a dataset containing math problem prompts and their corresponding options. Each sample includes the problem prompt (prompt), chosen option (chosen), chosen model (chosen_model), rejected option (rejected), rejected model (rejected_model), subset of data (subset), unique identifier (id), probability that the chosen option is the correct answer (chosen_yes_prob), probability that the chosen option is the wrong answer (chosen_no_prob), probability that the rejected option is the correct answer (rejected_yes_prob), and probability that the rejected option is the wrong answer (rejected_no_prob). The dataset is divided into the math_prm section, with a total of 447 examples.