Ayush-Singh/reward-bench-deepseek-math-7b-instruct-yes-no

该数据集包含了与数学问题相关的多个字段，包括问题提示(prompt)、选择的答案(chosen)、选择答案的模型(chosen_model)、被拒绝的答案(rejected)、拒绝答案的模型(rejected_model)、数据子集(subset)、唯一标识符(id)、选择答案为是的概率(chosen_yes_prob)、选择答案为不是的概率(chosen_no_prob)、拒绝答案为是的概率(rejected_yes_prob)和拒绝答案为不是的概率(rejected_no_prob)。数据集分为math_prm部分，共有447个示例。

The dataset includes multiple fields related to math problems, such as problem prompt (prompt), chosen answer (chosen), model for the chosen answer (chosen_model), rejected answer (rejected), model for the rejected answer (rejected_model), subset of data (subset), unique identifier (id), probability of the chosen answer being yes (chosen_yes_prob), probability of the chosen answer being no (chosen_no_prob), probability of the rejected answer being yes (rejected_yes_prob), and probability of the rejected answer being no (rejected_no_prob). The dataset is split into a math_prm section, containing a total of 447 examples.