Ayush-Singh/reward-bench-pythia-6.9b-set2-scores

该数据集是一个包含多个字段，如提示语(prompt)、选中项(chosen)、选中模型(chosen_model)、拒绝项(rejected)、拒绝模型(rejected_model)、子集(subset)、唯一标识符(id)、选中分数(chosen_score)、选中理由(chosen_justification)、拒绝分数(rejected_score)和拒绝理由(rejected_justification)的数据集。数据集分为alpacaeval_hard、hep_go、refusals_offensive等子集，每个子集包含不同数量的示例和对应的字节大小。数据集的下载大小为3.95MB，总大小为12.61MB。

The dataset consists of multiple fields such as prompt, chosen, chosen_model, rejected, rejected_model, subset, id, chosen_score, chosen_justification, rejected_score, and rejected_justification. It is divided into subsets like alpacaeval_hard, hep_go, refusals_offensive, etc., each with a different number of examples and corresponding byte sizes. The download size of the dataset is 3.95MB, and the total size is 12.61MB.