zjhhhh/game_iter2_zjhhhh__iter2_ver2_pad_adversary_min_step_301
收藏数据链接:
官方服务:
资源简介:
该数据集包含提示、要求以及对应的多个基础响应和模型响应,每个响应都有多个评判者的多数意见和平均评分。数据集仅包含训练集,共500个示例。
The dataset includes prompts, requirements, and corresponding multiple base and model responses, each with majority opinions and average ratings from multiple judges. The dataset contains only a training set with a total of 500 examples.
提供机构:
zjhhhh


