jdineen/pairsv2_annotated

该数据集包含多个特征，如数据来源、提示ID、提示内容、两个响应、温度参数、模型信息、生成时间戳、评分结构（包括无害性、帮助性、诚实性等原则评分）、比较结果（包括分数是否相等和分数差异）、两个模型的响应分数以及选择的响应。数据集划分为训练集，包含1个样本，总大小为705字节，下载大小为16161字节。

The dataset includes multiple features such as dataset source, prompt ID, prompt content, two responses, temperature parameters, model information, generation timestamp, scoring structure (including principle scores for harmlessness, helpfulness, honesty, etc.), comparison results (including whether scores are equal and score differences), response scores for two models, and the chosen response. The dataset is divided into a training set with 1 example, a total size of 705 bytes, and a download size of 16161 bytes.