InsultedByMathematics/llama3-ultrafeedback-armo-infoNCA-K-3
收藏数据链接:
官方服务:
资源简介:
该数据集包含多个响应及其对应的奖励值,以及提示和相关的token序列。数据集分为训练集和测试集,分别包含54328和1801个样本。数据集的下载大小为412739819字节,总大小为2321505543字节。
This dataset contains multiple responses with their corresponding rewards, prompts, and token information generated by the LLaMA model. The dataset is split into training and test sets for model training and evaluation.



