jihuny/pythia1.4b_10k_triplet_norm_penul_mean
收藏Hugging Face2025-11-03 更新2025-11-15 收录
下载链接：
https://hf-mirror.com/datasets/jihuny/pythia1.4b_10k_triplet_norm_penul_mean
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含了prompt文本及其索引，选择的文本及其索引，被拒绝的文本及其索引，选择的文本和被拒绝的文本的得分、对数概率和完成原因，以及两种文本的嵌入向量差异。数据集分为训练集，包含450000个示例，总大小为8157.53MB。
The dataset includes prompt text and its index, chosen text and its index, rejected text and its index, scores, log probabilities, and finish reasons for the chosen and rejected texts, as well as the embedding vector differences between the two texts. The dataset is split into a training set with 450,000 examples, totaling 8157.53MB in size.
提供机构：
jihuny
