ZixuanKe/fingpt_convfinqa_sup_sample_from_policy_v1.1_stepwise_dpo_binarized_val
收藏Hugging Face2024-11-23 更新2024-12-14 收录
下载链接:
https://hf-mirror.com/datasets/ZixuanKe/fingpt_convfinqa_sup_sample_from_policy_v1.1_stepwise_dpo_binarized_val
下载链接
链接失效反馈官方服务:
资源简介:
该数据集包含四个主要特征:提示(prompt)、被拒绝的响应(rejected)、被选择的响应(chosen)和理由(justification)。数据集仅包含一个训练集,共有1362个样本,总大小为8445126字节,下载大小为1412633字节。数据集的默认配置文件中,训练集的数据文件路径为data/train-*。
The dataset contains four main features: prompt, rejected, chosen, and justification. The dataset includes only a training set with 1362 examples, totaling 8445126 bytes in size, and a download size of 1412633 bytes. In the default configuration file, the data file path for the training set is data/train-*.
提供机构:
ZixuanKe



