gx-ai-architect/numinamath-178k-phi4-bon-verified-dpo-trl
收藏Hugging Face2025-02-03 更新2025-02-15 收录
下载链接：
https://hf-mirror.com/datasets/gx-ai-architect/numinamath-178k-phi4-bon-verified-dpo-trl
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含两个主要特征字段：选中的(chosen)和被拒绝的(rejected)，每个字段都包括内容(content)和角色(role)信息。此外，还有选中的得分(score_chosen)和被拒绝的得分(score_rejected)。数据集分为训练集和测试集，其中训练集有38033个样本，测试集有1000个样本。
The dataset includes two main feature fields: chosen and rejected, each with content and role information. Additionally, there are scores for the chosen and rejected items. The dataset is split into a training set with 38033 samples and a test set with 1000 samples.
提供机构：
gx-ai-architect
