LangAGI-Lab/qwen-7b-verified-8k-dpo-preference-set-filtered-400
收藏Hugging Face2025-02-11 更新2025-02-15 收录
下载链接：
https://hf-mirror.com/datasets/LangAGI-Lab/qwen-7b-verified-8k-dpo-preference-set-filtered-400
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含了指令和答案的配对，以及针对每个答案的选择和拒绝信息。每个答案都有是否正确的标记、答案文本和token长度。此外，数据集还包含了每个答案的辅助信息(o1-mini_thought_tokens)和答案的来源。数据集提供了一个训练集分割，可用于机器学习模型的训练。
The dataset consists of pairs of instructions and answers, along with information on the selection and rejection of each answer. Each answer includes a correctness label, answer text, and token length. Additionally, the dataset contains auxiliary information (o1-mini_thought_tokens) and the source of each answer. The dataset provides a training split for use in machine learning model training.
提供机构：
LangAGI-Lab
