avemio/German-RAG-ORPO-Long-Context-ShareGPT-HESSIAN-AI
收藏Hugging Face2025-02-06 更新2025-04-08 收录
下载链接：
https://hf-mirror.com/datasets/avemio/German-RAG-ORPO-Long-Context-ShareGPT-HESSIAN-AI
下载链接
链接失效反馈官方服务：
资源简介：
German-RAG-ORPO长期上下文任务数据集是一个专门用于微调语言模型的数据集，重点关注RAG特定能力。该数据集包含三个子集：具有多个引用的困难问答、以与会者和主题为焦点的会议问答以及以主题为焦点的会议问答。数据子集源自合成数据生成，基于腾讯的千万人格模拟方法，并经过自动验证和审校。
The German-RAG-ORPO Long Context Tasks Dataset is a specialized collection designed for fine-tuning language models with a focus on RAG-specific capabilities. It includes three subsets: hard QA with multiple references, QA meeting attendee topic, and QA meeting topic. These subsets are derived from synthetic data generation inspired by Tencents method of simulating one billion personas, and have been automatically validated and curated.
提供机构：
avemio
