DocAILab/FedE4RAG_Dataset

FedE4RAG_Dataset是为论文《Privacy-Preserving Federal Embedding Learning for Localized Retrieval-Augmented Generation》创建的数据集，旨在解决私有RAG系统中的数据稀缺和隐私挑战。该数据集使用联邦学习（FL）协作训练客户端RAG检索模型，保持原始数据的本地化。数据集包含训练数据和测试/验证数据，结构清晰，分别用于生成训练数据和下游问答测试。数据字段详细描述了每个字段的含义，包括上下文、索引、公司信息、问题、答案等。数据集的部分语料来源于开源数据集financebench。

FedE4RAG_Dataset is the dataset of the paper ***Privacy-Preserving Federal Embedding Learning for Localized Retrieval-Augmented Generation***. It addresses data scarcity and privacy challenges in private RAG systems. It uses federated learning (FL) to collaboratively train client-side RAG retrieval models, keeping raw data localized. The dataset contains training data and test/validation data, with clear structures for generating training data and downstream question & answer testing. Data fields are detailed, including context, index, company information, questions, answers, etc. Part of the corpus is derived from the open-source dataset financebench.