illuin-conteb/insurance

ConTEB-Insurance 数据集是 ConTEB（上下文感知文本嵌入基准）的一部分，旨在评估上下文嵌入模型的能力。该数据集聚焦于保险主题，特别是源自 EIOPA 实体文档的内容。数据集包括一份关于欧盟各国保险相关统计的文档，文档被分成多个文本块，并为这些块创建了需要结构化理解的手工编写的问题。该数据集提供了针对上下文嵌入的基准，包括经过策划的原始文档、源自它们的文本块以及查询。

The ConTEB-Insurance dataset is part of the ConTEB (Context-aware Text Embedding Benchmark), designed to evaluate the capabilities of contextual embedding models. It focuses on the insurance theme, particularly stemming from a document of the EIOPA entity. The dataset includes a long document with insurance-related statistics for each country of the European Union, split into chunks and manually crafted queries that require structural understanding for accurate chunk matching. It provides a focused benchmark for contextualized embeddings, including a curated set of original documents, chunks derived from them, and queries.