EBMQA
收藏资源简介:
EBMQA数据集是由Kahun Medical Ltd基于其医学知识图谱创建的,包含105,000条基于证据的医学问题与答案。该数据集主要分为数值型和语义型两类问题,旨在测试大型语言模型在处理医学知识时的性能。数据集的内容涵盖了从超过50,000篇同行评审文章中提取的医学关系,用于生成直接的多选题。EBMQA数据集的应用领域主要集中在评估和提升大型语言模型在临床问题解决中的准确性和效率,特别是在处理数值和语义医学知识方面。
The EBMQA dataset was developed by Kahun Medical Ltd based on its proprietary medical knowledge graph, containing 105,000 evidence-based medical questions and answers. It is primarily categorized into two types: numerical and semantic questions, aiming to evaluate the performance of large language models (LLMs) when processing medical knowledge. The dataset covers medical relationships extracted from more than 50,000 peer-reviewed academic articles, which are used to generate direct multiple-choice questions. The core application scenarios of the EBMQA dataset center on evaluating and enhancing the accuracy and efficiency of large language models in clinical problem-solving, particularly in handling both numerical and semantic medical knowledge.

- 1Performance of large language models in numerical vs. semantic medical knowledge: Benchmarking on evidence-based Q&As特拉维夫大学医学院 · 2024年



