Raw survey evaluation data for the manuscript “Comparing ChatGPT’s and Google Gemini’s Responses to Lung Cancer Questions”

This dataset includes raw survey responses collected from physicians, nurses, patients, and authors evaluating AI-generated answers to 40 lung cancer–related questions. The data were used to assess accuracy, relevance, sufficiency, understandability, and readability across two language models (ChatGPT and Gemini).In addition, the dataset contains raw evaluations of the references cited by the AI models, including classification of consistency, level of evidence, region, type and hallucination.

本数据集收录了由医生、护士、患者及作者针对40个肺癌相关问题的人工智能（AI）生成回复所给出的原始调研评价，用于评估两款语言模型（ChatGPT与Gemini）生成回复的准确性、相关性、充分性、易懂性与可读性。此外，本数据集还包含针对人工智能模型所引用参考文献的原始评价，涵盖一致性、证据等级、所属地域、文献类型及幻觉内容的分类标注。