A feasibility study on the automated generation of acceptance criteria for user stories using LLMs

This dataset supports a study exploring the use of large language models (LLMs) to automate the generation of acceptance criteria for user stories in agile software development. Eight software professionals contributed user stories and criteria, which were compared against output from OpenAI’s o1-mini model. The study includes qualitative evaluations from a focus group, highlighting both the potential and limitations of using LLMs for this task.

本数据集支撑一项探索大语言模型（Large Language Models，LLMs）在敏捷软件开发中自动化生成用户故事验收标准的研究。八名软件专业人士提供了用户故事及其验收标准，并将其与OpenAI的o1-mini模型生成的输出内容进行了对比。该研究包含焦点小组开展的定性评估，凸显了利用大语言模型完成此类任务的潜力与局限性。