LawFlow

该数据集名为LawFlow，由受过训练的法律专业学生提供的完整端到端法律工作流程组成，这些流程基于真实的商业实体成立场景。它捕捉了动态、模块化和迭代推理过程，这些过程反映了法律实践中存在的模糊性、修订和适应客户策略的特点。此外，该数据集还包括基于匿名化真实案例的标注场景，捕捉影响法律任务执行的结构性和人际因素，并辅以一个基于网页的应用程序进行数据收集。在规模上，该数据集通过深入访谈高级法律教员和学生，以及法律和计算机科学学生参与的模拟角色扮演来进行扩展。其任务是进行端到端的法律工作流程模拟与分析。

This dataset, named LawFlow, comprises full end-to-end legal workflows provided by trained law students, which are grounded in real business entity formation scenarios. It captures dynamic, modular, and iterative reasoning processes that reflect the ambiguity, revisions, and client strategy adaptation inherent in legal practice. Furthermore, the dataset includes annotated scenarios based on anonymized real cases, capturing structural and interpersonal factors that influence the execution of legal tasks, and is supported by a web-based application for data collection. In terms of scale, the dataset is expanded through in-depth interviews with senior legal faculty and students, as well as simulated role-playing activities involving law and computer science students. Its core objective is to enable end-to-end legal workflow simulation and analysis.