Synthetic Workflow Dataset

该数据集是为了模拟现实世界的工作流程，以评估Performant Agentic框架（PAF）在对话式人工智能应用中的有效性而生成的。每个数据集条目包含系统提示、对话历史和黄金响应。这些响应通过LLM-As-a-Judge以及人工评估进行验证。该数据集规模为模拟的100个对话，其任务是评估对话式人工智能中的图形遍历与节点选择方法。

This dataset is generated to simulate real-world workflows for evaluating the effectiveness of the Performant Agentic Framework (PAF) in conversational artificial intelligence applications. Each dataset entry contains system prompts, conversation histories, and golden responses. These responses are validated via LLM-As-a-Judge and human evaluation. The dataset comprises 100 simulated conversations, with the task of evaluating graph traversal and node selection methods in conversational AI.