HPC 支持前置:基于文档与历史工单的 Agentic LLM 助手
原标题:Before the Ticket: Grounding HPC User Support in Documentation and Resolved Tickets with an Agentic LLM Assistant
AI 摘要
普渡大学 Rosen 计算中心(RCAC)的研究者提出一个面向 HPC 用户支持的 agentic LLM 助手,在用户提交工单之前介入:一个编排器通过 Model Context Protocol 调用两个工具——对实时文档做 RAG,以及用 QLoRA 在 11,244 条历史工单问答对上微调的 Llama-3-8B 模型,并合并两者答案。所有模型均在本地或校内托管服务上运行,以满足工单数据不得离开机构的要求。在 400 条留出工单上的评估显示,将工单解决方案改写为自助指南后微调的模型与人工解决方案一致性最高(BERTScore F1 提升 0.06–0.27),但端到端合并没有超过单个工具,且 78% 的回答跟随文档工具。
正文节选
Before the Ticket: Grounding HPC User Support in Documentation and Resolved Tickets with an Agentic LLM Assistant Abstract Most questions that reach an HPC help desk have already been answered, in the center’s user guides or in a ticket someone else filed last semester; users still file new tickets because the guides are hard to navigate and the ticket history is not theirs to search. We present an assistant that answers from both sources: an orchestrator calls two tools over the Model Context P