METIS：面向深入探究与解决方案的导师引擎
METIS: Mentoring Engine for Thoughtful Inquiry & Solutions
AI总结：
METIS通过工具增强和阶段意识功能，帮助本科生从想法发展到论文，优于GPT-5和Claude Sonnet 4.5，在文档基础阶段表现更佳。
摘要：
许多学生缺乏接触专家研究导师的机会。我们探讨人工智能导师是否能帮助本科生从一个想法发展到一篇论文。我们构建了METIS，一个工具增强、阶段意识的助手，具备文献搜索、精选指南、方法学检查和记忆功能。我们通过六个写作阶段使用LLM-as-a-judge配对偏好、学生-角色评分表、短多轮辅导和证据/合规检查，评估METIS与GPT-5和Claude Sonnet 4.5的性能。在90个单轮提示中，LLM评委在71%的情况下更倾向于选择METIS而非Claude Sonnet 4.5，在54%的情况下更倾向于选择METIS而非GPT-5。学生评分（清晰度/可操作性/约束适配度；90个提示x3位评委）在各阶段均更高。在多轮会话（五个场景/代理）中，METIS的最终质量略高于GPT-5。收益集中在文档基础阶段（D-F），与阶段意识路由一致。阶段意识失败模式包括过早工具路由、浅层基础和偶尔的阶段误分类。
英文摘要：
Many students lack access to expert research mentorship. We ask whether an AI mentor can move undergraduates from an idea to a paper. We build METIS, a tool-augmented, stage-aware assistant with literature search, curated guidelines, methodology checks, and memory. We evaluate METIS against GPT-5 and Claude Sonnet 4.5 across six writing stages using LLM-as-a-judge pairwise preferences, student-persona rubrics, short multi-turn tutoring, and evidence/compliance checks. On 90 single-turn prompts, LLM judges preferred METIS to Claude Sonnet 4.5 in 71% and to GPT-5 in 54%. Student scores (clarity/actionability/constraint-fit; 90 prompts x 3 judges) are higher across stages. In multi-turn sessions (five scenarios/agent), METIS yields slightly higher final quality than GPT-5. Gains concentrate in document-grounded stages (D-F), consistent with stage-aware routing and groundings failure modes include premature tool routing, shallow grounding, and occasional stage misclassification.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
