小模型专攻语言技能，GraphRAG反超大模型
做知识问答的AI通常需要两步：先建知识图谱，再检索回答。但现有系统一次提取就建图，噪声多、检索脆弱。RAGU把提取和整合拆开：先两轮提取实体关系，再用聚类去重、LLM总结、社区检测，让图谱更干净。关键发现是：建图需要的理解、提取、推理能力本质是语言技能，随模型规模增长很慢，不像事实知识那样依赖大模型。于是他们训练了一个70亿参数的小模型Meno-Lite-0.1，专攻语言技能，在知识图谱构建上反超320亿参数的Qwen2.5-32B（调和平均+12.5%），在英文GraphRAG任务上持平。在医疗问答基准上，RAGU的证据召回率最高达0.84，而其他系统≤0.76。它不是你明天就能直接用的工具，但揭示了一个趋势：对于特定任务，小模型通过针对性训练可以匹敌甚至超越大模型，这为资源有限的场景提供了新可能。
📄 原文摘要(英文)
Graph retrieval-augmented generation (GraphRAG) enhances large language models with structured knowledge, yet existing systems construct knowledge graphs in a single extraction pass, producing noisy entities and brittle retrieval. RAGU, an open-source modular GraphRAG engine, addresses this by separating extraction from consolidation: entities and relations pass through two-stage typed extraction, DBSCAN-backed deduplication, LLM summarization, and Leiden community detection. A key insight motivates a compact extractor: the skills an in-pipeline LLM needs - comprehension, extraction, reasoning over context - are language skills that grow only weakly with model size, unlike factual world knowledge. Accordingly, we train Meno-Lite-0.1, a 7B model optimized for language skills, which outperforms Qwen2.5-32B on knowledge-graph construction (+12.5% relative harmonic mean) and matches it on English GraphRAG tasks. On GraphRAG-Bench (Medical), RAGU retrieves the most complete context at every factoid level (evidence recall up to 0.84 vs. leq0.76) and overtakes HippoRAG2 on synthesis tasks; on multi-hop factoid QA, the apparent HippoRAG2 advantage is shown to be largely an answer-format artifact. RAGU is installable via pip install graph_ragu, runs on a single GPU, and is released under MIT. The source code is publicly available at https://github.com/RaguTeam/RAGU, and the Meno-Lite-0.1 model can be obtained from https://huggingface.co/bond005/meno-lite-0.1.