English(EN) How six LLMs scored on 387 real construction questions, with and without retrieval

大型语言模型难以处理特定公司数据，但检索可将准确率提升至 90%

作者 PulseAugur 编辑部 · [1 个来源] · 2026-11-03 08:21

一项基准测试对六种大型语言模型 (LLM) 进行了 387 个特定建筑相关问题的测试，评估了它们在独立运行和检索增强生成 (RAG) 情况下的表现。结果显示，使用检索功能后性能大幅提升，模型单独的平均准确率为 19%，而在提供相关源文档后准确率达到 90%。研究强调，访问特定公司数据比选择哪种 LLM 更为关键，并指出模型经常会臆造行业标准答案，即使公司特定政策不同，且未表明其在猜测。 AI

影响强调了检索增强生成 (RAG) 系统在弥合通用 LLM 知识与特定企业数据之间差距的关键需求。

排序理由基准研究，评估大型语言模型在领域特定数据上的表现。[lever_c_demoted from research: ic=1 ai=1.0]

在 dev.to — LLM tag 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 23 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 基准研究，评估大型语言模型在领域特定数据上的表现。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: product, other; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: Breaking (< 6h); Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

报道来源 [1]

dev.to — LLM tag TIER_1 English(EN) · Metal America · 2026-11-03 08:21

How six LLMs scored on 387 real construction questions, with and without retrieval

<p>General-purpose models know a lot about construction in the abstract. Ask one what a vapor barrier does or why snow load matters and you'll get a good answer. But a construction business runs on specifics: how long a manufacturer takes, what a permit requires, which installati…

本文如何被排名

报道来源 [1]

How six LLMs scored on 387 real construction questions, with and without retrieval

相关实体

相关话题