返回全部动态
BIBIMBAP：面向BIM自动化编程的指令基准
原标题：BIBIMBAP: A Benchmark for Instructional BIM-Based Automated Programming
AI 摘要
研究者提出 BIBIMBAP，一个面向 Text-to-BIM 的基准，包含 100 个覆盖增删改查（CRUD）操作的任务，涉及空间、几何、拓扑、数值和概念类别。每个任务提供自然语言提示、IFC 模型、期望的结构化输出和自动化测试脚本。基线结果显示最佳性能仅 50.2%，模型在约束保持和空间推理上频繁失败，部分正确输出还伴随错误推理。
以上摘要由 AI 生成，可能存在误差。事实请以原文为准。
正文节选
BIBIMBAP: A Benchmark for Instructional BIM-Based Automated Programming Abstract Large Language Models (LLMs) are emerging as natural-language interfaces for Building Information Modeling (BIM), successfully achieving information retrieval, but no benchmark exists to evaluate their ability to edit Industry Foundation Classes (IFC). We introduce BIBIMBAP, a Text-to-BIM benchmark with 100 curated tasks covering Create-Read-Update-Delete (CRUD) operations across spatial, geometric, topological, num
发布时间：2026-11-10 13:00
抓取时间：2026-11-10 16:37
来源机构：arXiv