1
FloorplanQA: A Benchmark for Spatial Reasoning in LLMs using Structured Representations
空间推理能力评测新标杆:用楼层平面图结构化数据挑战大模型理解与推理极限
arXiv:2507.07644v4 Announce Type: replace Abstract: We introduce FloorplanQA, a diagnostic benchmark for evaluating spatial reasoning in large languag…