GUI-Primitives: Diagnosing Spatial Reasoning Failures in Vision-Language GUI Grounding

  • 类型:arxiv
  • 标识:2608.21832
  • 链接:https://arxiv.org/abs/2608.21832
  • 主分类:multimodal
  • 形态:benchmark
  • TLDR:Computer-use agents ground natural-language instructions in screenshots to locate interface elements, yet existing benchmarks do not isolate whether models bind relational language to the correct element. We introduce GUI-Primitives, a 994-item benchmark of contrastive instruction pairs over seven spatial relations in graphical user interfaces (left/right, above/below, containment, alignment, proximity, list ordinal, occlusion). Each pair holds the screenshot and anchor fixed while changing the relation expression, so the correct target moves between two designated candidates. Five annotators
  • 副分类:evaluation
  • 待LLM分类:否
  • 来源文件
  • /inbox/tom/_candidates/2026-08-28-agent-rag-longcontext-candidates.json