Lost on Campus benchmark aims to evaluate ESR of Vision-Language Model (VLM) in large-scale real-world outdoor 3D environments reconstructed by 3D Gaussian Splatting. Our benchmark introduces a unified reasoning-action evaluation framework that seamlessly integrates diagnostic QA for isolated reasoning with closed-loop interactive navigation for active reasoning under multimodal instructions.
