ראיה אינה ידיעה: האם VLMs יודעים מתי לא לענות שאלות חלליות (ולמה)?

תקציר מקורי באנגליתarXiv:2605.30557v2 Announce Type: replace-cross Abstract: Spatial reasoning benchmarks typically evaluate whether vision-language models can derive the correct answer from a visual observation. Yet in real 3D environments, the observation itself may be unreliable: occlusion can remove task-relevant evidence, while perspective can make visible geometry misleading. Reliable spatial reasoning therefore requires more than answering a question correctly. A model must also assess whether its current observation provides sufficient and trustworthy evidence for that answer. We introduce SPATIALUNCERTAIN, a controlled evaluation framework for studying viewpoint-dependent observational uncertainty. We study two complementary failure modes: missing evidence caused by occlusion and misleading evidence