datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
false_absence_detection_v01ClarusC64/false_absence_detection_v01
Dataset summary
This dataset tests whether models treat invisible entities as gone or still present.In each sequence, a target leaves the camera view.Some exits are real.Some are false absences with clear evidence that the target remains in the container.
Goal
check if models infer continued presence from indirect cues
avoid treating every disappearance as an exit
keep spatial grounding under occlusion and clutter
Key signals
absence_tag:… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/false_absence_detection_v01.reflection_boundary_challenge_v01ClarusC64/reflection_boundary_challenge_v01
Dataset summary
This dataset tests whether models handle mirrors and reflections without breaking container logic.Scenes contain real objects and mirrored views.Some reflections are consistent with the room.Others violate boundaries or basic geometry.
Main goals
detect when a reflection conflicts with the layout
keep track of entities visible only in mirrors
avoid placing entities across walls or through barriers
respect gravity and… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/reflection_boundary_challenge_v01.