Dataset Viewer
Auto-converted to Parquet Duplicate
Search is not available for this dataset
image
imagewidth (px)
512
512
label
class label
256 classes
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
000
101
101
101
101
101
101
101
101
101
101
101
101
101
101
101
101
101
101
101
101
101
101
101
101
101
101
101
101
101
101
101
101
101
101
101
101
101
101
101
101
101
101
101
End of preview. Expand in Data Studio

GeoPrecisionBench

GeoPrecisionBench is a dataset for testing whether a model can judge a coordinate pin against the building it is supposed to mark.

Each sample is a rural building in the contiguous United States (CONUS), shown on NAIP aerial imagery. A coordinate pin is rendered on the image at a controlled, signed offset from the building's footprint edge. Offsets range from −5 m (inside the footprint) to 80 m.

Each sample carries labels for two tasks:

On-building verification: whether the pin lies on the building. Distance estimation: the pin's distance to the building.

Contents

file content
pins.parquet one row per pin; image = path of the render the models saw
sources.parquet one row per source building
boxes.parquet sampling boxes
answers.parquet main-run model answers (verification, distance)
consistency_answers.parquet repeats 1-4 on the human-comparison pins
gold/ building check, human on-roof answers and review, failure-mode codes
configs/ prompts, bins, scoring settings (incl. excluded sources)

Footprints: Overture Maps 2026-08-19.0. Imagery: USDA NAIP.

Downloads last month
336