The dataset is currently empty. Upload or create new data files. Then, you will be able to explore them in the Dataset Viewer.

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

CivilBench

CivilBench: Assessing the Civil Intelligence of Vision-Language Models

CivilBench is a benchmark designed to systematically evaluate the capabilities of vision-language models (VLMs) in civil and structural engineering.

The benchmark covers three major capability categories:

  • Sequential Reasoning
  • Image-based Structural Understanding
  • Open-world Understanding

These categories further consist of 14 fine-grained evaluation dimensions covering structured reasoning, structural perception, and engineering-oriented reasoning.

Dataset Availability

The CivilBench dataset and associated resources will be publicly released upon acceptance of the paper.

Paper

The manuscript is currently under review.

Citation

Citation information will be provided upon publication.

Downloads last month
6