metadata
language:
- ko
license: apache-2.0
size_categories:
- n<1K
task_categories:
- visual-question-answering
- question-answering
- image-text-to-text
About this data
KOFFVQA is a general-purpose VLM benchmark in the Korean language. For more information, refer to our leaderboard page and the official evaluation code.
This contains the data for the benchmark consisting of images, their corresponding questions, and response grading criteria. The benchmark focuses on free-form visual question answering, evaluating the ability of large vision-language models to generate comprehensive and accurate text responses to questions about images.
Citation
@article{kim2025koffvqa,
title={KOFFVQA: An Objectively Evaluated Free-form VQA Benchmark for Large Vision-Language Models in the Korean Language},
author={Kim, Yoonshik and Jung, Jaeyoon},
journal={arXiv preprint arXiv:2503.23730},
year={2025}
}