---
license: apache-2.0
---

# ActiView: Evaluating Active Perception Ability for Multimodal Large Language Models

<p align="center"><img src="./figures/figure1.png" alt="Image" width=60% ></p>

<div align="center">
    <a href="https://wangphoebe.github.io/actiview_homepage/">🌐<strong>Homepage</strong></a> | <a href="https://arxiv.org/abs/2410.04659">📖<strong>arXiv</strong></a> | <a href="https://huggingface.co/datasets/wangphoebe/ActiView">🤗<strong>Dataset</strong></a>
</div>

## Overview

This is benchmark for paper "ActiView: Evaluating Active Perception Ability for Multimodal Large Language Models" ([arxiv](https://arxiv.org/abs/2410.04659))

Please refer to this [github](https://github.com/THUNLP-MT/ActiView) repo **for detail.**

To use this dataset, please download all the files and place them under the asset dir as in the above repo. The Archive_imgs.zip file contains 3 dirs of necessary image files for evaluating. Please unzip this file and place the resulting dirs under the asset dir in our github repo. 



## Citation
If you find our data useful, please consider citing:
```
@misc{wang2024activiewevaluatingactiveperception,
      title={ActiView: Evaluating Active Perception Ability for Multimodal Large Language Models}, 
      author={Ziyue Wang and Chi Chen and Fuwen Luo and Yurui Dong and Yuanchi Zhang and Yuzhuang Xu and Xiaolong Wang and Peng Li and Yang Liu},
      year={2024},
      eprint={2410.04659},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2410.04659}, 
}
```