README.md

MultiVerse: A Multi-Turn Conversation Benchmark for Evaluating Large Vision and Language Models

🏠 Homepage | 🤗 Dataset | 📄 Arxiv | 📕 PDF

📰 News

• The evaluation code will be released soon! Stay tuned!

Acknowledgement

This work was partly supported by Institute of Information & communications Technology Planning & Evaluation (IITP) grant funded by the Korea government(MSIT) (RS-2022-II220641, XVoice: Multi-Modal Voice Meta Learning)

Citation

If you find the resources in this repository useful, please cite our work:

@misc{lee2025multiversemultiturnconversationbenchmark,
      title={MultiVerse: A Multi-Turn Conversation Benchmark for Evaluating Large Vision and Language Models}, 
      author={Young-Jun Lee and Byung-Kwan Lee and Jianshu Zhang and Yechan Hwang and Byungsoo Ko and Han-Gyu Kim and Dongyu Yao and Xuankun Rong and Eojin Joo and Seung-Ho Han and Bowon Ko and Ho-Jin Choi},
      year={2025},
      eprint={2510.16641},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2510.16641}, 
}