ACM MM 2025Detecting Violations of Physical Common Sense in Images: A Challenge Dataset and Effective ModelWeibin Wu, Zitong Wang, Zhengjie Luo, Wenqing Chen, Zibin ZhengFirst student authorACM Multimedia 2025 (CCF-A) PDF Project
arXivDelta-K: Boosting Multi-Instance Generation via Cross-Attention AugmentationZitong Wang, Zijun Shen, Haohao Xu, Zhengjie Luo, Weibin WuarXiv preprint PDF arXiv
ICML SpotlightPonderLM-2: Pretraining LLM with Latent Thoughts in Continuous SpaceBoyi Zeng, He Li, Shixiang Song, Yixuan Wang, Zitong Wang, Ziwei He, Xinbing Wang, Zhouhan LinICML 2026 Spotlight (top 2.2%) PDF
PreprintEvery Step a Thought: Implicit Visual Reasoning in Diffusion Language ModelsZitong Wang, Haohao Xu, Zijun Shen, Weibin Wu, Zibin ZhengUnder review
arXivAdaPonderLM: Gated Pondering Language Models with Token-Wise Adaptive DepthShixiang Song, He Li, Zitong Wang, Boyi Zeng, Feichen Song, Yixuan Wang, Ziwei He, Zhouhan LinarXiv preprint PDF arXiv
arXivPretraining with Token-Level Adaptive Latent Chain-of-ThoughtBoyi Zeng, Yiqin Hao, He Li, Shixiang Song, Feichen Song, Zitong Wang, Siyuan Huang, Yi Xu, Ziwei He, Xinbing Wang, Zhouhan LinarXiv preprint PDF arXiv