Cheng Shi

3 papers ShanghaiTech University

• Discovering Compositional Hallucinations in LVLMs
• Eyes Wide Open: Ego Proactive Video-LLM for Streaming Video
• Vision Function Layer in Multimodal LLMs