2 papers
cs.CV2026
DreamRealm: Exploring and Materializing Hallucination Worlds for Trustworthy Vision-Language Models
Benlei Cui, Chunyang Chai, Yiliang Zhang +6
Multimodal large language models (MLLMs) remain prone to visual hallucinations when correct responses depend on subtle visual details. One line of post-training methods constructs…
cs.CL2026
Yuvion LLM: An Adversarially-Aware Large Language Model for Content And AI Safety
Ting Ma, Xiufeng Huang, Benlei Cui +43
As large language models are increasingly deployed in real-world systems, safety failures can still lead to harmful outputs and dangerous misuse. We argue that the essence of safet…