Xiaoyong Zhu

large language model (3) multimodal large language model (2) factuality evaluation (2) visual question answering (2) chinese language (2) image captioning (1) multimodal interaction (1) task generalization (1) adversarial attack (1) hidden state (1) safety evaluation (1) jailbreak attack (1) retrieval-augmented generation (1) large vision-language model (1) knowledge capacity (1) safety benchmark (1) modality interaction (1) llm agent (1) safety mechanism (1) activation pattern (1)