共计 42 篇文章
2026
Structural Graph Probing of Vision-Language Models
Q-MLLM： Vector Quantization for Robust Multimodal Large Language Model Security
Hallucination as Exploit： Evidence-Carrying Multimodal Agents
Vision Transformers Need More Than Registers
DYNAMIC MULTIMODAL ACTIVATION STEERING FOR HALLUCINATION MITIGATION IN LARGE VISION-LANGUAGE MODELS
Seeing Far and Clearly： Mitigating Hallucinations in MLLMs with Attention Causal Decoding
HALLUCINATION-AWARE INTERMEDIATE REPRESENTATION EDIT IN LARGE VISION-LANGUAGE MODELS
Beyond the Global Scores： Fine-Grained Token Grounding as a Robust Detector of LVLM Hallucinations
SAVE： Sparse Autoencoder-Driven Visual Information Enhancement for Mitigating Object Hallucination
Locate-then-Sparsify： Attribution Guided Sparse Strategy for Visual Hallucination Mitigation