作者位次 #4/10ACL 2026Findings· Findings of ACL
Global Context or Local Detail? Adaptive Visual Grounding for Hallucination Mitigation
Yubo Jiang · Xin Yang · Abudukelimu Wuerkaixi · Zheming Yuan · Xuxin Cheng · Cao Liu · Ke Zeng · Fengying Xie · Zhiguo Jiang · Haopeng Zhang
摘要
Large vision–language models (LVLMs) excel at multimodal reasoning but still suffer from object-existence hallucinations when multi-step deliberation decouples from visual evidence. Think-with-Images (TwI) attempts to counter this by generating auxiliary observations (e.g., zoomed crops or highli...