AI RESEARCH
When Sinks Help or Hurt: Unified Framework for Attention Sink in Large Vision-Language Models
arXiv CS.CV
•
ArXi:2604.03316v1 Announce Type: new Attention sinks are defined as tokens that attract disproportionate attention. While these have been studied in single modality transformers, their cross-modal impact in Large Vision-Language Models (LVLM) remains largely unexplored: are they redundant artifacts or essential global priors? This paper first categorizes visual sinks into two distinct categories: ViT-emerged sinks (V-sinks), which propagate from the vision encoder, and LLM-emerged sinks (L-sinks), which arise within deep LLM layers.