EDIT: Enhancing Vision Transformers by Mitigating Attention Sink through an Encoder-Decoder Architecture

Documenti analoghi

• Artifacts and Attention Sinks: Structured Approximations for Efficient Vision Transformers
  di: Lu, Andrew, et al.
  Pubblicazione: (2025)
• Mitigating Hallucination in Large Vision-Language Models through Aligning Attention Distribution to Information Flow
  di: Zhao, Jianfei, et al.
  Pubblicazione: (2025)
• SAGE: Sink-Aware Grounded Decoding for Multimodal Hallucination Mitigation
  di: Shukla, Tripti, et al.
  Pubblicazione: (2026)
• On the Nature of Attention Sink that Shapes Decoding Strategy in Omni-LLMs
  di: Yoo, Suho, et al.
  Pubblicazione: (2026)
• Tell Model Where to Look: Mitigating Hallucinations in MLLMs by Vision-Guided Attention
  di: Zhao, Jianfei, et al.
  Pubblicazione: (2025)