The Mirage of Performance Gains: Why Contrastive Decoding Fails to Mitigate Object Hallucinations in MLLMs?

Documenti analoghi

• ClearSight: Visual Signal Enhancement for Object Hallucination Mitigation in Multimodal Large language Models
  di: Yin, Hao, et al.
  Pubblicazione: (2025)
• Lifting the Veil on Visual Information Flow in MLLMs: Unlocking Pathways to Faster Inference
  di: Yin, Hao, et al.
  Pubblicazione: (2025)
• Mitigating Object Hallucinations in MLLMs via Multi-Frequency Perturbations
  di: Li, Shuo, et al.
  Pubblicazione: (2025)
• Mitigating Hallucinations in Large Vision-Language Models with Instruction Contrastive Decoding
  di: Wang, Xintong, et al.
  Pubblicazione: (2024)
• Delve into Visual Contrastive Decoding for Hallucination Mitigation of Large Vision-Language Models
  di: Lee, Yi-Lun, et al.
  Pubblicazione: (2024)