Image Tokens Matter: Mitigating Hallucination in Discrete Tokenizer-based Large Vision-Language Models via Latent Editing

Ähnliche Einträge

• SeCoKD: Aligning Large Language Models for In-Context Learning with Fewer Shots
  von: Wang, Weixing, et al.
  Veröffentlicht: (2024)
• MINT: Mitigating Hallucinations in Large Vision-Language Models via Token Reduction
  von: Wang, Chao, et al.
  Veröffentlicht: (2025)
• Feature Distribution Shift Mitigation with Contrastive Pretraining for Intrusion Detection
  von: Wang, Weixing, et al.
  Veröffentlicht: (2024)
• PAINT: Paying Attention to INformed Tokens to Mitigate Hallucination in Large Vision-Language Model
  von: Arif, Kazi Hasan Ibn, et al.
  Veröffentlicht: (2025)
• Generalized Categories Discovery for Long-tailed Recognition
  von: Li, Ziyun, et al.
  Veröffentlicht: (2023)