Prototypicality Bias Reveals Blindspots in Multimodal Evaluation Metrics

Documenti analoghi

• TikZilla: Scaling Text-to-TikZ with High-Quality Data and Reinforcement Learning
  di: Greisinger, Christian, et al.
  Pubblicazione: (2026)
• Qalam : A Multimodal LLM for Arabic Optical Character and Handwriting Recognition
  di: Bhatia, Gagan, et al.
  Pubblicazione: (2024)
• Uncovering Conceptual Blindspots in Generative Image Models Using Sparse Autoencoders
  di: Bohacek, Matyas, et al.
  Pubblicazione: (2025)
• 3DFroMLLM: 3D Prototype Generation only from Pretrained Multimodal LLMs
  di: Ahmed, Noor, et al.
  Pubblicazione: (2025)
• ScImage: How Good Are Multimodal Large Language Models at Scientific Text-to-Image Generation?
  di: Zhang, Leixin, et al.
  Pubblicazione: (2024)