Can Large Multimodal Models Actively Recognize Faulty Inputs? A Systematic Evaluation Framework of Their Input Scrutiny Ability

Ähnliche Einträge

• Refining Critical Thinking in LLM Code Generation: A Faulty Premise-based Evaluation Framework
  von: Li, Jialin, et al.
  Veröffentlicht: (2025)
• Don't Take the Premise for Granted: Evaluating the Premise Critique Ability of Large Language Models
  von: Li, Jinzhe, et al.
  Veröffentlicht: (2025)
• Asymmetric Co-Training for Source-Free Few-Shot Domain Adaptation
  von: Li, Gengxu, et al.
  Veröffentlicht: (2025)
• ActiView: Evaluating Active Perception Ability for Multimodal Large Language Models
  von: Wang, Ziyue, et al.
  Veröffentlicht: (2024)
• Can Visual Input Be Compressed? A Visual Token Compression Benchmark for Large Multimodal Models
  von: Peng, Tianfan, et al.
  Veröffentlicht: (2025)