Do VLMs Have a Moral Backbone? A Study on the Fragile Morality of Vision-Language Models

Documenti analoghi

• MORALISE: A Structured Benchmark for Moral Alignment in Visual Language Models
  di: Lin, Xiao, et al.
  Pubblicazione: (2025)
• Taming Knowledge Conflicts in Language Models
  di: Li, Gaotang, et al.
  Pubblicazione: (2025)
• LegalDrill: Diagnosis-Driven Synthesis for Legal Reasoning in Small Language Models
  di: Li, Tianchun, et al.
  Pubblicazione: (2026)
• On the Pros and Cons of Active Learning for Moral Preference Elicitation
  di: Keswani, Vijay, et al.
  Pubblicazione: (2024)
• Seeing but Not Believing: Probing the Disconnect Between Visual Attention and Answer Correctness in VLMs
  di: Liu, Zhining, et al.
  Pubblicazione: (2025)