PostAlign: Multimodal Grounding as a Corrective Lens for MLLMs

Documenti analoghi

• An Image Is Worth 1000 Lies: Adversarial Transferability across Prompts on Vision-Language Models
  di: Luo, Haochen, et al.
  Pubblicazione: (2024)
• Influencer Backdoor Attack on Semantic Segmentation
  di: Lan, Haoheng, et al.
  Pubblicazione: (2023)
• Self-Discovering Interpretable Diffusion Latent Directions for Responsible Text-to-Image Generation
  di: Li, Hang, et al.
  Pubblicazione: (2023)
• As Firm As Their Foundations: Can open-sourced foundation models be used to create adversarial examples for downstream tasks?
  di: Hu, Anjun, et al.
  Pubblicazione: (2024)
• Which Model Generated This Image? A Model-Agnostic Approach for Origin Attribution
  di: Liu, Fengyuan, et al.
  Pubblicazione: (2024)