Two Birds, One Projection: Harmonizing Safety and Utility in LVLMs via Inference-time Feature Projection

Documenti analoghi

• Angular Gradient Sign Method: Uncovering Vulnerabilities in Hyperbolic Networks
  di: Jo, Minsoo, et al.
  Pubblicazione: (2025)
• Attention-space Contrastive Guidance for Efficient Hallucination Mitigation in LVLMs
  di: Jo, Yujin, et al.
  Pubblicazione: (2026)
• Generalized and Personalized Federated Learning with Black-Box Foundation Models via Orthogonal Transformations
  di: Kong, Eun Gyung, et al.
  Pubblicazione: (2025)
• Contrastive Residual Energy Test-time Adaptation
  di: Han, Yewon, et al.
  Pubblicazione: (2025)
• Retaining and Enhancing Pre-trained Knowledge in Vision-Language Models with Prompt Ensembling
  di: Kim, Donggeun, et al.
  Pubblicazione: (2024)