Documenti analoghi

• From Shortcuts to Triggers: Backdoor Defense with Denoised PoE
  di: Liu, Qin, et al.
  Pubblicazione: (2023)
• MM-PoE: Multiple Choice Reasoning via. Process of Elimination using Multi-Modal Models
  di: Chakrabarty, Sayak, et al.
  Pubblicazione: (2024)
• Test-time Backdoor Mitigation for Black-Box Large Language Models with Defensive Demonstrations
  di: Mo, Wenjie, et al.
  Pubblicazione: (2023)
• Securing Multi-turn Conversational Language Models From Distributed Backdoor Triggers
  di: Tong, Terry, et al.
  Pubblicazione: (2024)
• Multiple Heads are Better than One: Mixture of Modality Knowledge Experts for Entity Representation Learning
  di: Zhang, Yichi, et al.
  Pubblicazione: (2024)