On the Geometric Limits of Transformer Defenses against Obfuscation Attacks: Latent Embedding Collapse & Performance Robustness Gap

Documenti analoghi

• Defensive Prompt Patch: A Robust and Interpretable Defense of LLMs against Jailbreak Attacks
  di: Xiong, Chen, et al.
  Pubblicazione: (2024)
• Query Provenance Analysis: Efficient and Robust Defense against Query-based Black-box Attacks
  di: Li, Shaofei, et al.
  Pubblicazione: (2024)
• PhantomFetch: Obfuscating Loads against Prefetcher Side-Channel Attacks
  di: Zhang, Xingzhi, et al.
  Pubblicazione: (2025)
• Rotated Robustness: A Training-Free Defense against Bit-Flip Attacks on Large Language Models
  di: Liu, Deng, et al.
  Pubblicazione: (2026)
• Removal Attack and Defense on AI-generated Content Latent-based Watermarking
  di: Lee, De Zhang, et al.
  Pubblicazione: (2025)