BitBypass: A New Direction in Jailbreaking Aligned Large Language Models with Bitstream Camouflage

Documenti analoghi

• Is On-Device AI Broken and Exploitable? Assessing the Trust and Ethics in Small Language Models
  di: Nakka, Kalyan, et al.
  Pubblicazione: (2024)
• A Machine Learning-Based Framework for Assessing Cryptographic Indistinguishability of Lightweight Block Ciphers
  di: Dani, Jimmy, et al.
  Pubblicazione: (2024)
• LiteLMGuard: Seamless and Lightweight On-Device Prompt Filtering for Safeguarding Small Language Models against Quantization-induced Risks and Vulnerabilities
  di: Nakka, Kalyan, et al.
  Pubblicazione: (2025)
• ForgeDAN: An Evolutionary Framework for Jailbreaking Aligned Large Language Models
  di: Cheng, Siyang, et al.
  Pubblicazione: (2025)
• CCJA: Context-Coherent Jailbreak Attack for Aligned Large Language Models
  di: Zhou, Guanghao, et al.
  Pubblicazione: (2025)