Deceptive Automated Interpretability: Language Models Coordinating to Fool Oversight Systems

Documenti analoghi

• Applying Refusal-Vector Ablation to Llama 3.1 70B Agents
  di: Lermen, Simon, et al.
  Pubblicazione: (2024)
• Too Big to Fool: Resisting Deception in Language Models
  di: Samsami, Mohammad Reza, et al.
  Pubblicazione: (2024)
• CryptoFormalEval: Integrating LLMs and Formal Verification for Automated Cryptographic Protocol Vulnerability Detection
  di: Curaba, Cristian, et al.
  Pubblicazione: (2024)
• Leveraging Large Language Models for Building Interpretable Rule-Based Data-to-Text Systems
  di: Warczyński, Jędrzej, et al.
  Pubblicazione: (2025)
• Language Model Re-rankers are Fooled by Lexical Similarities
  di: Hagström, Lovisa, et al.
  Pubblicazione: (2025)