Large Language Models can Strategically Deceive their Users when Put Under Pressure

Documenti analoghi

• Frontier Models are Capable of In-context Scheming
  di: Meinke, Alexander, et al.
  Pubblicazione: (2024)
• Me, Myself, and AI: The Situational Awareness Dataset (SAD) for LLMs
  di: Laine, Rudolf, et al.
  Pubblicazione: (2024)
• Forecasting Frontier Language Model Agent Capabilities
  di: Pimpale, Govind, et al.
  Pubblicazione: (2025)
• TracrBench: Generating Interpretability Testbeds with Large Language Models
  di: Thurnherr, Hannes, et al.
  Pubblicazione: (2024)
• The Reversal Curse: LLMs trained on "A is B" fail to learn "B is A"
  di: Berglund, Lukas, et al.
  Pubblicazione: (2023)