AI Sandbagging: Language Models can Strategically Underperform on Evaluations

van der Weij, Teun, Hofstätter, Felix, Jaffe, Ollie · arXiv · 2024