"Frontier Models are Capable of In-context Scheming."

Alexander Meinke et al. (2024)

Details and statistics

DOI: 10.48550/ARXIV.2412.04984

access: open

type: Informal or Other Publication

metadata version: 2025-01-14