Researchers introduce a new method to efficiently differentiate large language models (LLMs) in a black-box setting using ...