Linguists have, however, found clever ways to test LLMs' underlying knowledge, in effect tricking them with probing tests.

然而，语言学家已经找到了测试大型语言模型的潜在知识的巧妙办法，实际上就是用试探性的测试来欺骗它们。