The statistical-query model is a well-studied framework for analyzing the sample complexity of learning problems andcaptures mostknownlearning algorithms.

The statistical-query model is a well-studied framework for analyzing the sample complexity of learning problems andcaptures mostknownlearning algorithms.

The measure captures errors due to absent predicted requests as well as unpredicted actual requests; hence, predicted and actual inputs can be of arbitrary size.

It is widely acknowledged that large language models (LLMs) encode a vast reservoirofknowledge after being trained onmass data.