#generalization

기계 지원 번역 초안 (Korean) for "Zero-Shot": The ability of a model to perform a task it has never been explicitly trained or shown examples for. Zero-shot learning relies on the model's generalized understanding from pretraining to handle novel tasks based on instruction alone.

“예문 초안: The model classified customer sentiment zero-shot without any labeled training examples.”