learned representations

The high-level abstractions or features that a model has learned from data. These representations are crucial for tasks such as classification or generation, as they encapsulate the essential characteristics of the input data.

13 papers