object hallucination

A phenomenon in neural networks where models incorrectly predict or generate objects or features that do not exist in the input data, often related to overfitting or misinterpretation.

4 papers