data memorization

Data memorization refers to a model's tendency to overfit to the training data by memorizing specific examples instead of learning generalizable patterns. It is often identified as an issue that can harm a model's ability to generalize to new, unseen examples.