[notnullorvoid]

While I probably wouldn't classify it as cheating, it is an even bigger signal of concern for model quality.

Cheating by breaking the rules at least implies some learned patterns.

Repeating training data verbatim for narrow cases like this implies that the model is overfitting.