The test for overfitting is whether the model holds up for data outside the training set. If the biases are consistent with other real-world data, then it's no

The test for overfitting is whether the model holds up for data outside the training set. If the biases are consistent with other real-world data, then it's not overfitting.