A classification model is found to be largely underfitting the training data. Which of the following are correct? 1.Training error is low, validation error is high
2.Training error is low, validation error is low
3.Training error is high, validation error is low
4.Training error is high, validation error is high