Having 2 to 3 percent difference between the training and testing is not a sign of overfitting especially in models like random forest which is very robust in nature your model doing exceptional on training data and also on testing.In this case model make good.Overfitting occurs when there is a large gap between training and testing.If you still want to make sure if the model is not overfitting you can check with other models like decision tree or logistic regression and compare the scores.