2 Smoothing and Evaluation 20 Points Grading comment: We saw that test data could consist of some unseen words that you have never observed in the training data. To fix the issue, you are trying to use add- smoothing. Question 2.1 Q2.1 5 Points Grading comment: Think of what would happen if we use too small based on your work on previous questions. What would be the most likely to happen to the perplexity of the ngram model? Choice 1 of 2: goes higher Choice 2 of 2: goes lower