**Describe the bug** In explanation.py when creating perf dictionaries for interpret show method actual_score and predicted_score need to be probabilities **To Reproduce** See review comments in PR https://github.com/interpretml/interpret-community/pull/445 for details **Expected behavior** We need to get the probabilites and set them here for classification case