Counterfactual explanations for Reinforcement Learning agents on Atari - View it on GitHub