'Explaining RL Decisions with Trajectories': A Reproducibility Study