The Madden-Julian Oscillation (MJO) is an important driver of midlatitude subseasonal variability and predictability through its downstream effects on atmospheric circulation. However, there is considerable literature that also highlights El Niño Southern Oscillation (ENSO) teleconnections as a driver of midlatitude subseasonal variability, particularly during boreal winter in the Northern Hemisphere. These results raise the question of the relative importance of each phenomena for midlatitude subseasonal predictability. In this work, we explore the relative contribution of the MJO and ENSO to predictability of circulation anomalies over the North Pacific in boreal winter using the pre-industrial control runs of the Community Earth System Model version 2 (CESM2) and an interpretable neural network. We find that the interpretable network generally relies on the state of ENSO, rather than the MJO, to make correct predictions on a range of subseasonal lead times and predictand averaging windows. However, when ENSO is in a neutral state, our findings indicate that three distinct MJO evolutions provide some predictive information, offering insights into opportune forecasting windows for MJO teleconnections.