It's entirely possible that the GFS is right. If the spacing between the main shortwave and upstream shortwave are too close, a sheared out mess is possible.
So this above is also not surprising. The 00z GFS is probably the worst result, with narrow wavelength and extremely sheared mess by the time it reaches New England. So it stands to reason that the GEFS would have plenty of members that have greater spacing and more amplified solutions, dragging the mean north with it.