That’s not surprising since the GEFS mean tends to closely mirror the operational. If the 00z GFS swings back the other direction, I guarantee the GEFS will do likewise (if not, then the operational GFS would almost certainly be wrong).
The GEFS mean is even flatter than the operational run. Doesn’t seem to have any ridge at all.
Though the GEFS can be under-dispersed, so a systemic error could throw the whole mean off..theoretically.