Doing some quick math, looks like Seattle (5.9) and Arizona (4.5) have the largest average 2nd half point differential. Didn't see any other team above 4.
Thanks for that link. I'd guess it's pretty high variance from year to year (mostly random).
On both sides of the ball coordinators are playing cat and mouse all the way through the game, and it's not like they really have time to meaningfully flip the whole script during half time anyway.