Here is a city size and happiness generation comparision from my actual game.
I could imagine one of the made suggestions could help:
Remove Crime: I will compare some values in my game, if removing crime and not replacing it with something other might be a solution.
----> Completly removing crime would lead to 36 more happiness, sounds too much, but to compare it better, we remove also the 46 happiness from Religious beliefs (pacifismn with india on a pangäa map is simply overpowered).
Iam at the moment at +13 happiness... after removing both variables, I would stay at 3 happiness.... sounds realistic for the amount of big cities I have.
This would be maybe a solution, but I still think we should first try the Food/Hammer crime fighting, cause it sounds interesting and logical.
It will not be a solution to the big cities problem, cause even every building is constructed in my cites, I get more and more illiteracy, boredom and poverty. (If you wonder why I have only 10 urbanization, my priority in this game is: cultural specialists, tiles, scientists, rest of specialists. All my specialists are improved by mastery and humanismn (+1 culture per specialist). So the boredom should be really low, instead it rises as every other aspect.
In my opinion, my cities are not that big for a indian nation, bigger than any other civ (all other civs stopped their growth at around average 20 pop, except the celts, which have pantheon belief +2 happiness per celid hall), but I reached this size only by my unrealistic belief benefits. I hope this shows, how broken the exponential modifier by population is.