"The top of your folding range"

I actually do not know how to bluff with a pot limit...

thats okay because this is the Fixed Limit section

and OP i dont really have an answer because it all depends on opponents and cards. but generally I don't bluff a lot and when i do its something as simple as AKo on low boards. not that it could be called bluffing rather than valuebetting on my limits.

IMO ideas which sound like "bluff n% of your range", or either "top of your range" / "bottom of your range" are wrong. These come from a model of poker which tries to map cards onto an order of clear cut chances, and once you realize how complicated poker is you can see how bad that model applies to it.
First it's hard to establish a clear ranking of the hands. In which way are they the top/bottom ? Regarding showdown value ? Or draw equity ? Or implied odds ? Maybe relative fold equity ? Each of these qualities are mixed in versatile and often opposing ratios among hands in a range and they often. A low pair on some boards may have high showdown value and low implied odds +draw equity, on other boards like A347 a 55 with a flushdraw may have low SD value but a shitload of implied odds and draw equity. But its still that pair of fives, if you rank your hands in absolute hand strength as people usually do, like nothing<pairs<overpairs<2pairs etc.. .In the latter case in some spots you'll make a huge mistake like overrating that hand as a so called 'bluffcatcher', and maybe underrating it as a 'bluff'.
Secondly you can't just take on quality and create a ranking based upon it. You pick draw equity for example, you'll have only draws in your bluffing range, and on some boards when all draws hit you'll have no bluffs. If you go for SD value, you might end up making only worse hands fold. If you go for relative fold equity say you bluff with the nut low then you surely make better hands fold, but you may not have enough fold equity to bluff with plain air.

Individual hands move around different ranges as cards peel off. (Maybe this is what kav is saying.) A turn bluff with 6 high can turn into the stone cold nuts on the river, a TP turn value bet can turn into a questionable bluff catcher on the river. This means you have to reevaluate hands constantly. Only on the river is it clear what is the "bottom of your range". On any particular street, given a particular board, you can probably determine how many, and which hands to bluff fairly correctly, but always reevaluate when the next card peels. Unlike kav, I believe it is possible, while not easy, to define "top of folding range".

Another point. This is more of a speculation than established truth: Even good players call too much when faced with a check-raise bluff. They look at the pot odds they get and then determine how much to call. This is correct when faced with a bet (donk), but not when faced with a check-raise. The result is that they become exploitable. You can conceivably move some bluffs from a check-raise to a donk-bluff, especially on scare cards. They get worse odds and possibly call less. The result would be more succeeding bluffs and cheaper failures. This is an exploitative (and exploitable) move unless you balance it carefully.

You say you vastly underbluff on the river. This is easily solved. Tend to bluff (on turn) with draws containing low cards (a number of these easily belong to the top of the folding range on turn (because you (barely) don't have sufficient odds to call). When they bust on the river, just bluff with them. A small ace high flush draw, especially with a good kicker or a small pair is better used as a bluff-catcher on the river when they bust. They are senseless to bluff with.