OK, lets try a hypothetical.
You're DMing. Your (4th level PCs) players are exploring a dungeon. They're dragging their heels, and have been having a reasonably easy time of it; HP are close to full, the Barbarian has 1 rage left, most Hit Dice have been used, plus a few 2nd level slots. The PCs have enough time remaining for a Short rest however. The BBEG isnt far off (a few rooms away), and the players are getting a little side tracked with table talk and deciding what to do next. You decide to throw a quick 'random' encounter (to focus their attention, and weaken them a little for the BBEG) is in order. The session is due to end in around an hour or two.
Well, first off, I wouldn’t force an encounter just because I felt like the players were doing too well and the boss was going to be too easy, so this hypothetical is already pretty dubious for me.
You had nothing planned, so you roll on a table provided in the module you're running. You roll an '89'. The players dont see the roll, and have no access to the chart even if they did.
The result tells you '1d6 Werewolves'. You roll a 6. Werewolves don't really make sense in the context of the adventure.
If werewolves don’t make sense in the context of the adventure, why would I have included them on the random encounter table?
You know your PCs have no silver or magical weapons, and this encounter will almost certainly result in a TPK.
Again, why would I have even put 1d6(!!) werewolves on the random encounter table if I knew this?
You also note that on the same chart, the result of an '88' gives a much more manageable encounter for them, and also ties into the BBEG a lot better (it's the BBEG's henchman, obviously sent out to 'deal' with the PCs).
Do you:
A) Just throw the Werewolves at them because 'the dice say so'?
or
B) Alter the result to an 88, and use the more sensible encounter, that better serves the story and the outcome desired (and the reason you rolled in the first place) and that wont result in a crushing TPK?
Which is it?
Now, here’s what I would consider the important part of your question - “do you fudge rolls on random encounter tables?” To that question, I would answer “sometimes.” I try to seed my random encounter tables so that they are well-suited to the location and whatever I roll will fit. But, when a random encounter comes up, I don’t always even roll on the table. Sometimes I’ll just pick a result. Or if like a I’ve rolled the same encounter several times, I might decide to do a different one. This, in my view, is categorically different than fudging attack rolls, saves, checks, damage rolls, HP values, etc. because it doesn’t take agency from the players. The players’ choices have no real way of impacting the results of a random encounter roll (though they do impact
when such a roll is made, hence why I wouldn’t force a “random” encounter that wasn’t really random). Whether the result is chosen by me or a die, doesn’t make a practical difference to their experience, whereas a monster hitting more or less often or dying faster or slower because I decided so does.