Curious what your reasoning is?
In the real world you'd be before the amp, in the loop or after the speaker (cab).
Before or after the Cab doesn't matter for Linear Time Invariant effects like most time-based effects.
Obviously the digital world allows for different options, but wondering what your approach is and what benefit you are getting from it?
Also curious about the reasons for this one - never heard of anyone doing it before. I'll have to try it, but my expectation would have been low particularly on distorted tones given how nasty they sound to me cabless.
First let me clear up any confusion. The bulk of my amped/distorted tones go through a cab. Just wanted to be clear about that. Some of it may bypass into a filter block to create AITR effect. I do confess to using this technique more and more lately as the difference is significant and cool.
Second, let me say that I am a sound engineer, so I approach my presets as if they are an extension of my DAW/mixing board. My approach (most of the time) is to very much "build" a sound, and I deploy a ton of EQ, filter and mixer blocks to get there.
I am very careful with IRs because they are essentially a continious EQ, and as such, they are reductive in nature. Once it is gone...it's gone...you can't add it back. That usually what I am looking for on a distorted/amped signal, of course. Not usually so on effects. I usually want more of a frequency range on my effects (modulation and reverb) than what I can get post IR. If I start with a clean signal, modulate and then pass them through a PEQ instead, I have a lot more control. This, of course, requires more work in terms of setting up parallel paths and then bringing everything back together so that it sounds mixed and blended as opposed to disconnected tones. It also requires a certain amount of care not to walk on other instruments or vocals in a band setting.
So, for example, even on a heavily distorted tone, I may layer a clean signal chorus or trem on top, and the only thing the listener will be wise to is that they can now also hear the effects instead of them them getting lost/creating mud. Sorry, if that may last comment may be offensive to folks, but to my ears, modulation through a driven amp, or of a distorted signal in general, is often asking for trouble. In my opinion, it is often overused by guitar players and the fastest way to have a sound person dial you back in the mix. To be honest, it usually sounds like shit through a PA. Mix it back in with a clean, effected signal and suddenly you have both discernable modulation and distortion and hopefully a bit more PA friendly. That is one example, but hopefully helps explain my motivation.
These aren't hard rules, by any means. Delay roams my grid with abandon. The phaser block can lurk at the front of a driven amp, if that is the tone I am looking for. I guess my real point, is that I would encourage anyone with a fractal to experiment with parallel paths and to not be afraid of using the mixer block to bring in clean effects, amped signals, un-amped signals, dirt, or whatever it is that you are working with, into one big happy tone tent.