Only the Envelope/AHDSR devices have a release stage, but it’s not super-intuitive since you can effectively turn them all the way down to 0ms and they will still manage to pull off this little trick.
I’ve been confused about this a couple of times too
When I’d use a AHDSR - with 100% sustain only - multiplied with a fader the theoretical result would be exactly the same as without the AHDSR. But in practice it’s not. That’s indeed pretty confusing. I can see at least one reason, why not all modulation types share that behavior. But maybe it’d be a good idea to add the special feature of the AHDSR to the documentation then. Or even better… drop the behavior for the AHDSR and introduce different loop and trigger modes 
https://forum.renoise.com/t/new-sample-modes/35982… like I suggested almost 3 years ago. 
SCNR B)