maximum yield, don't trim off any bud sites... though the larfy stuff will only be good for extracts, cooking etc...
your light will dictate yield more than anything else. (and ambient co2 beign limiting factor in most cases...)
so, it depends on your intentions with the output. if you do extracts and stuff, i'd leave it all, or most of it. you can learn where the buds turn to absolute shit relative to your environment, and in future choose to cut that shit off, but it won't make a huge difference above. i.e. 100% of that growth will not be redirected elsewhere, but maybe some small percentage does? i have no use for larfy stuff so i remove lower bud sites that amount to 5-10 pistils and shit nug at end. that very first set of limbs that grow usually only grow trash even if you work your ass off to keep it level with the rest.. still pitiful. above that it seems to share better.
lst is alwys a good idea... more light hitting more areas, less wasted light etc etc... = more growth. also reduces risk of microbial growth you don't want. you can proll lst the side branches but it would take some HST for the apical meristem (trunk), which may or may not be helpful... more likely helpful with photoperiods than autos on a biological timer.
if you haven't done this stuff... a good option to to see an outcome without all the potential sophistry. no quality reseaarch has looked into effeciveness of these methods, their impact on time invested and outcomes... no comparison to control groups.. jsut a bunch of ppl that do somethign and feel it helps. i do think systematicall organizing the canopy for optimal light reception (as long as additional time required does not wipe out all those gains) is a helpful thing. beyond that it gets unclear.
i'm not trying to set production records.. i'm trying to grow buds conststently in a waay i appreciate the most... sometimes that does not align with highest yield. YMMV too