Well, my goal is to have a recording that I don't have to crank it up to hear the songs, then crank it down because the applause is so loud, etc. Should I maybe record with the applause peaking at 0, and then apply some sort of compression in post?
That's the path to "as good as you can make it sound", making those dynamics adjustments afterwards which allows you to check by ear to determine what works and what doesn't, how much of what to apply, and where to use what. It also allows you to use several methods in combination to better optimize the dynamic range of the recording without introducing audible artifacts. It's far more nuanced and allows for the ability to undo things and start over if you run astray, rather than using the blunt bludgeon of reducing the dynamics by using the limiter on the recorder. Those other dynamics techniques can also help avoid the need to constantly adjust the playback volume, even though they aren't working on the applause portions at all- they may be raising the level of the quiet stuff and bringing out the previous unheard details (see the parallel compression thread), or just reducing the loudness range of the musical portions between the applause sections (check out other compression and post-production sound editor technique threads here at TS).
It's not "wrong" to use the limiter on the recorder to squash the level of the applause down. If not having to take time to do the processing later is more important to you, heavy limiting of loud applause levels while recording will achieve the goal of reducing the overall dynamic range so you won't need to reach for the volume knob between each song. However, that won't lead to a recording which sounds as good as doing it the otherway. The limiter on a recorder is not intended to compress the overall dynamics of the recording, it is only meant as a safely device to prevent occasional "overs", softening the clip of the peaks instead of hard-clipping them.
A common recorder feature which
is intended to reduce the entire resulting dynamic range while recording is
Auto-record-gain, but that doesn't work very well for music and I don't think anyone here would recommend using it for the kind of recording we are doing. It's mostly intended for recording conversation and lectures. It's adjusting level "all the time" rather than just during peaks above a threshold level, so it's going to screw up the music parts along with the applause parts. If set appropriately, at least a limiter is only going to screw with the applause parts, even if it isn't really the most appropriate tool for the job.