Spend A LOT of time experimenting, and you will find what sound best. Use your own ears and your own judgements. That’s part of being a music engineer. We don’t learn and follow a set of perfectly crafted rules that lead to success… We use our own minds and opinions, and we decide what’s best in each situation. Every situation is unique because every song/mix is unique. Trusting your own opinion and judgement (while working on trustworthy speakers in a trustworthy tuned room) is how you learn and get good at what you’re doing.
For what it’s worth, because of your questions, it sounds like you’re just beginning and still learning so I’ll try to help a little more. Limiters can be helpful to control peaks and bursts of gain - so if you hear a bit too much of that, a limiter can be helpful in smoothing some of that out. They can have a slight “blanketing” effect, but sometimes this is exactly what you need. Set the limiter in the way it sounds best to you. If you don’t like the sound of the limiter, don’t use it. Clipping is cleaner than limiters to a point, but clipping has limitations too. Which is the best choice depends on your source material, your target level, your sound preferences, etc.
Converting a file to 16-bit is usually done last using either the Wavelab dither plugin, or maybe dither in another plugin (some limiters have dither). 44.1/16-bit are the specs used on CD, so if you’re preparing a master for CD, this should be your target. If you’re mastering for online release (not CD), it’s a matter of choice… Many places accept 24-bit files and Apple even recommends this: https://www.apple.com/au/itunes/mastered-for-itunes/docs/mastered_for_itunes.pdf. As far as sample rates, some people feel it’s best to keep sample rate as high as possible to the end, and then convert… but it’s kind of a matter of taste. For example, even if your source computer plays back at 96khz, if you’re going through D/A into A/D, you could capture music with your 2nd computer at 44khz/24-bit. Then process in Wavelab as you like, and then dither to whatever sample rate you need. You could also capture in Wavelab at 96khz/24-bit on your 2nd computer, and then use Wavelab to sample rate convert to 44khz (it may be best to do this right before the final limiter, but try and decide for yourself), and then dither to needed bit depth. There are many possible approaches… Another: If your source computer is connected digitally (not through an A/D–D/A), then obviously your digital clock on the 2nd computer must match the first computer, and you cannot capture at a different sample rate. In that case, resampling in Wavelab would be the best choice (if your target is less than 96khz). Try all of these options, make a critical decision using your own ears, and decide what’s best.