I’d imagine that upscaling refers to increasing bit-depth (like from 16 to 24 bits) which adds "vertical" resolution to the waveform... in theory, improving amplitude resolution... making subtle low-level detail more natural and audible (among other things) if done correctly.
Whereas upsampling likely refers to increasing sampling rate (like from 44.1 to 88.2 kHz) which also adds resolution but in the time-domain (or horizontal resolution of the waveform). I’m not sure if there are algorithms that can add high-frequency information that wasn’t captured in the original sampled result... but who knows... maybe so as AI and pattern-recognition software improves.
Also, in conjunction with upscaling, having more sample points could also serve to increase resolution because a bit-word needs to be calculated and saved for each new sample... and the upsampler/scaler might do a more accurate job of this than the typical "oversampling" filter in most DACs. Some DACs may also sound better in terms of being less affected by jitter if they are converting higher sampling rates. I’m just throwing out various ways that adding new samples could improve, or at least change, the perceived sound quality we hear.