If including multiple, numbered citations together, is it standard practice^ to order them correctly? For example, one might see

Several studies [1,3,5] show that quantity X is relevant to Y.

Or

Several studies [1][3][5] show that quantity X is relevant to Y.

It is also common to see

Several studies [5,1,3] show that quantity X is relevant to Y.

The last is easy to achieve if using a tool such as bibtex, with a bibliography ordered in some way other than order of mention in the text.

I personally find the lack of order jarring, but I am unsure if this is simply my personal taste.

^as far as this can be determined across fields. I'm personally interested in mathematics, computer science, computational science, physics, and related fields.

If including multiple, numbered citations together, is it standard practice^ to order them according to number? For example, one might see

Several studies [1,3,5] show that quantity X is relevant to Y.

Or

Several studies [1][3][5] show that quantity X is relevant to Y.

It is also common to see

Several studies [5,1,3] show that quantity X is relevant to Y.

The last is easy to achieve if using a tool such as bibtex, with a bibliography ordered in some way other than order of mention in the text.

Is there any authoritative reference that such ordering is preferable in general? To clarify, since this question was marked off-topic, the issue is whether, in the absence of guidelines from a journal, etc. there is any sufficiently-authorative concept of general style here. If the answer is "no" (which I believe it is), that is acceptable. If the answer is "yes," it can be defended with references to general academic style guides.

^as far as this can be determined across fields. I'm personally interested in mathematics, computer science, computational science, physics, and related fields.