You can’t extrapolate like this. You have to account for body surface area, and even then it’s a very poor approximation (given non-linear pharmacodynamics and inter species variability, etc). You also need an uncertainty factor from the NOAEL, at least a log fold. A better extrapolation would be:

For mouse: 0.05 x (3/37) x 75 = 0.304 mg

Summary on extrapolation in full text of this paper (though note that I don’t encourage extrapolation from animals to humans in most cases — it’s very imperfect):

Dose translation from animal to human studies revisited. - PubMed - NCBI