Yes, and the one of women drops below the one of men, and that's clearly what the initial reference to the paper was talking about.

But only in the case of 'outsiders'. In the case of 'insiders' the acceptance rate of men drops more than that of women when their gender is known (no p-value provided so hard to tell if statistically significant).

A person trying to use github as a hiring signal is going to be an outsider at some point.