Loss Function for logistic regression confusion

Remember that logs of numbers between 0 and 1 are all negative, right? So making a negative number larger (further to the right on the number line), makes it closer to 0, but from “underneath”.

Here’s a thread with a graph and some more discussion of the points here.