So from what I had initially understood, in a conversion problem, you go by whatever number has the smallest SF and round your answer to that number, even if that number is one.
In the examples on my homework it seems like my teacher chooses a random number of significant figures for each...