In the paper they use a ratio defined by comparing a hashtag's popularity on TikTok vs on Instagram. So it accounts for Instagram's biases since they're comparing to them. Not sure if that's what you're asking.
They’re using the difference between TikTok and Instagram to conclude (by implication at least) that TikTok has a bias. But the same data could just as easily imply Instagram has a bias, or both do. Or something else. Or nothing.