isn't mass supposed to tie to which ever flavours
Intuitively, yes, it's "supposed to"—but it doesn't. It's strange, but that's how it seems to work.
Here's a rough overview: Neutrinos are created in weak interactions. In these interactions, a charged lepton (like an electron) can emit a W boson and "turn into" a neutrino; likewise, a W boson can produce a neutrino/antilepton (or lepton/antineutrino) pair. These interactions conserve flavor: an electron will always appear with an electron-(anti)neutrino, a muon with a muon-(anti)neutrino, etc. These states, of a single, well-defined flavor, are called "flavor eigenstates".
Likewise, a neutrino with a single, well-defined mass is in a "mass eigenstate". If you've only studied non-relativistic QM, this might seem strange—isn't mass just a constant of the theory? Doesn't "mass eigenstate" imply a "mass operator"? What would that operator look like? To make sense of it, you need to move to a relativistic mindset. In relativity, a particle's energy and momentum are closely related; they form a four-vector, and this vector's length is a scalar quantity: $mc^2$. That suggests you can define a sort of "mass operator" using energy and momentum operators. Similar to how a particle's flavor state determines how it participates in weak interactions, a particle's mass state determines how its free states propagate through space.
It turns out, though, that for neutrinos, the flavor eigenstates are not, in fact, mass eigenstates. A flavor eigenstate is a superposition of mass eigenstates, and vice-versa. This means that the flavors of neutrinos don't actually have individual, well-defined masses. There simply are "three neutrino flavors" and "three neutrino masses", but we can't assign a one-to-one correspondence between them.
The reason this causes neutrino oscillations is because, as I mentioned, a particle's mass state determines how it propagates through space. A neutrino created by an electron interaction is necessarily an electron neutrino, so it must be in a superposition of three mass states. These three mass states then propagate (as plane waves) at different speeds, and as they do, their respective waves get out-of-phase and interfere with each other. The result ends up being that, if you measure the neutrino at some later time, it'll still be in a mixture of mass states, but won't necessarily be the same mixture. The new mixture, generally, won't correspond to the same flavor eigenstate the neutrino was created in, but will be a superposition of flavor states. When you measure it, then, there's some chance you'll get a different flavor than you started with. This is the source of "neutrino oscillations".