As You correctly guessed the "orbit plane" is actually a surface (no thickness), so the individual atoms of the space station would, if separate, be on different orbits and drift away over (short) time; of course they are kept together by direct interaction so structural strength of your space station won't have any problem in keeping it together.
This doesn't hold true for any free-floating object which, unless prevented, will try to follow his own orbit. This means objects "above" the "mean orbital surface" will drift to the ceiling, while those "below" (in Earth direction) will drift toward the floor.
Let us try to compute the actual forces.
On orbital plane (assumed being the Station Center of gravity) the two forces, gravity ($F_g = \frac{G M m}{r^2}$) and centrifugal ($F_c = m v^2/r$) are exactly the same "by definition" . This enables us to easily find relationship between speed and radius of a circular orbit.
What we need to do is a bit different, we need to find what is the difference of force if we change "a bit" the radius leaving the speed the same.
So we have:
$\Delta F = F_g^\prime - F_c^\prime = \frac{GMm}{(r+\delta)^2} - mv^2/(r+\delta) = ma$
$a = \frac{GM}{(r+\delta)^2} - v^2/(r+\delta)$
If we substitute the relevant data:
$G = 6.67 \times 10^{-11}$
$M = 5.972 \times 10^{24} kg$
$r = 35786 Km = 35786000 m$
$v = 3.07 Km/s = 3070 m/s$
and assume for $\delta$ 50 floors about $3m$ each: $\delta = 50 \times 3m = 150m$ we get:
$a = \frac{6.67 \times 10^{-11} \times 5.972 \times 10^{24}}{(35786000+150)^2} - 3070^2/(35786000+150) \approx 8 \times 10^{-7} m/s^2 $
Note: I had to "cheat" to get the result; the values I have for all constants are way too imprecise to correctly compute the result (we are speaking about $4 ppm = 4 \times 10{-6}$). What I did is to "massage" radius till the acceleration ($\delta = 0$) was null, then I added the displacement. Someone is welcome to cross-check my computations as I might have goofed somewhere.
From basic Physics we have:
$s = 1/2 at^2$ ... $t = \sqrt{2 s / a} = \sqrt{\frac{2 \times150}{8 \times 10^{-7}}} \approx 8470s \approx 2 h 20m$
As you see the "free floating" objects will drift to floor/ceiling in a matter of hours, days at most for objects nearer to orbital plane.
Note that the forces involved are quite tiny, so even the smallest friction will prevent it, so, in particular, you can expect surface tension to keep liquids plastered to whatever surface they touch and not to drift around.