Imagine a particle constrained to move along the surface of an infinitely long cylinder of radius $R$. When we zoom in close to the cylinder, this motion is clearly 2-dimensional, since the particle can move both along the axis of the cylinder and around the cylinder's circumference.

However, if we zoom out to distances far, far larger than $R$, then the motion starts to look 1-dimensional. When the circumference is extremely small, then motion around the circumference is very difficult to resolve - so for example, a particle which doesn't move along the axis but simply winds around the cirumference in circular loops will look as though it's basically sitting still.

A point on the surface of the cylinder can be labeled by two numbers - a real number $z$ which specifies a point along the axis of the cylinder, and an angle $\theta$ which specifies position around the circumference. Near the cylinder, the motion is clearly 2-dimensional, but at large distances, motion in the $\theta$ direction is impossible to see, and so for such distant observers, the only dynamical degree of freedom for the particle would be $z$ and the motion would be effectively 1-dimensional.
Our limited primate brains are not capable of directly visualizing it, but from a mathematical perspective it is trivially easy to extend this idea to higher dimensions. We might imagine, for example, that the position of a particle is labeled by four numbers - $x,y,z$, and $\theta$. The first three coordinates can take any real number as a value, but $\theta$ is once again restrict to an angle between $0$ and $360^\circ$. Loosely, you can imagine a tiny loop of radius $R$ at each point $(x,y,z)$. Just as before, at distances vastly larger than $R$, motion along the loop is impossible to resolve, and for distant observers it looks like the position of a particle is specified completely by $x,y,$ and $z$ alone. The motion along the loop only becomes apparent when we use probes which are sensitive to distances on the order of $R$.
Through such compact extra dimensions, it is possible to imagine a universe with (possibly many) more than the three "large" dimensions that we are intuitively familiar with, such that the compact extra dimensions are simply unobservable on the length scales we've been able to experimentally probe so far. If we postulate that ordinary matter simply cannot propagate in these additional dimensions, then it's possible that they aren't even particularly small. However, because gravity is intimately related with the structure of space and spacetime, it's possible that these otherwise-unobservable extra dimensions could have an important impact on gravitational interactions. In particular, they potentially go a long way toward explaining why gravity is (apparently) so phenomenally weak compared to the other elementary forces.