Lagrange points are stationary points (edit: and not extremae, as I originally wrote - thanks to pbuk for pointing that out) of the effective potential of a two-body system, whereas the diagrams "explaining" relativistic gravity as a rubber sheet are embeddings of the spatial plane of a one-body system. They don't fit together, really, because they're different aspects of different scenarios, despite superficial similarities.
I would suspect that Lagrange points are rather more complicated things in relativistic gravity. Two orbiting bodies emit gravitational radiation and their orbits decay, so I expect that the Lagrange points are only short-term stable (for some extremely long values of short term, with typical astrophysical bodies).