Could these be fake?
Is it a valid assumption that a MM2114N-2 would work in place of a D2114?
Yes, and yes.
The DSs I can find online tell me that the MM2114N-2 has a faster cycle time than the D2114 (200 ns vs. 400 ns), but this is usually not a bad thing - the other timings look compatible.
As these are small, DIP packaged, and static memories, it should be easy to just put them in a breadboard and manually test some address, or create a more sophisticated test rig with an MCU.
Nostalgic oldish man: I still remember when I expanded my self built 6502 board to a mind-blowing 4 kB, by stacking eight 2114s in two four chip high towers, all pins soldered in parallel but chip selects.