The name "vector" is actually just a special case of a one-dimensional matrix.

So a processor may have a vector of addresses.
Or a vector of instructions (that may be abused to store part of the program if that interrupt source isn't needed)

It's just a question of view. Some people consider both the above as interrupt vector tables. Some people consider only the table of addresses to be interrupt vector tables.

But the naming is irrelevant. The important thing here is to notice the huge difference in behaviour between a processor that loads an address from a vector element, and a processor to jumps to the vector elements and expects to find instructions there.