The number is the total atomic mass. Helium 2 is 2 protons. Helium 3 is 2 protons and a neutron.
Regular helium is Helium 4. By far the most abundant, over helium 3 (one proton).
That’s Helium 3, it’s rare but it exists.
A diproton is Helium 2. It’s extremely unstable and doesn’t hang around. (We assume it exists during fusion).
Relative to the amount of boring old He-4, rare.
He - 3 may be more common on the surface of the moon, due to deposition from the solar wind. People have speculated that He - 3 could be mined on the moon and used in fusion reactors. As of now it’s a rather far fetched idea, but it’s not totally absurd.
To add to @Chronos’s point: How strongly attracted single neutrons and/or protons are to one another depends on the quantum mechanical spin of the two particles. All three pairings (nn, pp, np) are unstable in the case that the nucleon spins are anti-parallel. So for the deuteron (np), the neutron and proton always have parallel spins.
So, why doesn’t this lead to comparable nn and pp bound states? The problem is that these involve identical particles and thus they run afoul of the Pauli exclusion principle. The nuclear force requires the spins to be parallel to have a potentially bound state, but the exclusion principle doesn’t allow identical particles to occupy the same state, which means that spatial overlap (which is part of the quantum state) is disfavored in the parallel-spin nn and pp cases.
So, the np case is the uniquely bound case because the neutron and proton can have the same spin without penalty.