There are two handles on this: reconstructing the energy/direction and making sure it's consistent with the beam and gating on the beam duty cycle (since at least the beams I'm familiar with produce neutrinos in short bunches, but I'm no expert).

You can also add modulation to the beam to encode more information to help discriminate against backgrounds but that might require even more rate...