Helium-3 coming from active seafloor vents has something like 10x (to in some locations 30x) the ratio of 3He/4He in the atmosphere. This is so consistent that it is often used as a way to detect underwater vulcanism. Yeah, it's still an order of magnitude less than in the lunar regolith, but we have lots of experience handling seawater at scale, and we have zero experience handling rocks in 1/6th G at scale, so I would probably bet on it taking a long time for the lunar mining to get to the same effectiveness as the seawater handling.

Deuterium and Tritium probably.

There's very little tritium in seawater either (and it was mostly put there by hydrogen bomb tests.)