If the RPV steam from the pressure release valves is always released into the wetwell, the air space there should be much more contaminated than the nitrogen in the drywell. The drywell would get more severely contaminated later on, once the RPV melts through or the seals around it start leaking, but initially all the good stuff ends up in the torus.
There may be some exchange of pressure via the downcomer tubes between the two if there is a large pressure differential (will water get pushed up the tubes if pressure gets too high in the torus, opening a path for contaminated gas to leak from the torus into the drywell?), but my understanding is that until the RPV gets damaged by excessive temperatures the wetwell would be the more contaminated of the two spaces.
It says they prohibited entry to R/B 1 at 21:51 on 2011-03-11 because of radiation (about seven hours after the quake). Tepco estimates that fuel was exposed five hours after the quake because the water level had fallen too far. So by the time they prohibited entrance, the meltdown had been in progress for about two hours. I don't know what the containment pressure was like by then, but 4 hours later, at 02:00 JST it was 0.6 MPa and at 05:30 it was at 0.82 MPa. The containment may already have been near or at design pressure (0.42 MPa = 4.2 bar) when the radiation went up.
For radiation levels inside the building to go up without venting there must have been some leaks. But in a way it is not surprising that the unit 1 containment was leaking in 2011 (when it was 40 years old and stressed to the max by a melting core) when it had already leaked unacceptably in 1992 during routine inspections when tested at 3 bar:
2.5% of several thousand cubic meters of nitrogen at 3 bar is several hundred cubic meters that would have leaked per day.
Probably because most of the pipes coming out or going into the containment go next door to the turbine building. It's like the belly button of the reactor. The turbine hall is also more spacious.