I don't know why people are enthused about "passive heat removal" for a reactor that's operating and generating electricity (as opposed to post-shutdown decay heat removal). That requirement rules out every possible type of fission reactor there is, other than extremely small ones with very low power densities (like OP).
Forced convection is how you fit gigawatts of power in the volume of a small room; it's what makes nuclear fission practical at scale.
Gigawatt reactors do not fit in a small room. Even the reactor itself is the size of a large room. Then there is a huge amount of infrastructure around them, then enormous cooling towers. Forget about power density when you have sites measured in tens or hundreds of acres with 300 foot towers--ultimately heat has to be dissipated to the environment, and that takes space.
The entire heat source, the active part of the core, very easily fits in a small room (at a mean power density somewhere around 100 MW/m^3 for water-cooled reactors).
I'm talking about the difficultly of heat removal, not trying to minimize size or complexity or anything.
Sure, but you gotta consider the entire system, so adding the heat pumping system actually may decrease the overall power generated per unit of volume; at least, certainly increase cost and decrease reliability. I wonder if thousands of SMRs spread over an area wouldn't actually generate more power per acre.
Forced convection is how you fit gigawatts of power in the volume of a small room; it's what makes nuclear fission practical at scale.