I get the impression that people tend to use the "sweet spot" terminology a bit loosely when it comes to MF and LF listening.You are right to doubt it arising from phase cancellation. The numbers don't make sense in terms of wavelength. Instead, a "sweet spot" at LF is a broader concept...pretty much just anywhere the noise happens to be at its least practical value simultaneously from all the multiple individual noise sources in the vicinity. (That is, how far can I get away from the TV set and furnace igniter noise in my own house before the neighbor's TV and RF ballasted lights start to become the dominant noise source?)
The fact that this can work at all is based on the very phenomenon you observed with the power line hum. You know that a propagating radio wave in free space decreases in intensity in inverse proportion to distance from the source (1/r). But strength within the near field region of a given source falls off inversely with the square of the distance (1/r2). (And very close to the source, the relationship is even steeper, 1/r3 or greater under some circumstances.) So basically, anywhere within the near field, distance is your friend. Doubling it means decreasing the field by a factor of four, be the offending field electric or magnetic in nature.
And you are also correct in thinking the near field region is wavelength dependent. It is usually defined, in fact, as λ/2π, or .159 wavelength. It's not a magic, precise dividing line between the induction zone and the radiative zone, and the free-space attenuation does not instantly transition from 1/r2 to 1/r right there either, but it's a useful rule of thumb.
I hope that helps.
John