Yeah, so GPS was designed to not apply leap seconds from the start, which is important because there's a field built into the protocol to send the current offset. The reason I describe it as "no longer applying" is because the GPS time is monotonic from UTC at an arbitrary starting point, so it doesn't match up with any of the strictly monotonic time bases either - that is, GPS time is monotonic but does not match the agreed upon international monotonic time standard, TAI, because GPS time has some of the UTC leap seconds applied.

TAI is also, of course, fundamentally just a divergence from UTC caused by not applying leap seconds. However, formally, UTC is defined as TAI plus or minus the leap second offset in order to minimize divergence from UT1 (which is a smoothed astronomical time base rather than monotonic). GPS time is TAI plus or minus a different offset caused by some of the leap seconds being considered.

There is a basis for this, LORAN did the exact same thing but with an earlier divergence point.