P0327 code set - KS's replaced - Still throwing code

98 XJ with an 05 LM7 5.3L from a Sierra 1500, 35x12.5" shoes, 4.56 gears, 4l65e1. Swap went smooth (finished around 8-10 months ago).

I recently noticed fuel economy slowly tanking. Was getting 20mpg when I first completed the swap, but sometime in the past few months it's slowly dropped. Currently getting around 16.5mpg highway.
For the past few weeks, I've been throwing a P0327 intermittently. It'll clear on it's own but return a few days later. If I clear it manually, it usually comes back in about the same amount of time.

Tested the KS's by tapping the heads and the front KS didn't really seem to respond reliably, so I swapped out both sensors and harness. Old KS's were rusty as hell, and there was heaps of mud in there, so I gave it all a good cleaning, and built up some RTV dams for both sensors before replacing. Got it all back together, and still throwing the P0327.

Today I had time to do some testing. At cold idle, both KS's are making around 150-100mV. Both respond roughly the same to tapping on the heads. Verified good continuity to the ECM connector pins 11 and 51.

So, I'm already aware of what SHOULD set a P0327 under normal conditions, but starting to think something abnormal is going on. All my grounds seem good, not seeing short to voltage, just can't find anything wrong, other than smells rich, tanking fuel economy, and the P0327.

Any ideas?

I've seen excessive false knock cause knock sensor codes.

Occasionally, I'll notice certain calibrations are more sensitive to false knock than most.
For example the 99's had a lot of false knock.

If you can reflash the Pcm with another or newer OS and see if this fixes it.
The other option is to desensitize the knock sensors in the calibration along with all the limit timing tables.

Yup, we're on the same page.
I think I figured it out. My original theory was that the exhaust leak was causing excessive noise (not enough for the ECM to think it's a knock, but enough to raise the average KS voltage). The high average KS voltage then drops below the threshold required to set a P0327 after driving for a bit and coming back down to idle (exhaust leak makes far less noise at idle than at higher RPM).
I'd already listened to everyone who said there was no way the exhaust leak was the problem, and replaced both KS's and harness to no avail. After sealing up the exhaust though, I threw the code one more time on the first drive (expected, as the noise and thus average voltage would be low until the ECM figured out the new average). But it's been a week since fixing the leak and no more code.