Here's my take on all of this...
The end goal is to maintain mash temps (or make small adjustments). Say you mash-in at 148F but want a target temp of 152F. In your case, reading the wort as it exits the mash, your PID will turn on and continue to heat the element until it starts to near 152F - at which point it will start to throttle back the heating cycle so you don't over shoot the temperature (assuming it has already been calibrated). But, during that time, the wort may reach temps much higher than 152F. In other words, the PID doesn't care that it's raising the temperature past your set point, it only cares that the wort exiting your mash tun gets to the set temp. So you could be heating your wort up to even 160F before everything equalizes in the mash-tun nearing your target temp.
Now take the other scenario, you are reading the temperature as it exits the RIMS chamber. In this case, the PID will never over-heat past 152 (or maybe a little more). As you continue to recirculate the wort, you eventually heat the entire mash BUT without ever exposing the wort to much higher temperatures.
Hope I'm making sense. Anyways, that is why I chose to read my temp. as it exits the HERMS chamber. But maybe both work just as well?