MSRs have negative temperature reactivity coefficient and outlet temps around 700C at atm pressure. PWR is at measly 300C and 150 Bar.
If all control is lost, the salt expands as it heats up pushing the expanded volume out from the reactor core. The fission stops once the fuel is leaves the core region where the moderator is. Reverse is also true: you pull heat off from the loop, so the fuel-salt becomes denser, increasing reactivity. MSRs can naturally “follow” the load, if done right.
I’ll just comment about one thing that keeps popping up in the discussions: grid-level storage. There is no such thing yet really that would last a full day cycle, and the 100MW or so units we are building are mostly for frequency stabilization and for buying enough time to turn on a base-load plant when the renewables drop out. I’m not arguing against storage - it is absolutely needed.
The problem is the scale, which people don’t seem to get. Largest amount of energy we can currently repeatedly store and release is with pumped hydro, and the locations where this is possible are few and far between. Once the batteries reach this level-of-capacity, then we have a possibility to use them as grid-level storage that lasts a few days instead of hours.