The first trials with NaCl didn't destroy the glass tanks and remember that NaCl melts at 801 °C, a temperature far superior to the maximum temperature of current MSR concepts. He had to go at a higher temperature to get a steam explosion, maybe superior to 883 °C the temperature where sodium boils.
Those points are all true. However it is interesting that a physical effect induced purely by temperature created an explosion much larger than the impact from the chemical effect of Sodium hitting water. It really does show the need to test such things.
For the sodium case the explosion occurs outside of the tank, at the surface of the water so there is no mechanical stress to the tank. The sodium didn't enter into the water because of the chemical reaction. In a sodium-water steam generator, if there is a leak, the pressure of the water leads to a better mixing with the sodium which can lead the demise of the steam generator (although there is no explosion here since there is no oxygen to react with the hydrogen normally).