Hi,
There are many ways to do what you want, I think you originally didn't want it to shut down but you wanted the "short circuit protection" to divert excess current away from the sensor line - is that correct? If so, you could mess around with either of these circuits which are/were in the ST LM78xx datasheet in the applications section:
An overcurrent shutdown circuit would be quite different, and off the top of my head I'd consider a PMOS pass device after the regulator rather than a relay; how you control the MOSFET gate is another matter I haven't thought through - maybe a comparator (with a reference voltage) driven by an op amp used to sense the current passing to the sensor, I think that is - downside - more parts but - upside - less messy than relying on random BJT Vbe turn-on voltages and less temperature-dependent. PICs have comparators in them, maybe, to save on one part?
Look at - I think they're called - powerpath ICs and things like that if you want a small, ready-made and time-saving solution.
- - - Updated - - -
...Figure 13 divides the current proportionally between regulator and transistor, e.g. 10% through former and 90% through latter - maybe that could be used to shunt excess current to ground via a resistor (rather than pass it to the load), not sure if it would still work, 'though, you'd need to test that theory.
Figure 14 does the same as 13 and it turns off Q1 when current across Rsc creates sufficient voltage drop to turn on Q2. Afraid I can't remember what it does with the regulator. Hopefully another member could explain that part.
- - - Updated - - -
14 looks like it does hiccup mode functionality until the problem/short circuit current either goes away or power is removed, maybe not what you're looking for... The more the rusty mental cogs whirr, the more I think that that is a current limiting circuit.