The mount on the top of the 747 (the Shuttle Carrier Aircraft) for the drop tests was very similar to the one used throughout the program for ferry flight. The major difference was that the legs on the forward support strut for the drop tests were 13 feet long instead of the normal 8.5 feet long used for ferry flight, and included a fairing/drag strut on the aft of each leg. The extra length was to pitch the Orbiter up relative to the SCA to aid in vehicle separation.
You can compare the drop test and operational forward attach in these photos.
Both mounts interfaced with the Orbiter in the same fittings that it used to be mounted on the External Tank for launch.
The release mechanisms were indeed pyrotechnic. From the Approach and Landing Test Final Report:
A.1.5.1 Mechanical Separation
The separation system provided the capability to release the orbiter from the carrier aircraft . This was accomplished by pyrotechnic frangible bolts at three structural attachments, one forward and
two aft. Load sensors at each of the structural attachment interfaces provided measurement of the loads between the orbiter and carrier. Separation of electrical umbilicals was accomplished by pull-apart connectors subsequent to structural attachment separation using relative separation motion.
One problem was discussed
Pyrotechnic functions that were required to operate did so
normally. These consisted of (1) actuation of the strut
that assisted the hydraulics in deploying the nose
landing gear and ( 2 ) carrier aircraft / orbiter
separation. Shock from the separation system explosive bolts
caused the electrical connector to be damaged during
separation. On orbital flights, the orbiter / external tank separation system electrical connector will be replaced after each flight . A potential problem identified prior to flight 1 was that, with a
single-point failure, a pyrotechnic initiator controller at the forward or aft attach points could fire when armed without the "fire 1" or "fire 2" command
being present. To preclude this possibility,
modifications were made on the pyrotechnic initiator controller separation circuits so that, even with a failure, a controller would not operate prematurely.
Additional reference: Jenkins, Space Shuttle, 1992 edition, page 134