Add a message to either the Logging Text on the General tab or to the Final Message Text on the Terminator tab and see if it ends up in the Sequence's log - that would be your biggest clue the stage is actually getting triggered.
Then on to the STOP requests, assuming the stage gets triggered - how long post abort do the other two jobs run before they finish? If they finish up rather quickly, the STOP request may not have time to take effect. It can also depend on what exactly the other jobs are doing, some activities are harder to stop than others.
If the TA stage isn't actually getting called, you'll need to check your trigger values and also any 'automatically handle' type options you've enabled in the Sequence job.
Rather than having three in the job (which is perfectly legitimate) you could consider another approach to handle this 'globally'. Add an Exception Activity stage to the job and link the Terminator Activity to that. This would give you one place where any aborted job could branch to and shut down everything else without having to have a spider web of links from all the Job Activity stages. Just something to ponder.
