In some situations you can keep people from fighting the hypothetical by asking a question which explicitly states the point of the hypothetical, instead of asking something vague.

E.g., for Newcomb's paradox, instead of asking "What do you choose?" (potential answer: "I don't really need a million dollars, so I'll just take the box with the $1000") ask "which choice of box(es) maximizes expected monetary gain?"

E.g., for the Monty Hall problem, instead of asking "Would you switch doors?" (potential answers: "I'd probably just stick with my gut" or "I like goats!") ask "does switching or not switching maximize your chances of getting the car?"

E.g., for the Prisoner's Dilemma, instead of "What do you do?" (potential answer: "try to escape!") ask "Which of the offered choices minimizes expected jail time?"

Additionally, since they are framed in a less personal way, these kinds of questions may be less likely to be perceived as traps or set-ups.

Unfortunately it's much trickier to apply this strategy to hypotheticals about moral intuition, because usually the purpose of these is to see what considerations the other person in particular attaches to a question like "what do you do?" (E.g., you can't just ask "which choice maximizes utility?" instead of "what do you do?" in a trolly problem, without circumventing the whole original point of the question). You may still be able to rein in responses by specifically asking something like "would you flip or not flip the lever?", though.