How would they know for sure that some details were not part of some other training data they use? The authors may have discussed some tangential details on a f

How would they know for sure that some details were not part of some other training data they use? The authors may have discussed some tangential details on a forum for example, in which case you might argue that the model picked up on these details the authors assumed were benign but novel and worked out how to apply them to the problem.