How can the results be reproduced, for example, if I ask a question about a particular scene?
How can the results be reproduced, for example, if I ask a question about a particular scene?