This offline facility lets people compare interpretations under different message-timing rules. Its cases are prepared examples. It does not run an LLM, test the frozen DUAL-P0 experiment, or establish cognitive effects. Keep the original DUAL-P0 contract, reserved seeds and data separate.
Prepare the comparison
Choose an example scenario and an exchange regime, then start. You can play both roles or let two people take turns. The interface controls when each role receives the other's words.
Export contains committed interpretations and messages, not unsaved drafts. JSON is inspectable and editable; import checks structure but does not authenticate authorship. Files are saved only when you request them.
Three distinct research possibilities
DUAL-X1 · Independent interpretations
Does the opportunity to exchange before or after commitment change what people retain, revise or overlook?
DUAL-X2 · The tiny N0 gate
Can a restricted coordinator protect independent information better than unconstrained sharing or a powerful referee?
DUAL-X3 · The unseen third
Can a third participant use a bounded product of A/B interaction on new tasks without seeing the hidden answers?
The negative result matters
Record when exchange makes no difference or distorts an interpretation. These exploratory notes do not change the separate DUAL-P0 protocol.