A demonstration that the instrument is reconfigurable. Same constant-sum allocation, same runner, same scoring as the four-axis pilot — but here the dimensions are five adjudicative logics (the currency used to resolve a conflict: Welfare, Duty, Honor, Sacred, Self-interest), a compositional set rather than bipolar axes. The five logics are identical in every scenario; only the dilemma and the country framing change.
Which logic resolves the conflict — and how framing moves it
Each bar is the mean allocation across the five logics, averaged over six dilemmas. Unframed, the models reason by impartial welfare and duty. Ask them to answer "as a typical person living in X" and the composition fans out — evidence the instrument picks up real signal on a dimension set it wasn't designed around.
Read the prompts and the actual reasoning
Pick a dilemma and a framing. You get the scenario, the options, and for each model its judgment allocation, its allocation across the five logics, and its own written reasoning (the fullest of the reruns).