Today
-
Present starter example scores as prose, matching Mediation card
-
Streamline dataset card and starter documentation to match Mediation layout
-
Clarify starter split purposes and paired ATE/CATE workflow
-
Add independent 96-study public starter and CausalPFN paper reference
-
Match test-only publication scope: observed inputs, opaque IDs, folds and blank prediction templates
-
Release public paired ATE/CATE test collection, generator, labels and scoring tools
-
initial commit