Confirmatory trial evaluates AMÁLIA-9B's performance in annotating moral foundations in European Portuguese, suggesting effective application in text analysis.
Pre-registration of the confirmatory phase of a study evaluating AMÁLIA-9B (Portugal's open national LLM) as a theory-driven annotator of the moral foundation Authority/Subversion on a transcreated European-Portuguese corpus (MFRC-derived, n=748). Registers the primary endpoint (the recovery gap between holistic and clause-level stated-formula performance), directional hypotheses, numerical success criteria, confirmatory secondaries, and the strictly out-of-sample test set (448 previously unseen texts), before any model annotation of those texts. A completed 300-item pilot is disclosed in full; the frozen instruments (English and Portuguese constructs) and the study's append-only method ledger are attached.
No takes yet. Share an insight, caveat, or question.
Manuel Marques-Pita (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: