Can a model represent genuine preference pluralism at cluster level without collapsing to an average — and would that be legible enough to anyone to matter?jordanSASite Admin·...importedcommunication · 9.7UpTrust ranks posts by who trusts the writer, topic by topic. Sign up ›machine-learningai-safetyinterpretabilityUptrustDowntrust0Reactmore options