4 papers
Model vs. User: Whose Values Prevail? Inference-Time Value Alignment in Multi-Turn Conversations
Enrico Liscio, Anne Arzberger, Aditya Shankar +1
Human values shape how we expect language models to behave. Dominant alignment approaches treat values as global objectives, although the appropriate value orientation is personal…
Beyond a Global Norm: Personalizing Toxicity Sensitivity in Language Models Without Retraining
Rares A. C. Diaconescu, Iulia Slanina, Alina Florea +5
Reducing toxicity is often framed as a global alignment problem, yet perceptions of harmful language are subjective and context-dependent. We present the first comparative evaluati…
"Label from Somewhere": Reflexive Annotating for Situated AI Alignment
Anne Arzberger, Celine Offerman, Ujwal Gadiraju +2
AI alignment relies on annotator judgments, yet annotation pipelines often treat annotators as interchangeable, obscuring how their social position shapes annotation. We introduce…
Co-Constructing Alignment: A Participatory Approach to Situate AI Values
Anne Arzberger, Enrico Liscio, Maria Luce Lupetti +2
As AI systems become embedded in everyday practice, value misalignment has emerged as a pressing concern. Yet, dominant alignment approaches remain model centric, treating users as…