trends.barnas.net

task · first seen 11 Oct 2026

pluralistic alignment

Aligning language models to represent diverse, competing societal values rather than a single dominant viewpoint.

Mentions per week

0419 May11 Aug10 NovPapersCodeDiscussion

PapersCodeDiscussion

Earliest sources

Latest sources

Mentioned together with

LLM-based annotationpreference learninghypernetworkactivation steeringreward learning from preferences