CuMA: Aligning LLMs with Sparse Cultural Values via Demographic-Aware Mixture of Adapters

arXiv:2601.04885v2 Announce Type: replace-cross
Abstract: As Large Language Models (LLMs) serve a global audience, alignment must transition from enforcing universal consensus to respecting cultural pluralism. We demonstrate that dense models, when forced to fit conflicting value distributions, suffer from textbfMean Collapse, converging to a generic average that fails to represent diverse groups. We attribute this to textbfCultural Sparsity, where gradient interference prevents dense parameters from spanning distinct cultural modes. To resolve this, we propose textbftextscCuMA (textbfCultural textbfMixture of textbfAdapters), a framework that frames alignment as a textbfconditional capacity separation problem. By incorporating demographic-aware routing, textscCuMA internalizes a textitLatent Cultural Topology to explicitly disentangle conflicting gradients into specialized expert subspaces. Extensive evaluations on WorldValuesBench, Community Alignment, and PRISM demonstrate that textscCuMA achieves state-of-the-art performance, significantly outperforming both dense baselines and semantic-only MoEs. Crucially, our analysis confirms that textscCuMA effectively mitigates mean collapse, preserving cultural diversity. Our code is available at https://github.com/Throll/CuMA.