Toward Culturally Aligned LLMs through Ontology-Guided Multi-Agent Reasoning
Wonduk Seo, Wonseok Choi, Junseo Koh, Juhyeon Lee, Hyunjin An, Minhyeong Yu, Jian Park, Qingshan Zhou, Seunghyun Lee, Yi Bu
Why It Matters
What makes this one worth your time
This research addresses the critical issue of cultural misalignment in LLMs, which is essential for developing AI systems that are sensitive to diverse cultural values and improve decision-making processes.
OG-MAR enhances cultural alignment in LLMs through structured ontology-guided reasoning.
Summary
The paper proposes OG-MAR, an Ontology-Guided Multi-Agent Reasoning framework that enhances cultural alignment in LLMs by utilizing structured value representations from the World Values Survey and instantiating value-persona agents to improve decision-making consistency and interpretability.
Key contributions
- Introduction of the OG-MAR framework for ontology-guided reasoning in LLMs.
- Demonstration of improved cultural alignment and robustness through experiments on regional social-survey benchmarks.
- Development of a judgment agent that enforces ontology consistency and demographic proximity.
Notable insights
- The use of a global cultural ontology to guide multi-agent reasoning is a novel approach to enhancing interpretability in LLM outputs.
- The synthesis of outputs from multiple value-persona agents allows for a more nuanced representation of cultural values.
Possible limitations
- Not stated in the abstract.
Abstract
arXiv:2601.21700v3 Announce Type: replace-cross Abstract: Large Language Models (LLMs) increasingly support culturally sensitive decision making, yet often exhibit misalignment due to skewed pretraining data and the absence of structured value representations. Existing methods can steer outputs, but often lack demographic grounding and treat values as independent, unstructured signals, reducing consistency and interpretability. We propose OG-MAR, an Ontology-Guided Multi-Agent Reasoning framework. OG-MAR summarizes respondent-specific values from the World Values Survey (WVS) and constructs a global cultural ontology by eliciting relations over a fixed taxonomy via competency questions. At inference time, it retrieves ontology-consistent relations and demographically similar profiles to instantiate multiple value-persona agents, whose outputs are synthesized by a judgment agent that enforces ontology consistency and demographic proximity. Experiments on regional social-survey benchmarks across four LLM backbones show that OG-MAR improves cultural alignment and robustness over competitive baselines, while producing more transparent reasoning traces.