CHAIRO: Contextual Hierarchical Analogical Induction and Reasoning Optimization for LLMs

TL;DR

This paper introduces CHAIRO, a framework that uses analogical examples to improve content moderation with LLMs, enhancing accuracy, interpretability, and adaptability to diverse and unseen content scenarios.

Contribution

The paper presents a novel analogical reasoning-based moderation framework that optimizes rule induction and decision-making for better generalization and interpretability.

Findings

Outperforms rule-injected fine-tuning and static RAG pipelines in moderation accuracy.

Produces more interpretable and applicable moderation rules.

Demonstrates robustness through human and external model evaluations.

Abstract

Content moderation in online platforms faces persistent challenges due to the evolving complexity of user-generated content and the limitations of traditional rule-based and machine learning approaches. While recent advances in large language models (LLMs) have enabled more sophisticated moderation via direct prompting or fine-tuning, these approaches often exhibit limited generalization, interpretability, and adaptability to unseen or ambiguous cases. In this work, we propose a novel moderation framework that leverages analogical examples to enhance rule induction and decision reliability. Our approach integrates end-to-end optimization of analogical retrieval, rule generation, and moderation classification, enabling the dynamic adaptation of moderation rules to diverse content scenarios. Through comprehensive experiments, we demonstrate that our method significantly outperforms both rule-injected fine-tuning baselines and multi-stage static RAG pipelines in terms of moderation accuracy and rule quality. Further evaluations, including human assessments and external model generalization tests, confirm that our framework produces rules with better clarity, interpretability, and applicability. These findings show that analogical example-driven methods can advance robust, explainable, and generalizable content moderation in real-world applications.