PolicyLLM: Towards Excellent Comprehension of Public Policy for Large Language Models
This work provides a benchmark and model for improving LLMs' understanding of public policy, addressing an underexplored area critical for real-world decision-making.
The paper introduces PolicyBench, the first large-scale cross-system benchmark for evaluating LLMs' policy comprehension, and proposes PolicyMoE, a domain-specialized Mixture-of-Experts model. PolicyMoE outperforms baselines on application-oriented policy tasks, achieving the highest accuracy on structured reasoning tasks.
Large Language Models (LLMs) are increasingly integrated into real-world decision-making, including in the domain of public policy. Yet, their ability to comprehend and reason about policy-related content remains underexplored. To fill this gap, we present \textbf{\textit{PolicyBench}}, the first large-scale cross-system benchmark (US-China) evaluating policy comprehension, comprising 21K cases across a broad spectrum of policy areas, capturing the diversity and complexity of real-world governance. Following Bloom's taxonomy, the benchmark assesses three core capabilities: (1) \textbf{Memorization}: factual recall of policy knowledge, (2) \textbf{Understanding}: conceptual and contextual reasoning, and (3) \textbf{Application}: problem-solving in real-life policy scenarios. Building on this benchmark, we further propose \textbf{\textit{PolicyMoE}}, a domain-specialized Mixture-of-Experts (MoE) model with expert modules aligned to each cognitive level. The proposed models demonstrate stronger performance on application-oriented policy tasks than on memorization or conceptual understanding, and yields the highest accuracy on structured reasoning tasks. Our results reveal key limitations of current LLMs in policy understanding and suggest paths toward more reliable, policy-focused models.