DeonticBench: A Benchmark for Reasoning over Rules

TL;DR

DEONTICBENCH is a comprehensive benchmark designed to evaluate large language models' ability to perform complex deontic reasoning over real-world legal and policy rules, incorporating both language-based and symbolic approaches.

Contribution

It introduces a large, diverse set of tasks for deontic reasoning, along with a framework for combining language models with symbolic computation, and provides baseline results and analysis.

Findings

Best models achieve only around 44-46% accuracy on benchmark tasks.

Training improves symbolic program generation but does not reliably solve tasks.

Benchmark covers real-world legal and policy domains with over 6,200 tasks.

Abstract

Reasoning with complex, context-specific rules remains challenging for large language models (LLMs). In legal and policy settings, this manifests as deontic reasoning: reasoning about obligations, permissions, and prohibitions under explicit rules. While many recent benchmarks emphasize short-context mathematical reasoning, fewer focus on long-context, high-stakes deontic reasoning. To address this gap, we introduce DEONTICBENCH, a benchmark of 6,232 tasks across U.S. federal taxes, airline baggage policies, U.S. immigration administration, and U.S. state housing law. These tasks can be approached in multiple ways, including direct reasoning in language or with the aid of symbolic computation. Besides free-form chain-of-thought reasoning, DEONTICBENCH enables an optional solver-based workflow in which models translate statutes and case facts into executable Prolog, leading to formal problem interpretations and an explicit program trace. We release reference Prolog programs for all instances. Across frontier LLMs and coding models, best hard-subset performance reaches only 44.4% on SARA Numeric and 46.6 macro-F1 on Housing. We further study training with supervised fine-tuning and reinforcement learning for symbolic program generation. Although training improves Prolog generation quality, current RL methods still fail to solve these tasks reliably. Overall, DEONTICBENCH provides a benchmark for studying context-grounded rule reasoning in real-world domains under both symbolic and non-symbolic settings.