Do LLMs Truly Understand When a Precedent Is Overruled?
Li Zhang, Jaromir Savelka, Kevin Ashley
TL;DR
This paper introduces a long-context legal benchmark to evaluate LLMs on overruling relationships in U.S. Supreme Court opinions, using 236 case pairs from 1789 to 2025. It deploys three tasks (open-ended identification, closed-ended verification, and reversed verification) to stress deep legal comprehension and temporal reasoning across lengthy texts. Empirical results show three major limitations: era sensitivity, reliance on shallow heuristics, and context-dependent temporal reasoning failures, with best models achieving only about 73% accuracy on the core task. The work highlights a critical gap between current LLM capabilities and the nuanced demands of legal practice, and it points to future directions in domain-specific pretraining, uncertainty quantification, retrieval-augmented approaches, and broader cross-jurisdiction benchmarks.
Abstract
Large language models (LLMs) with extended context windows show promise for complex legal reasoning tasks, yet their ability to understand long legal documents remains insufficiently evaluated. Developing long-context benchmarks that capture realistic, high-stakes tasks remains a significant challenge in the field, as most existing evaluations rely on simplified synthetic tasks that fail to represent the complexity of real-world document understanding. Overruling relationships are foundational to common-law doctrine and commonly found in judicial opinions. They provide a focused and important testbed for long-document legal understanding that closely resembles what legal professionals actually do. We present an assessment of state-of-the-art LLMs on identifying overruling relationships from U.S. Supreme Court cases using a dataset of 236 case pairs. Our evaluation reveals three critical limitations: (1) era sensitivity -- the models show degraded performance on historical cases compared to modern ones, revealing fundamental temporal bias in their training; (2) shallow reasoning -- models rely on shallow logical heuristics rather than deep legal comprehension; and (3) context-dependent reasoning failures -- models produce temporally impossible relationships in complex open-ended tasks despite maintaining basic temporal awareness in simple contexts. Our work contributes a benchmark that addresses the critical gap in realistic long-context evaluation, providing an environment that mirrors the complexity and stakes of actual legal reasoning tasks.