Civil litigation is inherently a life-cycle process: what a lawyer drafts on day one constrains what unfolds at trial months later. Yet existing legal benchmarks evaluate isolated subtasks, and prior legal-agent simulators reinitialize each scenario from shared ground truth, leaving cross-stage causal dependencies unmodeled.
As agents grow more capable, legal-domain LLM agents promise to turn document-heavy matters into reviewable work products -- yet reliable deployment faces three obstacles: no large-scale evidence on how today's strongest model-and-harness combinations behave on end-to-end legal matters; no agent architecture adapted to the legal vertical, only general-purpose harnesses; and, in a setting that keeps shifting with new facts, authorities, and deadlines, no mechanism for systems to learn from their own outcomes. We address each.
arXiv:2606. 04602v1 Announce Type: new Abstract: As agents grow more capable, legal-domain LLM agents promise to turn document-heavy matters into reviewable work products -- yet reliable deployment faces three obstacles: no large-scale evidence on how today's strongest model-and-harness combinations behave on end-to-end legal matters; no agent architecture adapted to the legal vertical, only general-purpose harnesses; and, in a setting that keeps shifting with new facts, authorities, and deadlines, no mechanism for systems to learn from their own outcomes.
arXiv:2608. 14610v1 Announce Type: new Abstract: Legal reasoning tasks such as legal judgment prediction (LJP) require identifying the temporally correct version of the law governing a case -- a capability we term temporal applicable-law determination.
LEGO is a dual‑module framework that combines a Legal Expert GraphRAG system with an expert Chain‑of‑Thought approach to enhance complex legal reasoning. The GraphRAG component uses an expert‑annotated civil code graph and a greedy normative‑coverage retrieval algorithm to extract relevant provision subgraphs, while the Chain‑of‑Thought module structures retrieved provisions and case facts into a Provision‑Fact‑Conclusion reasoning flow. Using a Qwen3‑8B backbone, LEGO achieves 40.53% exact‑match accuracy on LawExamQA_Civil, surpassing baseline RAG and CoT models and matching larger models on multi‑hop and open‑ended benchmarks, with ablation studies confirming the complementary benefits of both modules.
LegalEvolver is a query‑optimization framework for legal judgment prediction that generates legally informative queries to improve retrieval quality. It uses heuristic query seeding, summarizes rollout experiences into rubrics, and applies rubric‑guided reinforcement learning for iterative self‑evolution. Experiments on large benchmarks show that evolving rubrics outperform fixed ones and that rubric guidance boosts performance beyond direct supervised fine‑tuning plus RL.
arXiv:2605.25920v2 Announce Type: replace Abstract: While large language models (LLMs) augmented with agentic search capabilities show promise for legal reasoning, they overlook a fundamental constra...
arXiv:2610.19427v1 Announce Type: new Abstract: Legal reasoning is a fundamental capability for large language models (LLMs) to complete legal AI tasks such as legal judgment prediction. In legal pra...
The study examines whether large language models (LLMs) can perform legally meaningful reasoning by testing OpenAI GPT 5.4 on European Court of Human Rights case forecasting. Using various prompting strategies, the authors find that the model produces structurally complete but substantively shallow analyses, and that LLM-as-a-Judge evaluators are internally consistent yet only weakly aligned with human annotators. The expert-curated prompt yields more comprehensive reasoning but does not improve prediction accuracy, leading the authors to caution against relying solely on automated LLM evaluation or using task accuracy as a proxy for reasoning quality.
arXiv:2606. 30906v1 Announce Type: new Abstract: Artificial Intelligence is increasingly applied to the field of law, and has the potential to increase access to justice.
As large language models are increasingly deployed as tool-augmented legal agents, they introduce agentic hallucinations where tool-call and reasoning errors cascade into fabricated holdings and misci...
The paper investigates how large language models (LLMs) internally encode legal knowledge, specifically the 26 CATO factors used in case-based legal reasoning. Experiments on three open models (Qwen3-1.7B, Qwen3-8B, Gemma-2 2B) demonstrate that all factors are linearly decodable and can be localized to single-layer readouts, with high test AUCs (0.86–1.00). The study identifies 78 model‑specific layer‑feature coordinates, linking symbolic legal abstractions (HYPO and CATO) to learned internal representations in LLMs.