Legal$Δ$: Enhancing Legal Reasoning in LLMs via Reinforcement Learning with Chain-of-Thought Guided Information Gain

• URL: http://arxiv.org/abs/2508.12281v2
• Date: Tue, 19 Aug 2025 02:05:06 GMT
• Title: Legal$Δ$: Enhancing Legal Reasoning in LLMs via Reinforcement Learning with Chain-of-Thought Guided Information Gain
• Authors: Xin Dai, Buqiang Xu, Zhenghao Liu, Yukun Yan, Huiyuan Xie, Xiaoyuan Yi, Shuo Wang, Ge Yu,
• Abstract summary: We propose Legal$Delta$ to enhance legal reasoning through chain-of-thought guided information gain.<n>Legal$Delta$ employs a dual-mode input setup-comprising direct answer and reasoning-augmented modes.<n>It consistently produces more robust and trustworthy legal judgments without relying on labeled preference data.
• Score: 21.20249684727035
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Legal Artificial Intelligence (LegalAI) has achieved notable advances in automating judicial decision-making with the support of Large Language Models (LLMs). However, existing legal LLMs still struggle to generate reliable and interpretable reasoning processes. They often default to fast-thinking behavior by producing direct answers without explicit multi-step reasoning, limiting their effectiveness in complex legal scenarios that demand rigorous justification. To address this challenge, we propose Legal$\Delta$, a reinforcement learning framework designed to enhance legal reasoning through chain-of-thought guided information gain. During training, Legal$\Delta$ employs a dual-mode input setup-comprising direct answer and reasoning-augmented modes-and maximizes the information gain between them. This encourages the model to acquire meaningful reasoning patterns rather than generating superficial or redundant explanations. Legal$\Delta$ follows a two-stage approach: (1) distilling latent reasoning capabilities from a powerful Large Reasoning Model (LRM), DeepSeek-R1, and (2) refining reasoning quality via differential comparisons, combined with a multidimensional reward mechanism that assesses both structural coherence and legal-domain specificity. Experimental results on multiple legal reasoning tasks demonstrate that Legal$\Delta$ outperforms strong baselines in both accuracy and interpretability. It consistently produces more robust and trustworthy legal judgments without relying on labeled preference data. All code and data will be released at https://github.com/NEUIR/LegalDelta.

Related papers

• Towards Generalizable Reasoning: Group Causal Counterfactual Policy Optimization for LLM Reasoning [50.352417879912515]
  Large language models (LLMs) excel at complex tasks with advances in reasoning capabilities.<n>We propose Group Causal Counterfactual Policy Optimization to explicitly train LLMs to learn generalizable reasoning patterns.<n>We then construct token-level advantages from this reward and optimize the policy, encouraging LLMs to favor reasoning patterns that are process-valid and counterfactually robust.
  arXiv  Detail & Related papers  (2026-02-06T08:03:11Z)
• Multi-Source Retrieval and Reasoning for Legal Sentencing Prediction [50.6851250608938]
  Legal sentencing prediction (LSP) remains difficult due to its need for fine-grained objective knowledge and flexible subjective reasoning.<n>We propose $MSR2$, a framework that integrates multi-source retrieval and reasoning in LLMs with reinforcement learning.<n>Experiments on two real-world datasets show that $MSR2$ improves both accuracy and interpretability in LSP.
  arXiv  Detail & Related papers  (2026-02-04T15:55:55Z)
• LegalOne: A Family of Foundation Models for Reliable Legal Reasoning [54.57434222018289]
  We present LegalOne, a family of foundational models specifically tailored for the Chinese legal domain.<n>LegalOne is developed through a comprehensive three-phase pipeline designed to master legal reasoning.<n>We publicly release the LegalOne weights and the LegalKit evaluation framework to advance the field of Legal AI.
  arXiv  Detail & Related papers  (2026-01-31T10:18:32Z)
• Adversarial Yet Cooperative: Multi-Perspective Reasoning in Retrieved-Augmented Language Models [72.4149653187766]
  We propose a Reasoner-Verifier framework named Adrialversa Reasoning RAG (ARR)<n>The Reasoner and Verifier engage in reasoning on retrieved evidence and critiquing each other's logic while being guided by process-aware advantage.<n> Experiments on multiple benchmarks demonstrate the effectiveness of our method.
  arXiv  Detail & Related papers  (2026-01-08T06:57:03Z)
• When Reasoning Meets Its Laws [45.03949352588766]
  This paper presents the Laws of Reasoning (LoRe), a unified framework that characterizes intrinsic reasoning patterns in LRMs.<n>We first propose compute law with the hypothesis that the reasoning compute should scale linearly with question complexity.<n>Since the question complexity is difficult to quantify in practice, we examine these hypotheses by two properties of the laws, monotonicity and compositionality.
  arXiv  Detail & Related papers  (2025-12-19T18:59:11Z)
• Benchmarking Multi-Step Legal Reasoning and Analyzing Chain-of-Thought Effects in Large Language Models [8.769542756426786]
  We introduce M SLR, the first Chinese multi-step legal reasoning dataset grounded in real-world judicial decision making.<n>M SLR adopts the IRAC framework (Issue, Rule, Application, Conclusion) to model structured expert reasoning from official legal documents.<n>We design a scalable Human-LLM collaborative annotation pipeline that efficiently produces fine-grained step-level reasoning annotations.<n>Further experiments demonstrate that Self-Initiated Chain-of-Thought prompts generated by models autonomously improve reasoning coherence and quality, outperforming human-designed prompts.
  arXiv  Detail & Related papers  (2025-11-11T08:45:29Z)
• GLARE: Agentic Reasoning for Legal Judgment Prediction [60.13483016810707]
  Legal judgment prediction (LJP) has become increasingly important in the legal field.<n>Existing large language models (LLMs) have significant problems of insufficient reasoning due to a lack of legal knowledge.<n>We introduce GLARE, an agentic legal reasoning framework that dynamically acquires key legal knowledge by invoking different modules.
  arXiv  Detail & Related papers  (2025-08-22T13:38:12Z)
• Reason-to-Recommend: Using Interaction-of-Thought Reasoning to Enhance LLM Recommendation [9.282278040339138]
  $textbfR2Rec$ is a reasoning-enhanced recommendation framework.<n>It samples interaction chains from the user-item graph and converts them into structured interaction-of-thoughts.
  arXiv  Detail & Related papers  (2025-06-05T14:16:44Z)
• CoThink: Token-Efficient Reasoning via Instruct Models Guiding Reasoning Models [56.40065909544213]
  Large language models (LLMs) benefit from increased test-time compute, a phenomenon known as test-time scaling.<n>However, reasoning-optimized models often overthink even simple problems, producing excessively verbose outputs and leading to low token efficiency.<n>We identify two key causes of this verbosity: (1) reinforcement learning reduces the information density of forward reasoning, and (2) backward chain-of thought training encourages redundant and often unnecessary verification steps.
  arXiv  Detail & Related papers  (2025-05-28T06:24:45Z)
• RLJP: Legal Judgment Prediction via First-Order Logic Rule-enhanced with Large Language Models [58.69183479148083]
  Legal Judgment Prediction (LJP) is a pivotal task in legal AI.<n>Existing LJP models integrate judicial precedents and legal knowledge for high performance.<n>But they neglect legal reasoning logic, a critical component of legal judgments requiring rigorous logical analysis.<n>This paper proposes a rule-enhanced legal judgment prediction framework based on first-order logic (FOL) formalism and comparative learning (CL)
  arXiv  Detail & Related papers  (2025-05-27T14:50:21Z)
• Reinforced Latent Reasoning for LLM-based Recommendation [83.18146814163308]
  Large Language Models (LLMs) have demonstrated impressive reasoning capabilities in complex problem-solving tasks.<n>Existing methods typically rely on fine-tuning with explicit chain-of-thought (CoT) data.<n>In this work, we explore an alternative approach that shifts from explicit CoT reasoning to compact, information-dense latent reasoning.
  arXiv  Detail & Related papers  (2025-05-25T11:03:45Z)
• Beyond 'Aha!': Toward Systematic Meta-Abilities Alignment in Large Reasoning Models [86.88657425848547]
  Large reasoning models (LRMs) already possess a latent capacity for long chain-of-thought reasoning.<n>We explicitly align models with three meta-abilities: deduction, induction, and abduction, using automatically generated, self-verifiable tasks.<n>Our three stage-pipeline individual alignment, parameter-space merging, and domain-specific reinforcement learning, boosts performance by over 10% relative to instruction-tuned baselines.
  arXiv  Detail & Related papers  (2025-05-15T17:58:33Z)
• An Explicit Syllogistic Legal Reasoning Framework for Large Language Models [5.501226256903341]
  Large language models (LLMs) can answer legal questions, but often struggle with explicit syllogistic reasoning.<n>We introduce SyLeR, a novel framework designed to enable LLMs to perform explicit syllogistic legal reasoning.<n>SyLeR employs a tree-structured hierarchical retrieval mechanism to synthesize relevant legal statutes and precedents.
  arXiv  Detail & Related papers  (2025-04-05T03:34:51Z)
• Legal Mathematical Reasoning with LLMs: Procedural Alignment through Two-Stage Reinforcement Learning [12.90492832643565]
  Legal mathematical reasoning is essential for applying large language models (LLMs) in high-stakes legal contexts.<n>We present LexNum, the first Chinese legal mathematical reasoning benchmark.<n>We also propose LexPam, a two-stage reinforcement learning framework for efficient legal reasoning training.
  arXiv  Detail & Related papers  (2025-04-03T13:54:53Z)
• Reasoning with Latent Thoughts: On the Power of Looped Transformers [52.84192961524481]
  We show that for many synthetic reasoning problems, a $k$-layer transformer looped $L$ times nearly matches the performance of a $kL$-layer non-looped model.<n>Our empirical analysis reveals an intriguing phenomenon: looped and non-looped models exhibit scaling behavior that depends on their effective depth.
  arXiv  Detail & Related papers  (2025-02-24T18:49:05Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.