DianJin-R1: Evaluating and Enhancing Financial Reasoning in Large Language Models

• URL: http://arxiv.org/abs/2504.15716v1
• Date: Tue, 22 Apr 2025 09:01:04 GMT
• Title: DianJin-R1: Evaluating and Enhancing Financial Reasoning in Large Language Models
• Authors: Jie Zhu, Qian Chen, Huaixia Dou, Junhui Li, Lifan Guo, Feng Chen, Chi Zhang,
• Abstract summary: We propose DianJin-R1, a reasoning-enhanced framework for large language models (LLMs) in the financial domain.<n>Central to our approach is DianJin-R1-Data, a high-quality dataset constructed from CFLUE, FinQA, and a proprietary compliance corpus (Chinese Compliance Check, CCC)<n>Our models, DianJin-R1-7B and DianJin-R1-32B, are fine-tuned from Qwen2.5-7B-Instruct and Qwen2.5-32B-Instruct using a structured format that generates both reasoning steps and final answers.
• Score: 13.567516575993546
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Effective reasoning remains a core challenge for large language models (LLMs) in the financial domain, where tasks often require domain-specific knowledge, precise numerical calculations, and strict adherence to compliance rules. We propose DianJin-R1, a reasoning-enhanced framework designed to address these challenges through reasoning-augmented supervision and reinforcement learning. Central to our approach is DianJin-R1-Data, a high-quality dataset constructed from CFLUE, FinQA, and a proprietary compliance corpus (Chinese Compliance Check, CCC), combining diverse financial reasoning scenarios with verified annotations. Our models, DianJin-R1-7B and DianJin-R1-32B, are fine-tuned from Qwen2.5-7B-Instruct and Qwen2.5-32B-Instruct using a structured format that generates both reasoning steps and final answers. To further refine reasoning quality, we apply Group Relative Policy Optimization (GRPO), a reinforcement learning method that incorporates dual reward signals: one encouraging structured outputs and another rewarding answer correctness. We evaluate our models on five benchmarks: three financial datasets (CFLUE, FinQA, and CCC) and two general reasoning benchmarks (MATH-500 and GPQA-Diamond). Experimental results show that DianJin-R1 models consistently outperform their non-reasoning counterparts, especially on complex financial tasks. Moreover, on the real-world CCC dataset, our single-call reasoning models match or even surpass the performance of multi-agent systems that require significantly more computational cost. These findings demonstrate the effectiveness of DianJin-R1 in enhancing financial reasoning through structured supervision and reward-aligned learning, offering a scalable and practical solution for real-world applications.

Related papers

• Agentar-Fin-R1: Enhancing Financial Intelligence through Domain Expertise, Training Efficiency, and Advanced Reasoning [12.548390779247987]
  We introduce the Agentar-Fin-R1 series of financial large language models.<n>Our optimization approach integrates a high-quality, systematic financial task label system.<n>Our models undergo comprehensive evaluation on mainstream financial benchmarks.
  arXiv  Detail & Related papers  (2025-07-22T17:52:16Z)
• ReasonFlux-PRM: Trajectory-Aware PRMs for Long Chain-of-Thought Reasoning in LLMs [56.32212611983997]
  We introduce ReasonFlux-PRM, a novel trajectory-aware PRM to evaluate trajectory-response type of reasoning traces.<n>ReasonFlux-PRM incorporates both step-level and trajectory-level supervision, enabling fine-grained reward assignment aligned with structured chain-of-thought data.<n>Our derived ReasonFlux-PRM-7B yields consistent performance improvements, achieving average gains of 12.1% in supervised fine-tuning, 4.5% in reinforcement learning, and 6.3% in test-time scaling.
  arXiv  Detail & Related papers  (2025-06-23T17:59:02Z)
• GRPO-CARE: Consistency-Aware Reinforcement Learning for Multimodal Reasoning [53.894789613838654]
  We introduce SEED-Bench-R1, a benchmark with complex real-world videos requiring balanced perception and reasoning.<n>Using SEED-Bench-R1, we find that standard GRPO, while improving answer accuracy, often reduces logical coherence between reasoning steps and answers, with only a 57.9% consistency rate.<n>We propose GRPO-CARE, a consistency-aware RL framework optimizing both answer correctness and reasoning coherence without explicit supervision.
  arXiv  Detail & Related papers  (2025-06-19T08:49:13Z)
• RKEFino1: A Regulation Knowledge-Enhanced Large Language Model [3.3128149968030436]
  We propose RKEFino, a regulation knowledge-enhanced financial reasoning model built upon Fino.<n>We formulate two QA tasks-based and mathematical reasoning-and introduce a novel Numerical NER task covering financial entities in both sentences and tables.<n> Experimental results demonstrate the effectiveness and generalization capacity of RKEFino1 in compliance-critical financial tasks.
  arXiv  Detail & Related papers  (2025-06-06T03:02:52Z)
• Discriminative Policy Optimization for Token-Level Reward Models [55.98642069903191]
  Process reward models (PRMs) provide more nuanced supervision compared to outcome reward models (ORMs)<n>Q-RM explicitly learns token-level Q-functions from preference data without relying on fine-grained annotations.<n>Reinforcement learning with Q-RM significantly enhances training efficiency, achieving convergence 12 times faster than ORM on GSM8K and 11 times faster than step-level PRM on MATH.
  arXiv  Detail & Related papers  (2025-05-29T11:40:34Z)
• General-Reasoner: Advancing LLM Reasoning Across All Domains [64.70599911897595]
  Reinforcement learning (RL) has recently demonstrated strong potential in enhancing the reasoning capabilities of large language models (LLMs)<n>We propose General-Reasoner, a novel training paradigm designed to enhance LLM reasoning capabilities across diverse domains.<n>We train a series of models and evaluate them on a wide range of datasets covering wide domains like physics, chemistry, finance, electronics etc.
  arXiv  Detail & Related papers  (2025-05-20T17:41:33Z)
• RM-R1: Reward Modeling as Reasoning [81.50471199906738]
  Reasoning Reward Models (ReasRMs) formulate reward modeling as a reasoning task.<n>We propose a reasoning-oriented training pipeline and train a family of ReasRMs, RM-R1.<n>Our models achieve state-of-the-art performance across three reward model benchmarks on average.
  arXiv  Detail & Related papers  (2025-05-05T06:11:12Z)
• NEMOTRON-CROSSTHINK: Scaling Self-Learning beyond Math Reasoning [62.88540902786668]
  Large Language Models (LLMs) have shown strong reasoning capabilities, particularly when enhanced through Reinforcement Learning (RL)<n>We propose NEMOTRON-CROSSTHINK, a framework that systematically incorporates multi-domain corpora, including both synthetic and real-world question-answer pairs, into RL training to improve generalization across diverse reasoning tasks.
  arXiv  Detail & Related papers  (2025-04-15T21:37:13Z)
• Fin-R1: A Large Language Model for Financial Reasoning through Reinforcement Learning [17.649686407321923]
  We introduce Fin-R1, a reasoning large language model specifically designed for the financial sector.<n>Fin-R1 is built using a two-stage architecture, leveraging a financial reasoning dataset distilled and processed based on DeepSeek-R1.<n>It demonstrates performance close to DeepSeek-R1 with a parameter size of 7 billion across a range of financial reasoning tasks.
  arXiv  Detail & Related papers  (2025-03-20T15:46:18Z)
• A Survey on Post-training of Large Language Models [185.51013463503946]
  Large Language Models (LLMs) have fundamentally transformed natural language processing, making them indispensable across domains ranging from conversational systems to scientific exploration.<n>These challenges necessitate advanced post-training language models (PoLMs) to address shortcomings, such as restricted reasoning capacities, ethical uncertainties, and suboptimal domain-specific performance.<n>This paper presents the first comprehensive survey of PoLMs, systematically tracing their evolution across five core paradigms.
  arXiv  Detail & Related papers  (2025-03-08T05:41:42Z)
• Reward Models Identify Consistency, Not Causality [54.987590763737145]
  State-of-the-art reward models prioritize structural consistency over causal correctness.<n>Removing the problem statement has minimal impact on reward scores.<n> altering numerical values or disrupting the reasoning flow significantly affects RM outputs.
  arXiv  Detail & Related papers  (2025-02-20T14:57:14Z)
• FinMTEB: Finance Massive Text Embedding Benchmark [18.990655668481075]
  We introduce the Finance Massive Text Embedding Benchmark (FinMTEB), a specialized counterpart to MTEB designed for the financial domain.<n>FinMTEB comprises 64 financial domain-specific embedding datasets across 7 tasks.<n>We show three key findings: (1) performance on general-purpose benchmarks shows limited correlation with financial domain tasks; (2) domain-adapted models consistently outperform their general-purpose counterparts; and (3) surprisingly, a simple Bag-of-Words approach outperforms sophisticated dense embeddings in financial Semantic Textual Similarity tasks.
  arXiv  Detail & Related papers  (2025-02-16T04:23:52Z)
• Fino1: On the Transferability of Reasoning Enhanced LLMs to Finance [32.516564836540745]
  Large language models (LLMs) have shown strong general reasoning capabilities, but their effectiveness in financial reasoning remains underexplored.<n>We evaluate 24 state-of-the-art general and reasoning-focused LLMs across four complex financial reasoning tasks.<n>We propose two domain-adapted models, Fino1-8B and FinoB, trained with chain-of-thought (CoT) fine-tuning and reinforcement learning.
  arXiv  Detail & Related papers  (2025-02-12T05:13:04Z)
• Large Language Models Meet Symbolic Provers for Logical Reasoning Evaluation [24.081573908824353]
  First-order logic (FOL) reasoning is pivotal for intelligent systems. Existing benchmarks often rely on extensive human annotation or handcrafted templates. We propose a novel framework called ProverGen that synergizes the generative strengths of Large Language Models with the rigor and precision of symbolic provers.
  arXiv  Detail & Related papers  (2025-02-10T15:31:54Z)
• MR-Ben: A Meta-Reasoning Benchmark for Evaluating System-2 Thinking in LLMs [55.20845457594977]
  Large language models (LLMs) have shown increasing capability in problem-solving and decision-making.<n>We present a process-based benchmark MR-Ben that demands a meta-reasoning skill.<n>Our meta-reasoning paradigm is especially suited for system-2 slow thinking.
  arXiv  Detail & Related papers  (2024-06-20T03:50:23Z)
• Advancing LLM Reasoning Generalists with Preference Trees [119.57169648859707]
  We introduce Eurus, a suite of large language models (LLMs) optimized for reasoning. Eurus models achieve state-of-the-art results among open-source models on a diverse set of benchmarks.
  arXiv  Detail & Related papers  (2024-04-02T16:25:30Z)
• FinGPT: Instruction Tuning Benchmark for Open-Source Large Language Models in Financial Datasets [9.714447724811842]
  This paper introduces a distinctive approach anchored in the Instruction Tuning paradigm for open-source large language models. We capitalize on the interoperability of open-source models, ensuring a seamless and transparent integration. The paper presents a benchmarking scheme designed for end-to-end training and testing, employing a cost-effective progression.
  arXiv  Detail & Related papers  (2023-10-07T12:52:58Z)
• PIXIU: A Large Language Model, Instruction Data and Evaluation Benchmark for Finance [63.51545277822702]
  PIXIU is a comprehensive framework including the first financial large language model (LLMs) based on fine-tuning LLaMA with instruction data. We propose FinMA by fine-tuning LLaMA with the constructed dataset to be able to follow instructions for various financial tasks. We conduct a detailed analysis of FinMA and several existing LLMs, uncovering their strengths and weaknesses in handling critical financial tasks.
  arXiv  Detail & Related papers  (2023-06-08T14:20:29Z)

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.