Librarian Bot. I found the following papers similar to this paper.

The following papers were recommended by the Semantic Scholar API

Fin-R1: A Large Language Model for Financial Reasoning through Reinforcement Learning (2025)
SFT or RL? An Early Investigation into Training R1-Like Reasoning Large Vision-Language Models (2025)
SARI: Structured Audio Reasoning via Curriculum-Guided Reinforcement Learning (2025)
Exploring the Effect of Reinforcement Learning on Video Understanding: Insights from SEED-Bench-R1 (2025)
Leveraging Reasoning Model Answers to Enhance Non-Reasoning Model Capability (2025)
Nemotron-CrossThink: Scaling Self-Learning beyond Math Reasoning (2025)
Chain-of-Thought Matters: Improving Long-Context Language Models with Reasoning Path Supervision (2025)

Please give a thumbs up to this comment if you found it helpful!

If you want recommendations for any Paper on Hugging Face checkout this Space

You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: \n\n@librarian-bot\n\t recommend

\n","updatedAt":"2025-04-29T01:34:59.922Z","author":{"_id":"63d3e0e8ff1384ce6c5dd17d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg","fullname":"Librarian Bot (Bot)","name":"librarian-bot","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":264}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.7316564321517944},"editors":["librarian-bot"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2504.15716","authors":[{"_id":"680dcc5d3478de07603a8036","user":{"_id":"642656cbad1e3b0e6e91b752","avatarUrl":"/avatars/3bf0ee15fd528e09b2b889f5cce3cbd0.svg","isPro":false,"fullname":"Jie Zhu","user":"amazingj","type":"user"},"name":"Jie Zhu","status":"claimed_verified","statusLastChangedAt":"2025-04-28T07:39:02.713Z","hidden":false},{"_id":"680dcc5d3478de07603a8037","name":"Qian Chen","hidden":false},{"_id":"680dcc5d3478de07603a8038","name":"Huaixia Dou","hidden":false},{"_id":"680dcc5d3478de07603a8039","name":"Junhui Li","hidden":false},{"_id":"680dcc5d3478de07603a803a","user":{"_id":"62f0bb7f47cf7b3344fe9ad6","avatarUrl":"/avatars/e509f31af551cfd30e319b03d4c92740.svg","isPro":false,"fullname":"guo","user":"lifan","type":"user"},"name":"Lifan Guo","status":"admin_assigned","statusLastChangedAt":"2025-04-29T12:41:24.106Z","hidden":false},{"_id":"680dcc5d3478de07603a803b","name":"Feng Chen","hidden":false},{"_id":"680dcc5d3478de07603a803c","name":"Chi Zhang","hidden":false}],"publishedAt":"2025-04-22T09:01:04.000Z","submittedOnDailyAt":"2025-04-28T06:16:26.234Z","title":"DianJin-R1: Evaluating and Enhancing Financial Reasoning in Large\n Language Models","submittedOnDailyBy":{"_id":"642656cbad1e3b0e6e91b752","avatarUrl":"/avatars/3bf0ee15fd528e09b2b889f5cce3cbd0.svg","isPro":false,"fullname":"Jie Zhu","user":"amazingj","type":"user"},"summary":"Effective reasoning remains a core challenge for large language models (LLMs)\nin the financial domain, where tasks often require domain-specific knowledge,\nprecise numerical calculations, and strict adherence to compliance rules. We\npropose DianJin-R1, a reasoning-enhanced framework designed to address these\nchallenges through reasoning-augmented supervision and reinforcement learning.\nCentral to our approach is DianJin-R1-Data, a high-quality dataset constructed\nfrom CFLUE, FinQA, and a proprietary compliance corpus (Chinese Compliance\nCheck, CCC), combining diverse financial reasoning scenarios with verified\nannotations. Our models, DianJin-R1-7B and DianJin-R1-32B, are fine-tuned from\nQwen2.5-7B-Instruct and Qwen2.5-32B-Instruct using a structured format that\ngenerates both reasoning steps and final answers. To further refine reasoning\nquality, we apply Group Relative Policy Optimization (GRPO), a reinforcement\nlearning method that incorporates dual reward signals: one encouraging\nstructured outputs and another rewarding answer correctness. We evaluate our\nmodels on five benchmarks: three financial datasets (CFLUE, FinQA, and CCC) and\ntwo general reasoning benchmarks (MATH-500 and GPQA-Diamond). Experimental\nresults show that DianJin-R1 models consistently outperform their non-reasoning\ncounterparts, especially on complex financial tasks. Moreover, on the\nreal-world CCC dataset, our single-call reasoning models match or even surpass\nthe performance of multi-agent systems that require significantly more\ncomputational cost. These findings demonstrate the effectiveness of DianJin-R1\nin enhancing financial reasoning through structured supervision and\nreward-aligned learning, offering a scalable and practical solution for\nreal-world applications.","upvotes":12,"discussionId":"680dcc5e3478de07603a807e","ai_summary":"A reasoning-enhanced framework, DianJin-R1, is proposed to improve financial domain reasoning in LLMs through augmented supervision and reinforcement learning, achieving superior performance on financial benchmarks.","ai_keywords":["reasoning-augmented supervision","reinforcement learning","DianJin-R1-Data","Qwen2.5-7B-Instruct","Qwen2.5-32B-Instruct","Group Relative Policy Optimization","GRPO","structured outputs","answer correctness","CFLUE","FinQA","CCC","MATH-500","GPQA-Diamond","single-call reasoning","multi-agent systems"]},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"642656cbad1e3b0e6e91b752","avatarUrl":"/avatars/3bf0ee15fd528e09b2b889f5cce3cbd0.svg","isPro":false,"fullname":"Jie Zhu","user":"amazingj","type":"user"},{"_id":"67b1a0d17c5844c6bcf091db","avatarUrl":"/avatars/bf933b81eb35b529741c1704ad8ef73b.svg","isPro":false,"fullname":"benqi","user":"Aidabenk","type":"user"},{"_id":"648eb1eb59c4e5c87dc116e0","avatarUrl":"/avatars/c636cea39c2c0937f01398c94ead5dad.svg","isPro":false,"fullname":"fdsqefsgergd","user":"T-representer","type":"user"},{"_id":"67b19ed4626cd81034706905","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/67b19ed4626cd81034706905/VZrmCTkaBtT0IiRCVKATJ.png","isPro":false,"fullname":"Daixiny","user":"DaiyuXX","type":"user"},{"_id":"620783f24e28382272337ba4","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/620783f24e28382272337ba4/zkUveQPNiDfYjgGhuFErj.jpeg","isPro":false,"fullname":"GuoLiangTang","user":"Tommy930","type":"user"},{"_id":"665b133508d536a8ac804f7d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/Uwi0OnANdTbRbHHQvGqvR.png","isPro":false,"fullname":"Paulson","user":"Pnaomi","type":"user"},{"_id":"666aad704649c647cfa4a87e","avatarUrl":"/avatars/da24a587a58d65c1f92e27d2f779eee1.svg","isPro":false,"fullname":"Muthuraj Ponnusamy","user":"radirajjj","type":"user"},{"_id":"64d4615cf8082bf19b916492","avatarUrl":"/avatars/8e1b59565ec5e4b31090cf1b911781b9.svg","isPro":false,"fullname":"wongyukim","user":"wongyukim","type":"user"},{"_id":"5f43448a79c1ba4c353d0d8f","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/5f43448a79c1ba4c353d0d8f/DiSygV3dn7A_OjmGVTrHD.jpeg","isPro":true,"fullname":"Sugato Ray","user":"sugatoray","type":"user"},{"_id":"5e4d7e6a37cb5b49818287c1","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/5e4d7e6a37cb5b49818287c1/-lu-COKEu04rLFkowvJ93.png","isPro":false,"fullname":"Pratik Bhavsar","user":"pratikbhavsar","type":"user"},{"_id":"663ccbff3a74a20189d4aa2e","avatarUrl":"/avatars/83a54455e0157480f65c498cd9057cf2.svg","isPro":false,"fullname":"Nguyen Van Thanh","user":"NguyenVanThanhHust","type":"user"},{"_id":"6578f357e390cfd409bda675","avatarUrl":"/avatars/c513b5f953dd787ed8bdb9e7cf31ed9e.svg","isPro":false,"fullname":"Chris Concannon","user":"choncan","type":"user"}],"acceptLanguages":["*"],"dailyPaperRank":0}">

Papers

arxiv:2504.15716

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

Published on Apr 22

· Submitted by

Jie Zhu on Apr 28

Upvote

Authors:

Jie Zhu ,

Lifan Guo ,

Abstract

A reasoning-enhanced framework, DianJin-R1, is proposed to improve financial domain reasoning in LLMs through augmented supervision and reinforcement learning, achieving superior performance on financial benchmarks.

AI-generated summary

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.