WarriorMath: Enhancing the Mathematical Ability of Large Language Models with a Defect-aware Framework
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866909718748856320 |
|---|---|
| author | Chen, Yue He, Minghua Yang, Fangkai Zhao, Pu Wang, Lu Kang, Yu Dong, Yifei Zhan, Yuefeng Sun, Hao Lin, Qingwei Rajmohan, Saravan Zhang, Dongmei |
| author_facet | Chen, Yue He, Minghua Yang, Fangkai Zhao, Pu Wang, Lu Kang, Yu Dong, Yifei Zhan, Yuefeng Sun, Hao Lin, Qingwei Rajmohan, Saravan Zhang, Dongmei |
| contents | Large Language Models (LLMs) excel in solving mathematical problems, yet their performance is often limited by the availability of high-quality, diverse training data. Existing methods focus on augmenting datasets through rephrasing or difficulty progression but overlook the specific failure modes of LLMs. This results in synthetic questions that the model can already solve, providing minimal performance gains. To address this, we propose WarriorMath, a defect-aware framework for mathematical problem solving that integrates both targeted data synthesis and progressive training. In the synthesis stage, we employ multiple expert LLMs in a collaborative process to generate, critique, and refine problems. Questions that base LLMs fail to solve are identified and iteratively improved through expert-level feedback, producing high-quality, defect-aware training data. In the training stage, we introduce a progressive learning framework that iteratively fine-tunes the model using increasingly challenging data tailored to its weaknesses. Experiments on six mathematical benchmarks show that WarriorMath outperforms strong baselines by 12.57% on average, setting a new state-of-the-art. Our results demonstrate the effectiveness of a defect-aware, multi-expert framework for improving mathematical ability. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2508_01245 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | WarriorMath: Enhancing the Mathematical Ability of Large Language Models with a Defect-aware Framework Chen, Yue He, Minghua Yang, Fangkai Zhao, Pu Wang, Lu Kang, Yu Dong, Yifei Zhan, Yuefeng Sun, Hao Lin, Qingwei Rajmohan, Saravan Zhang, Dongmei Computation and Language Large Language Models (LLMs) excel in solving mathematical problems, yet their performance is often limited by the availability of high-quality, diverse training data. Existing methods focus on augmenting datasets through rephrasing or difficulty progression but overlook the specific failure modes of LLMs. This results in synthetic questions that the model can already solve, providing minimal performance gains. To address this, we propose WarriorMath, a defect-aware framework for mathematical problem solving that integrates both targeted data synthesis and progressive training. In the synthesis stage, we employ multiple expert LLMs in a collaborative process to generate, critique, and refine problems. Questions that base LLMs fail to solve are identified and iteratively improved through expert-level feedback, producing high-quality, defect-aware training data. In the training stage, we introduce a progressive learning framework that iteratively fine-tunes the model using increasingly challenging data tailored to its weaknesses. Experiments on six mathematical benchmarks show that WarriorMath outperforms strong baselines by 12.57% on average, setting a new state-of-the-art. Our results demonstrate the effectiveness of a defect-aware, multi-expert framework for improving mathematical ability. |
| title | WarriorMath: Enhancing the Mathematical Ability of Large Language Models with a Defect-aware Framework |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2508.01245 |