Elamur Research⌕
☰
home / papers / 2025 / Superficial Self-Improved Reasoners Benefit from Model Merging

Superficial Self-Improved Reasoners Benefit from Model Merging

Paper · 2025 · Xiangchi Yuan, Chunhui Zhang, Zheyuan Liu et al.

Summary

Self-improvement, where a language model is fine-tuned on reasoning traces it generated itself, is often proposed as a remedy for scarce high-quality data. Prior work has mostly worried about model collapse caused by shrinking output diversity. Yuan et al. identify a separate problem they call "Superficial Self-Improved Reasoners": accuracy on in-domain (ID) math benchmarks goes up, but out-of-domain (OOD) reasoning gets worse, which points to memorization rather than real skill gains.

The authors generate chain-of-thought solutions, keep only those whose final answer is correct, and run SFT on them. They then compare each layer's importance for reasoning, estimated from how much the loss changes when its weights are removed, with how much that layer actually changes during training. The largest updates land in layers that matter least for reasoning.

Their fix is Iterative Model Merging (IMM). After each round, IMM blends the weights of the original model with those of the self-improved model. Across iterations, IMM prevents collapse and recovers OOD accuracy, which makes self-improvement loops more reliable.

Key points

  • The models studied are Qwen2.5-0.5B-Instruct, Qwen2.5-1.5B-Instruct and Llama2-7B.
    • GSM8K and MATH serve as the ID datasets.
    • SAT_Math, MAWPS and ASDiv serve as the OOD datasets.
  • Over three self-improvement iterations, the baseline methods collapse by the first or second round. IMM avoids this and gives the best ID scores, including higher pass@k as k grows.
  • Smaller models are hit harder by collapse, and all three model sizes gain from merging.
  • IMM restores OOD accuracy after each merge and sometimes exceeds the base model. The one exception is Qwen2.5-0.5B-Instruct on MAWPS, which the authors suspect is close to the ID data.
  • IMM also helps when distilling from the stronger Qwen2.5-7B-Instruct into Qwen2.5-1.5B-Instruct and Llama-2-7B. ID accuracy holds roughly steady, while OOD accuracy often improves.
  • Limitations noted by the authors:
    • The merge uses fixed weights; dynamic or layer-adaptive merging is left for future work.
    • The best ratio of real to synthetic data is not explored.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓Georgia Institute of TechnologyAtlanta · United States133 works · 24,497 citations🎓University of Notre DameNotre Dame · United States51 works · 36,717 citations🎓Dartmouth CollegeHanover · United States37 works · 42,625 citations

 Fields, topics and concepts

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 works

🧭 Topics 1Research subjects and languages the work is about.Reasoning · 1,426 works

Cite

Cite as Xiangchi Yuan, Chunhui Zhang, Zheyuan Liu, Dachuan Shi, Leyan Pan, Soroush Vosoughi and Wenke Lee. 2025. Superficial Self-Improved Reasoners Benefit from Model Merging. Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2025.emnlp-main.301

@inproceedings{yuan2025superficial,
  title = {{Superficial Self-Improved Reasoners Benefit from Model Merging}},
  author = {Xiangchi Yuan, Chunhui Zhang, Zheyuan Liu, Dachuan Shi, Leyan Pan, Soroush Vosoughi and Wenke Lee},
  booktitle = {Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing},
  year = {2025},
  doi = {10.18653/v1/2025.emnlp-main.301},
  url = {https://doi.org/10.18653/v1/2025.emnlp-main.301}
}

References

78 references, 31 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Sina Alemohammad, Josue Casco-Rodriguez, Lorenzo Luzi, Ahmed Imtiaz Humayun, Hossein Babaei, Daniel LeJeune, Ali Siahkoohi, and Richard Baraniuk. 2024. Self-consuming generative models go MAD. In International Conference on Learning Representations. find
  2. Aida Amini, Saadia Gabriel, Peter Lin, Rik Koncel-Kedziorski, Yejin Choi, and Hannaneh Hajishirzi. 2019. Mathqa: Towards interpretable math word problem solving with operation-based formalisms. arXiv preprint arXiv:1905.13319.
  3. Hritik Bansal, Arian Hosseini, Rishabh Agarwal, Vinh Q Tran, and Mehran Kazemi. 2024. Smaller, weaker, yet better: Training llm reasoners via compute-optimal sampling. arXiv preprint arXiv:2408.16737.
  4. Quentin Bertrand, Avishek Joey Bose, Alexandre Duplessis, Marco Jiralerspong, and Gauthier Gidel. 2023. On the stability of iterative retraining of generative models on their own data. arXiv preprint arXiv:2310.00429. — [arXiv:2310.00429](https://arxiv.org/abs/2310.00429) find
  5. Quentin Bertrand, Joey Bose, Alexandre Duplessis, Marco Jiralerspong, and Gauthier Gidel. 2024. On the stability of iterative retraining of generative models on their own data. In The Twelfth International Conference on Learning Representations. find
  6. Matyas Bohacek and Hany Farid. 2023. Nepotistically trained generative-ai models collapse. arXiv preprint arXiv:2311.12202. — [arXiv:2311.12202](https://arxiv.org/abs/2311.12202) find
  7. Karl Cobbe, Vineet Kosaraju, Mohammad Bavarian, Mark Chen, Heewoo Jun, Lukasz Kaiser, Matthias Plappert, Jerry Tworek, Jacob Hilton, Reiichiro Nakano, et al. 2021. Training verifiers to solve math word problems, 2021. URL https://arxiv.org/abs/2110.14168. ↗
  8. DeepMind. 2024a. Ai solves imo problems at a silver medal level. URL https://deepmind.google/discover/blog/ai-solves-imo-problems-at-silver-medal-level/. ↗
  9. DeepMind. 2024b. Alphazero: Shedding new light on chess, shogi, and go. URL https://deepmind.google/discover/blog/alphazero-shedding-new-light-on-chess-shogi-and-go/. ↗
  10. Xingjian Diao, Chunhui Zhang, Keyi Kong, Weiyi Wu, Chiyu Ma, Zhongyu Ouyang, Peijun Qing, Soroush Vosoughi, and Jiang Gui. 2025. Soundmind: Rl-incentivized logic reasoning for audio-language models. arXiv preprint arXiv:2506.12935.
  11. Elvis Dohmatob, Yunzhen Feng, Arjun Subramonian, and Julia Kempe. 2025. Strong model collapse. In The Thirteenth International Conference on Learning Representations. find
  12. Yunzhen Feng, Elvis Dohmatob, Pu Yang, Francois Charton, and Julia Kempe. 2025. Beyond model collapse: Scaling up with synthesized data requires verification. In The Thirteenth International Conference on Learning Representations. find
  13. Shi Fu, Sen Zhang, Yingjie Wang, Xinmei Tian, and Dacheng Tao. 2024a. Towards theoretical understandings of self-consuming generative models. In Forty-first International Conference on Machine Learning. find
  14. Yonggan Fu, Zhongzhi Yu, Junwei Li, Jiayi Qian, Yongan Zhang, Xiangchi Yuan, Dachuan Shi, Roman Yakunin, and Yingyan Celine Lin. 2024b. Amoeballm: Constructing any-shape large language models for efficient and instant deployment. Advances in Neural Information Processing Systems, 37:78299– 78319. find
  15. Matthias Gerstgrasser, Rylan Schaeffer, Apratim Dey, Rafael Rafailov, Tomasz Korbak, Henry Sleight, Rajashree Agrawal, John Hughes, Dhruv Bhandarkar Pai, Andrey Gromov, Dan Roberts, Diyi Yang, David L. Donoho, and Sanmi Koyejo. 2024. Is model collapse inevitable? breaking the curse of recursion by a… find
  16. Nate Gillman, Michael Freeman, Daksh Aggarwal, Chia-Hong Hsu, Calvin Luo, Yonglong Tian, and Chen Sun. 2024. Self-correcting self-consuming loops for generative model training. In International Conference on Machine Learning. find
  17. Caglar Gulcehre, Tom Le Paine, Srivatsan Srinivasan, Ksenia Konyushkova, Lotte Weerts, Abhishek Sharma, Aditya Siddhant, Alex Ahern, Miaosen Wang, Chenjie Gu, et al. 2023. Reinforced self-training (rest) for language modeling. arXiv preprint arXiv:2308.08998.
  18. Daya Guo, Dejian Yang, Haowei Zhang, Junxiao Song, Ruoyu Zhang, Runxin Xu, Qihao Zhu, Shirong Ma, Peiyi Wang, Xiao Bi, et al. 2025. Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning. arXiv preprint arXiv:2501.12948.
  19. Ruohao Guo, Wei Xu, and Alan Ritter. 2024a. Meta-tuning LLMs to leverage lexical knowledge for generalizable language style understanding. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics. find
  20. Yanzhu Guo, Guokan Shang, Michalis Vazirgiannis, and Chloé Clavel. 2023. The curious decline of linguistic diversity: Training language models on synthetic text. arXiv preprint arXiv:2311.09807. find
  21. Yanzhu Guo, Guokan Shang, Michalis Vazirgiannis, and Chloé Clavel. 2024b. The curious decline of linguistic diversity: Training language models on synthetic text. In Findings of the Association for Computational Linguistics: NAACL 2024. find
  22. Ryuichiro Hataya, Han Bao, and Hiromi Arai. 2023. Will large-scale generative models corrupt future datasets? In Proceedings of the IEEE/CVF International Conference on Computer Vision. find
  23. Ruifei He, Shuyang Sun, Xin Yu, Chuhui Xue, Wenqing Zhang, Philip Torr, Song Bai, and Xiaojuan Qi. 2022. Is synthetic data from generative models ready for image recognition? arXiv preprint arXiv:2210.07574. — [arXiv:2210.07574](https://arxiv.org/abs/2210.07574) find
  24. Dan Hendrycks, Collin Burns, Steven Basart, Andy Zou, Mantas Mazeika, Dawn Song, and Jacob Steinhardt. 2020. Measuring massive multitask language understanding. arXiv preprint arXiv:2009.03300.
  25. Dan Hendrycks, Collin Burns, Saurav Kadavath, Akul Arora, Steven Basart, Eric Tang, Dawn Song, and Jacob Steinhardt. 2021. Measuring mathematical problem solving with the MATH dataset. In Neural Information Processing Systems Datasets and Benchmarks Track.
  26. Arian Hosseini, Xingdi Yuan, Nikolay Malkin, Aaron Courville, Alessandro Sordoni, and Rishabh Agarwal. 2024. V-star: Training verifiers for self-taught reasoners. arXiv preprint arXiv:2402.06457.
  27. Lanxiang Hu, Qiyu Li, Anze Xie, Nan Jiang, Ion Stoica, Haojian Jin, and Hao Zhang. 2024. Gamearena: Evaluating llm reasoning through live computer games. arXiv preprint arXiv:2412.06394. find
  28. Jiaxin Huang, Shixiang Shane Gu, Le Hou, Yuexin Wu, Xuezhi Wang, Hongkun Yu, and Jiawei Han. 2022. Large language models can self-improve. arXiv preprint arXiv:2210.11610.
  29. Aaron Jaech, Adam Kalai, Adam Lerer, Adam Richardson, Ahmed El-Kishky, Aiden Low, Alec Helyar, Aleksander Madry, Alex Beutel, Alex Carney, et al. 2024. Openai o1 system card. arXiv preprint arXiv:2412.16720.
  30. Can Jin, Tong Che, Hongwu Peng, Yiyuan Li, Dimitris Metaxas, and Marco Pavone. 2024. Learning from teaching regularization: Generalizable correlations should be easy to imitate. Advances in Neural Information Processing Systems, 37:966–994. find
  31. Can Jin, Hongwu Peng, Qixin Zhang, Yujin Tang, Dimitris N Metaxas, and Tong Che. 2025. Two heads are better than one: Test-time scaling of multi-agent collaborative reasoning. arXiv preprint arXiv:2504.09772. — [arXiv:2504.09772](https://arxiv.org/abs/2504.09772) find
  32. Takeshi Kojima, Shixiang Shane Gu, Machel Reid, Yutaka Matsuo, and Yusuke Iwasawa. 2022. Large language models are zero-shot reasoners. Advances in neural information processing systems.
  33. Rik Koncel-Kedziorski, Subhro Roy, Aida Amini, Nate Kushman, and Hannaneh Hajishirzi. 2016. MAWPS: A math word problem repository. In Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. find
  34. Woosuk Kwon, Zhuohan Li, Siyuan Zhuang, Ying Sheng, Lianmin Zheng, Cody Hao Yu, Joseph E. Gonzalez, Hao Zhang, and Ion Stoica. 2023. Efficient memory management for large language model serving with pagedattention. In Proceedings of the ACM SIGOPS 29th Symposium on Operating Systems Principles.
  35. Yujia Li, David Choi, Junyoung Chung, Nate Kushman, Julian Schrittwieser, Rémi Leblond, Tom Eccles, James Keeling, Felix Gimeno, Agustin Dal Lago, et al. 2022. Competition-level code generation with alphacode. Science.
  36. Zhaoyi Li, Gangwei Jiang, Hong Xie, Linqi Song, Defu Lian, and Ying Wei. 2024. Understanding and patching compositional reasoning in llms. Findings of the Association for Computational Linguistics. find
  37. Hunter Lightman, Vineet Kosaraju, Yura Burda, Harri Edwards, Bowen Baker, Teddy Lee, Jan Leike, John Schulman, Ilya Sutskever, and Karl Cobbe. 2023. Let’s verify step by step. arXiv preprint arXiv:2305.20050.
  38. Zheyuan Liu, Guangyao Dou, Xiangchi Yuan, Chunhui Zhang, Zhaoxuan Tan, and Meng Jiang. 2025. Modality-aware neuron pruning for unlearning in multimodal large language models. The 63rd Annual Meeting of the Association for Computational Linguistics.
  39. Zimu Lu, Aojun Zhou, Houxing Ren, Ke Wang, Weikang Shi, Junting Pan, Mingjie Zhan, and Hongsheng Li. 2024. Mathgenie: Generating synthetic data with question back-translation for enhancing mathematical reasoning of llms. arXiv preprint arXiv:2402.16352. find
  40. Peijun Qing, Chongyang Gao, Yefan Zhou, Xingjian Diao, Yaoqing Yang, and Soroush Vosoughi. 2024. AlphaLoRA: Assigning LoRA experts based on layer training quality. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing. find
  41. Xinyin Ma, Gongfan Fang, and Xinchao Wang. 2023. Llm-pruner: On the structural pruning of large language models. Advances in neural information processing systems, 36:21702–21720.
  42. Samyam Rajbhandari, Jeff Rasley, Olatunji Ruwase, and Yuxiong He. 2020. Zero: memory optimizations toward training trillion parameter models. In Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis. find
  43. Amil Merchant, Elahe Rahimtoroghi, Ellie Pavlick, and Ian Tenney. 2020. What happens to BERT embeddings during fine-tuning? In Proceedings of the Third BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP.
  44. Shen-yun Miao, Chao-Chun Liang, and Keh-Yih Su. 2020. A diverse corpus for evaluating and developing English math word problem solvers. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics. find
  45. Marius Mosbach, Anna Khokhlova, Michael A. Hedderich, and Dietrich Klakow. 2020. On the interplay between fine-tuning and sentence-level probing for linguistic knowledge in pre-trained transformers. In Proceedings of the Third BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NL… find
  46. OpenAI. 2025. Openai o3-mini. https://openai.com/index/openai-o3-mini/. ↗
  47. URL Vishakh Padmakumar and He He. 2024. Does writing with language models reduce content diversity? In The Twelfth International Conference on Learning Representations. find
  48. Richard Yuanzhe Pang, Weizhe Yuan, He He, Kyunghyun Cho, Sainbayar Sukhbaatar, and Jason E Weston. 2024. Iterative reasoning preference optimization. In The Thirty-eighth Annual Conference on Neural Information Processing Systems.
  49. Arkil Patel, Satwik Bhattamishra, and Navin Goyal. 2021. Are NLP models really able to solve simple math word problems? In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies.
  50. Archiki Prasad, Weizhe Yuan, Richard Yuanzhe Pang, Jing Xu, Maryam Fazel-Zarandi, Mohit Bansal, Sainbayar Sukhbaatar, Jason Weston, and Jane Yu. 2024. Self-consistency preference optimization. arXiv preprint arXiv:2411.04109. find
  51. Zhenting Qi, Mingyuan MA, Jiahang Xu, Li Lyna Zhang, Fan Yang, and Mao Yang. 2025. Mutual reasoning makes smaller LLMs stronger problem-solver. In The Thirteenth International Conference on Learning Representations. find
  52. Mohamed El Amine Seddik, Suei-Wen Chen, Soufiane Hayou, Pierre Youssef, and Merouane Debbah. 2024. How bad is training on synthetic data? a statistical analysis of language model collapse. arXiv preprint arXiv:2404.05090. — [arXiv:2404.05090](https://arxiv.org/abs/2404.05090) find
  53. Dachuan Shi, Yonggan Fu, Xiangchi Yuan, Zhongzhi Yu, Haoran You, Sixu Li, Xin Dong, Jan Kautz, Pavlo Molchanov, and Yingyan Celine Lin. 2025. Lacache: Ladder-shaped kv caching for efficient long-context modeling of large language models. In Forty-second International Conference on Machine Learning. find
  54. Dachuan Shi, Chaofan Tao, Anyi Rao, Zhendong Yang, Chun Yuan, and Jiaqi Wang. 2023. Cross-get: Cross-guided ensemble of tokens for accelerating vision-language transformers. arXiv preprint arXiv:2305.17455. — [arXiv:2305.17455](https://arxiv.org/abs/2305.17455) find
  55. Ilia Shumailov, Zakhar Shumaylov, Yiren Zhao, Yarin Gal, Nicolas Papernot, and Ross Anderson. 2023. The curse of recursion: Training on generated data makes models forget. arXiv preprint arXiv:2305.17493.
  56. David Silver, Thomas Hubert, Julian Schrittwieser, Ioannis Antonoglou, Matthew Lai, Arthur Guez, Marc Lanctot, Laurent Sifre, Dharshan Kumaran, Thore Graepel, et al. 2018. A general reinforcement learning algorithm that masters chess, shogi, and go through self-play. Science, 362(6419):1140–1144. find
  57. Avi Singh, John D Co-Reyes, Rishabh Agarwal, Ankesh Anand, Piyush Patil, Xavier Garcia, Peter J Liu, James Harrison, Jaehoon Lee, Kelvin Xu, et al. 2023. Beyond human data: Scaling self-training for problem-solving with language models. arXiv preprint arXiv:2312.06585.
  58. Charlie Victor Snell, Jaehoon Lee, Kelvin Xu, and Aviral Kumar. 2025. Scaling LLM test-time compute optimally can be more effective than scaling parameters for reasoning. In The Thirteenth International Conference on Learning Representations. find
  59. Ilya Sutskever. 2024. Sequence to sequence learning with neural networks: what a decade. Keynote at NeurIPS. find
  60. Hugo Touvron, Louis Martin, Kevin Stone, Peter Albert, Amjad Almahairi, Yasmine Babaei, Nikolay Bashlykov, Soumya Batra, Prajjwal Bhargava, Shruti Bhosale, et al. 2023. Llama 2: Open foundation and fine-tuned chat models. arXiv preprint arXiv:2307.09288.
  61. Yue Wang, Hung Le, Akhilesh Deepak Gotmare, Nghi DQ Bui, Junnan Li, and Steven CH Hoi. 2023. Codet5+: Open code large language models for code understanding and generation. arXiv preprint arXiv:2305.07922. find
  62. Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Fei Xia, Ed Chi, Quoc V Le, Denny Zhou, et al. 2022. Chain-of-thought prompting elicits reasoning in large language models. Advances in neural information processing systems.
  63. Tianhao Wu, Weizhe Yuan, Olga Golovneva, Jing Xu, Yuandong Tian, Jiantao Jiao, Jason Weston, and Sainbayar Sukhbaatar. 2024. Meta-rewarding language models: Self-improving alignment with llm-as-a-meta-judge. arXiv preprint arXiv:2407.19594.
  64. Weiyi Wu, Xinwen Xu, Chongyang Gao, Xingjian Diao, Siting Li, Lucas A Salas, and Jiang Gui. 2025. Assessing and mitigating medical knowledge drift and conflicts in large language models. arXiv preprint arXiv:2505.07968. find
  65. An Yang and et al. 2024. Qwen2.5 technical report. arXiv preprint arXiv:2412.15115. find
  66. Shunyu Yao, Dian Yu, Jeffrey Zhao, Izhak Shafran, Tom Griffiths, Yuan Cao, and Karthik Narasimhan. 2024. Tree of thoughts: Deliberate problem solving with large language models. Advances in Neural Information Processing Systems. find
  67. Zhifan Ye, Kejing Xia, Yonggan Fu, Xin Dong, Jihoon Hong, Xiangchi Yuan, Shizhe Diao, Jan Kautz, Pavlo Molchanov, and Yingyan Celine Lin. 2025. Longmamba: Enhancing mamba’s long context capabilities via training-free receptive field enlargement. The Thirteenth International Conference on Learning Re… find
  68. Eric Zelikman, Yuhuai Wu, Jesse Mu, and Noah Goodman. 2022. Star: Bootstrapping reasoning with reasoning. Advances in Neural Information Processing Systems.
  69. Chunhui Zhang, Yiren Jian, Zhongyu Ouyang, and Soroush Vosoughi. 2024. Working memory identifies reasoning limits in language models. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing. find
  70. Chunhui Zhang, Zhongyu Ouyang, Kwonjoon Lee, Nakul Agarwal, Sean Dae Houlihan, Soroush Vosoughi, and Shao-Yuan Lo. 2025a. Overcoming multi-step complexity in multimodal theory-of-mind reasoning: A scalable bayesian planner. In Forty-second International Conference on Machine Learning. find
  71. Chunhui Zhang, Zhongyu Ouyang, Xiangchi Yuan, Soroush Vosoughi, et al. 2025b. Growing through experience: Scaling episodic grounding in language models. The 63rd Annual Meeting of the Association for Computational Linguistics.
  72. Huaixiu Steven Zheng, Swaroop Mishra, Xinyun Chen, Heng-Tze Cheng, Ed H. Chi, Quoc V Le, and Denny Zhou. 2024. Take a step back: Evoking reasoning via abstraction in large language models. In The Twelfth International Conference on Learning Representations. find
  73. Wanjun Zhong, Ruixiang Cui, Yiduo Guo, Yaobo Liang, Shuai Lu, Yanlin Wang, Amin Saied, Weizhu Chen, and Nan Duan. 2024. AGIEval: A human-centric benchmark for evaluating foundation models. In Findings of the Association for Computational Linguistics: NAACL 2024.
  74. Yichu Zhou and Vivek Srikumar. 2021. A closer look at how fine-tuning changes bert. arXiv preprint arXiv:2106.14282. find
  75. Le Yu, Bowen Yu, Haiyang Yu, Fei Huang, and Yongbin Li. 2024a. Language models are super mario: Absorbing abilities from homologous models as a free lunch. In International Conference on Machine Learning.
  76. Longhui Yu, Weisen Jiang, Han Shi, Jincheng YU, Zhengying Liu, Yu Zhang, James Kwok, Zhenguo Li, Adrian Weller, and Weiyang Liu. 2024b. Metamath: Bootstrap your own mathematical questions for large language models. In ICLR.
  77. Weizhe Yuan, Richard Yuanzhe Pang, Kyunghyun Cho, Sainbayar Sukhbaatar, Jing Xu, and Jason Weston. 2024. Self-rewarding language models. arXiv preprint arXiv:2401.10020. find
  78. Zheng Yuan, Hongyi Yuan, Chengpeng Li, Guanting Dong, Keming Lu, Chuanqi Tan, Chang Zhou, and Jingren Zhou. 2023. Scaling relationship on learning mathematical reasoning with large language models. arXiv preprint arXiv:2308.01825.

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/yuan-2025-superficial-self-improved-reasoners-benefit/