Elamur Research⌕
☰
home / papers / 2025 / Case2Code: Scalable Synthetic Data for Code Generation

Case2Code: Scalable Synthetic Data for Code Generation

Paper · 2025 · Yunfan Shao, Linyang Li, Yichuan Ma et al.

Summary

Code LLMs are often improved with synthetic instruction data, but this data usually comes from a strong teacher model. That makes it expensive and limits it to what the teacher can do. Training data also rarely contains inductive tasks, where a program must be inferred from examples of its behaviour. The authors propose Case2Code, a task in which a model reconstructs an unknown program from several input-output cases. They built an automatic data pipeline:

  1. Extract self-contained Python functions from The Stack.
  2. Have a small "writer" LLM propose example inputs.
  3. Run the functions in an interpreter to get the true outputs.

Because outputs come from execution, data correctness does not depend on the writer model's skill. Untrained models, including strong ones, struggle with this kind of induction. Fine-tuning on Case2Code improves both in-distribution induction and general coding benchmarks. This suggests a cheap, scalable source of training signal that supports weak-to-strong learning. It may stay useful once human-written data runs out.

Key points

  • Functions are kept only if they pass a syntax check, take at least one argument and return a value, and avoid third-party packages and external I/O. This allows a simple executor with no sandbox.
  • LLaMA3-70B, GPT-3.5 and GPT-4 all find case-to-code induction hard without dedicated training.
  • Training on Case2Code also helps on general benchmarks: HumanEval, HumanEval+ and MBPP.
  • Writer comparison:
    • InternLM2-7B: about 1.3M samples, 1600 tokens/s.
    • LLaMA3-70B: about 700K samples, 720 tokens/s, roughly 4.5× the total cost.
    • LLaMA3-70B inputs are higher quality per sample, but the 7B writer stays efficient.
  • Data generated by InternLM2-7B improved InternLM2-1.8B, 7B and 20B. The 20B model is nearly three times the size of the generator.

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

🎓Fudan UniversityShanghai · China167 works · 20,987 citations🎓Chinese University of Hong KongHong Kong · China126 works · 14,073 citations🔬Shanghai Artificial Intelligence LaboratoryShanghai · China117 works · 11,281 citations

 Fields, topics and concepts

📚 Shelves 6The 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 worksInformation Systems · 1,437 worksComputer Science Applications · 102 works

🧭 Topics 1Research subjects and languages the work is about.Code & programming languages · 129 works

Cite

Cite as Yunfan Shao, Linyang Li, Yichuan Ma, Peiji Li, Demin Song, Qinyuan Cheng, Shimin Li, Xiaonan Li, Pengyu Wang, Qipeng Guo, Hang Yan, Xipeng Qiu, Xuanjing Huang and Dahua Lin. 2025. Case2Code: Scalable Synthetic Data for Code Generation. Proceedings of the 31st International Conference on Computational Linguistics.

The paper itself, not this page. BibTeX from the library's record of the work.

@inproceedings{shao2025case2code,
  title = {{Case2Code: Scalable Synthetic Data for Code Generation}},
  author = {Yunfan Shao, Linyang Li, Yichuan Ma, Peiji Li, Demin Song, Qinyuan Cheng, Shimin Li, Xiaonan Li, Pengyu Wang, Qipeng Guo, Hang Yan, Xipeng Qiu, Xuanjing Huang and Dahua Lin},
  booktitle = {Proceedings of the 31st International Conference on Computational Linguistics},
  year = {2025},
  url = {https://aclanthology.org/2025.coling-main.733/}
}

References

35 references, 17 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. AI@Meta. 2024. Llama 3 model card. find
  2. Matej Balog, Alexander L Gaunt, Marc Brockschmidt, Sebastian Nowozin, and Daniel Tarlow. 2016. Deepcoder: Learning to write programs. arXiv preprint arXiv:1611.01989. find
  3. Lukas Berglund, Meg Tong, Max Kaufmann, Mikita Balesni, Asa Cooper Stickland, Tomasz Korbak, and Owain Evans. 2023. The reversal curse: Llms trained on "a is b" fail to learn "b is a". CoRR, abs/2309.12288. find
  4. Zheng Cai, Maosong Cao, Haojiong Chen, Kai Chen, Keyu Chen, Xin Chen, Xun Chen, Zehui Chen, Zhi Chen, Pei Chu, Xiaoyi Dong, Haodong Duan, Qi Fan, Zhaoye Fei, Yang Gao, Jiaye Ge, Chenya Gu, Yuzhe Gu, Tao Gui, Aijia Guo, Qipeng Guo, Conghui He, Yingfan Hu, Ting Huang, Tao Jiang, Penglong Jiao, Zhenjia… find
  5. Shang, Yunfan Shao, Demin Song, Zifan Song, Zhihao Sui, Peng Sun, Yu Sun, Huanze Tang, Bin Wang, Guoteng Wang, Jiaqi Wang, Jiayu Wang, Rui Wang, Yudong Wang, Ziyi Wang, Xingjian Wei, Qizhen Weng, Fan Wu, Yingtong Xiong, and et al. 2024. Internlm2 technical report. CoRR, abs/2403.17297.
  6. Sahil Chaudhary. 2023. Code alpaca: An instruction-following llama model for code generation. https://github.com/sahil280114/codealpaca. ↗
  7. Guoxin Chen, Minpeng Liao, Chengxi Li, and Kai Fan. 2024. Alphamath almost zero: process supervision without process. ArXiv, abs/2405.03553.
  8. Jacob Devlin, Jonathan Uesato, Surya Bhupatiraju, Rishabh Singh, Abdel-rahman Mohamed, and Pushmeet Kohli. 2017. Robustfill: Neural program learning under noisy I/O. In Proceedings of the 34th International Conference on Machine Learning, ICML 2017, Sydney, NSW, Australia, 6-11 August 2017, volume 7… find
  9. Kevin Ellis, Catherine Wong, Maxwell Nye, Mathias Sablé-Meyer, Lucas Morales, Luke Hewitt, Luc Cary, Armando Solar-Lezama, and Joshua B Tenenbaum. 2021. Dreamcoder: Bootstrapping inductive program synthesis with wake-sleep library learning. In Proceedings of the 42nd acm sigplan international confer… find
  10. Leo Gao, Stella Biderman, Sid Black, Laurence Golding, Travis Hoppe, Charles Foster, Jason Phang, Horace He, Anish Thite, Noa Nabeshima, Shawn Presser, and Connor Leahy. 2021. The pile: An 800gb dataset of diverse text for language modeling. CoRR, abs/2101.00027. find
  11. Sirui Hong, Mingchen Zhuge, Jonathan Chen, Xiawu Zheng, Yuheng Cheng, Jinlin Wang, Ceyao Zhang, Zili Wang, Steven Ka Shing Yau, Zijuan Lin, Liyang Zhou, Chenyu Ran, Lingfeng Xiao, Chenglin Wu, and Jürgen Schmidhuber. 2024. Metagpt: Meta programming for A multi-agent collaborative framework. In The T…
  12. Jiaxin Huang, Shixiang Gu, Le Hou, Yuexin Wu, Xuezhi Wang, Hongkun Yu, and Jiawei Han. 2023. Large language models can self-improve. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, EMNLP 2023, Singapore, December 6-10, 2023, pages 1051–1068. Association for…
  13. Denis Kocetkov, Raymond Li, Loubna Ben Allal, Jia Li, Chenghao Mou, Carlos Muñoz Ferrandis, Yacine Jernite, Margaret Mitchell, Sean Hughes, Thomas Wolf, Dzmitry Bahdanau, Leandro von Werra, and Harm de Vries. 2022. The stack: 3 TB of permissively licensed source code. CoRR, abs/2211.15533. find
  14. Wen-Ding Li and Kevin Ellis. 2024. Is programming by example solved by llms? CoRR, abs/2406.08316. — [arXiv:2406.08316](https://arxiv.org/abs/2406.08316) find
  15. Ziyang Luo, Can Xu, Pu Zhao, Qingfeng Sun, Xiubo Geng, Wenxiang Hu, Chongyang Tao, Jing Ma, Qingwei Lin, and Daxin Jiang. 2023. Wizardcoder: Empowering code large language models with evol-instruct. CoRR, abs/2306.08568. find
  16. Arindam Mitra, Hamed Khanpour, Corby Rosset, and Ahmed Awadallah. 2024. Orca-math: Unlocking the potential of slms in grade school math. CoRR, abs/2402.14830.
  17. OpenAI. 2023. GPT-4 technical report. abs/2303.08774.
  18. CoRR, Long Ouyang, Jeffrey Wu, Xu Jiang, Diogo Almeida, Carroll L. Wainwright, Pamela Mishkin, Chong Zhang, Sandhini Agarwal, Katarina Slama, Alex Ray, John Schulman, Jacob Hilton, Fraser Kelton, Luke Miller, Maddie Simens, Amanda Askell, Peter Welinder, Paul F. Christiano, Jan Leike, and Ryan Lowe.… — Training language models to follow instructions with human feedback
  19. Baptiste Rozière, Jonas Gehring, Fabian Gloeckle, Sten Sootla, Itai Gat, Xiaoqing Ellen Tan, Yossi Adi, Jingyu Liu, Tal Remez, Jérémy Rapin, Artyom Kozhevnikov, Ivan Evtimov, Joanna Bitton, Manish Bhatt, Cristian Canton-Ferrer, Aaron Grattafiori, Wenhan Xiong, Alexandre Défossez, Jade Copet, Faisal… — Code Llama: Open Foundation Models for Code
  20. Kensen Shi, Joey Hong, Manzil Zaheer, Pengcheng Yin, and Charles Sutton. 2023. Exedec: Execution decomposition for compositional generalization in neural program synthesis. CoRR, abs/2307.13883. find
  21. David Silver, Aja Huang, Chris J. Maddison, Arthur Guez, Laurent Sifre, George van den Driessche, Julian Schrittwieser, Ioannis Antonoglou, Vedavyas Panneershelvam, Marc Lanctot, Sander Dieleman, Dominik Grewe, John Nham, Nal Kalchbrenner, Ilya Sutskever, Timothy P. Lillicrap, Madeleine Leach, Koray… find
  22. Qiushi Sun, Zhirui Chen, Fangzhi Xu, Kanzhi Cheng, Chang Ma, Zhangyue Yin, Jianing Wang, Chengcheng Han, Renyu Zhu, Shuai Yuan, Qipeng Guo, Xipeng Qiu, Pengcheng Yin, Xiaoli Li, Fei Yuan, Lingpeng Kong, Xiang Li, and Zhiyong Wu. 2024. A survey of neural code intelligence: Paradigms, advances and bey… — [arXiv:2403.14734](https://arxiv.org/abs/2403.14734) find
  23. Rohan Taori, Ishaan Gulrajani, Tianyi Zhang, Yann Dubois, Xuechen Li, Carlos Guestrin, Percy Liang, and Tatsunori B. Hashimoto. 2023. Stanford alpaca: An instruction-following llama model. https://github.com/tatsu-lab/stanford_alpaca. ↗
  24. Hugo Touvron, Louis Martin, Kevin Stone, Peter Albert, Amjad Almahairi, Yasmine Babaei, Nikolay Bashlykov, Soumya Batra, Prajjwal Bhargava, Shruti Bhosale, Dan Bikel, Lukas Blecher, Cristian Canton-Ferrer, Moya Chen, Guillem Cucurull, David Esiobu, Jude Fernandes, Jeremy Fu, Wenyin Fu, Brian Fuller,… — Llama 2: Open Foundation and Fine-Tuned Chat Models
  25. Peiyi Wang, Lei Li, Zhihong Shao, R. X. Xu, Damai Dai, Yifei Li, Deli Chen, Y. Wu, and Zhifang Sui. 2023a. Math-shepherd: Verify and reinforce llms step-by-step without human annotations. CoRR, abs/2312.08935.
  26. Xuezhi Wang, Jason Wei, Dale Schuurmans, Quoc Le, Ed Chi, Sharan Narang, Aakanksha Chowdhery, and Denny Zhou. 2022. Self-consistency improves chain of thought reasoning in language models. arXiv preprint arXiv:2203.11171.
  27. Yizhong Wang, Yeganeh Kordi, Swaroop Mishra, Alisa Liu, Noah A. Smith, Daniel Khashabi, and Hannaneh Hajishirzi. 2023b. Self-instruct: Aligning language models with self-generated instructions. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long… find
  28. Jason Wei, Maarten Bosma, Vincent Y. Zhao, Kelvin Guu, Adams Wei Yu, Brian Lester, Nan Du, Andrew M. Dai, and Quoc V. Le. 2022. Finetuned language models are zero-shot learners. In The Tenth International Conference on Learning Representations, ICLR 2022, Virtual Event, April 25-29, 2022. OpenReview…
  29. Jiarong Wu, Lili Wei, Yanyan Jiang, Shing-Chi Cheung, Luyao Ren, and Chang Xu. 2024. Programming by example made easy. ACM Trans. Softw. Eng. Methodol., 33(1):4:1–4:36. find
  30. Zonglin Yang, Li Dong, Xinya Du, Hao Cheng, Erik Cambria, Xiaodong Liu, Jianfeng Gao, and Furu Wei. 2022. Language models as inductive reasoners. arXiv preprint arXiv:2212.10923.
  31. Shunyu Yao, Dian Yu, Jeffrey Zhao, Izhak Shafran, Thomas L. Griffiths, Yuan Cao, and Karthik Narasimhan. 2023. Tree of thoughts: Deliberate problem solving with large language models. ArXiv, abs/2305.10601. find
  32. Longhui Yu, Weisen Jiang, Han Shi, Jincheng Yu, Zhengying Liu, Yu Zhang, James T Kwok, Zhenguo Li, Adrian Weller, and Weiyang Liu. 2023a. Metamath: Bootstrap your own mathematical questions for large language models. arXiv preprint arXiv:2309.12284.
  33. Yue Yu, Yuchen Zhuang, Jieyu Zhang, Yu Meng, Alexander J. Ratner, Ranjay Krishna, Jiaming Shen, and Chao Zhang. 2023b. Large language model as attributed training data generator: A tale of diversity and bias. In Advances in Neural Information Processing Systems 36: Annual Conference on Neural Inform…
  34. Zheng Yuan, Hongyi Yuan, Chengpeng Li, Guanting Dong, Chuanqi Tan, and Chang Zhou. 2023. Scaling relationship on learning mathematical reasoning with large language models. CoRR, abs/2308.01825.
  35. Terry Yue Zhuo, Minh Chien Vu, Jenny Chim, Han Hu, Wenhao Yu, Ratnadira Widyasari, Imam Nur Bani Yusuf, Haolan Zhan, Junda He, Indraneil Paul, Simon Brunner, Chen Gong, Thong Hoang, Armel Randy Zebaze, Xiaoheng Hong, Wen-Ding Li, Jean Kaddour, Ming Xu, Zhihan Zhang, Prateek Yadav, Naman Jain, Alex G… — BigCodeBench: Benchmarking Code Generation with Diverse Function Calls and Complex Instructions

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/shao-2025-case2code-scalable-synthetic-data-code/