Elamur Research
☰
home / papers / 2025 / OpenThoughts: Data Recipes for Reasoning Models

OpenThoughts: Data Recipes for Reasoning Models

Written for the Research library (claude-opus-5-5) from the paper text. Original — arxiv.org/pdf/2506.04178

Summary

The strongest reasoning models are usually distilled from training data that is not public, so researchers know little about which data choices actually matter. The OpenThoughts project addresses this by building fully open supervised fine-tuning (SFT) datasets for reasoning in math, code and science. Earlier releases came first: OpenThoughts-114K, then OpenThoughts2-1M, whose 32B model matched DeepSeek-R1-Distill-32B. The authors then ran more than 1,000 controlled ablations covering each stage of the pipeline: question sourcing, mixing of sources, question filtering, deduplication and repeated answer sampling, answer filtering, and choice of teacher. They scaled the best recipe into OpenThoughts3-1.2M, using QwQ-32B to generate the answers, and fine-tuned Qwen2.5-7B-Instruct on it to get OpenThinker3-7B. Among 7B models trained on open data, it is the strongest they report. Averaged over 12 benchmarks, it beats DeepSeek-R1-Distill-Qwen-7B by about 12 points. All datasets and models are published at openthoughts.ai. This gives a reproducible baseline and an evidence-based guide for curating reasoning data.

Key points

  • OpenThinker3-7B scores 53% on AIME 2025, 51% on LiveCodeBench 06/24–01/25 and 54% on GPQA Diamond. That is 15.3, 17.2 and 20.5 points higher than DeepSeek-R1-Distill-Qwen-7B, and 2.1 points higher on average than Nemotron-Nano-8B.
  • The best teacher is not the model with the best benchmark scores: QwQ-32B produced better training data than DeepSeek-R1, and Phi-4-reasoning-plus was clearly worse, in all three domains.
  • Generating up to 16 answers for each question expands a data source cheaply and helps performance. Exact deduplication or no deduplication worked better than fuzzy deduplication.
  • None of the answer-filtering methods gave a clear gain, including GPT-based verification, majority consensus and fastText. Training on unfiltered answers did as well or better.
  • Taking questions from only the top 1–2 sources beat mixing 8–16 sources for diversity. Filtering questions by LLM-rated difficulty or by response length beat embedding-based and fastText filters.

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

Fields: Mathematics & statistics

Cited by

112 papers of our library cite it (of the 140673 whose bibliographies we read); 2 of them have a page here, 110 not yet.

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.48550/arxiv.2506.04178 · arXiv 2506.04178

@misc{guha2025openthoughts,
  title = {{OpenThoughts: Data Recipes for Reasoning Models}},
  author = {Guha, Etash and Marten, Ryan and Keh, Sedrick and Raoof, Negin and Smyrnis, Georgios and Bansal, Hritik and Nezhurina, Marianna and Mercat, Jean and Vu, Trung and Sprague, Zayne and Suvarna, Ashima and Feuer, Benjamin and Chen, Liangyu and Khan, Zaid and Frankel, Eric and Grover, Sachin and Choi, Caroline and Muennighoff, Niklas and Su, Shiye and Zhao, Wanjia and Yang, John and Pimpalgaonkar, Shreyas and Sharma, Kartik and Ji, Charlie Cheng-Jie and Deng, Yichuan and Pratt, Sarah and Ramanujan, Vivek and Saad-Falcon, Jon and Li, Jeffrey and Dave, Achal and Albalak, Alon and Arora, Kushal and Wulfe, Blake and Hegde, Chinmay and Durrett, Greg and Oh, Sewoong and Bansal, Mohit and Gabriel, Saadia and Grover, Aditya and Chang, Kai-Wei and Shankar, Vaishaal and Gokaslan, Aaron and Merrill, Mike A. and Hashimoto, Tatsunori and Choi, Yejin and Jitsev, Jenia and Heckel, Reinhard and Sathiamoorthy, Maheswaran and Dimakis, Alexandros G. and Schmidt, Ludwig},
  year = {2025},
  doi = {10.48550/arxiv.2506.04178},
  eprint = {2506.04178},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.48550/arxiv.2506.04178}
}

References

66 references, 21 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Marah Abdin, Jyoti Aneja, Hany Awadalla, Ahmed Awadallah, Ammar Ahmad Awan, Nguyen Bach, Amit Bahree, Arash Bakhtiari, Jianmin Bao, Harkirat Behl, et al. Phi-3 technical report: A highly capable language model locally on your phone. arXiv preprint arXiv:2404.14219, 2024. find our card
  2. Marah Abdin, Sahaj Agarwal, Ahmed Awadallah, Vidhisha Balachandran, Harkirat Behl, Lingjiao Chen, Gustavo de Rosa, Suriya Gunasekar, Mojan Javaheripi, Neel Joshi, et al. Phi-4-reasoning technical report. arXiv preprint arXiv:2504.21318, 2025. find
  3. Bo Adler, Niket Agarwal, Ashwath Aithal, Dong H Anh, Pallab Bhattacharya, Annika Brundyn, Jared Casper, Bryan Catanzaro, Sharon Clay, Jonathan Cohen, et al. Nemotron-4 340b technical report. arXiv preprint arXiv:2406.11704, 2024. find
  4. Wasi Uddin Ahmad, Sean Narenthiran, Somshubra Majumdar, Aleksander Ficek, Siddhartha Jain, Jocelyn Huang, Vahid Noroozi, and Boris Ginsburg. Opencodereasoning: Advancing data distillation for competitive coding. arXiv preprint arXiv:2504.01943, 2025. URL https://arxiv.org/abs/2504.01943. ↗
  5. Daman Arora, Himanshu Singh, and Mausam. Have llms advanced enough? a challenging problem solving benchmark for large language models. In Houda Bouamor, Juan Pino, and Kalika Bali (eds.), Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pp. 7527–7543, Singapore… find
  6. Mislav Balunović, Jasper Dekoninck, Ivo Petrov, Nikola Jovanović, and Martin Vechev. Matharena: Evaluating llms on uncontaminated math competitions, February 2025. URL https://math arena.ai/. find
  7. Akhiad Bercovich, Itay Levy, Izik Golan, Mohammad Dabbah, Ran El-Yaniv, Omri Puny, Ido Galil, Zach Moshe, Tomer Ronen, Najeeb Nabwani, et al. Llama-nemotron: Efficient reasoning models. arXiv preprint arXiv:2505.00949, 2025. find
  8. Bespoke-Labs. Bespoke-stratos: The unreasonable effectiveness of reasoning distillation. www.bespokelabs.ai/blog/bespoke-stratos-the-unreasonable-effectiveness-of-reasoning-distillation, 2025. Accessed: 2025-01-22. find
  9. Sahil Chaudhary. Code alpaca: An instruction-following llama model for code generation. https: //github.com/sahil280114/codealpaca, 2023. find
  10. Xingyu Chen, Jiahao Xu, Tian Liang, Zhiwei He, Jianhui Pang, Dian Yu, Linfeng Song, Qiuzhi Liu, Mengfei Zhou, Zhuosheng Zhang, et al. Do not think that much for 2+ 3=? on the overthinking of o1-like llms. arXiv preprint arXiv:2412.21187, 2024. find our card
  11. Yang Chen, Zhuolin Yang, Zihan Liu, Chankyu Lee, Mohammad Shoeybi, Bryan Catanzaro, and Wei Ping. Acereason-nemotron: Advancing math and code reasoning through reinforcement learning. arXiv preprint, 2025. find
  12. Yihe Deng, Hritik Bansal, Fan Yin, Nanyun Peng, Wei Wang, and Kai-Wei Chang. Openvlthinker: An early exploration to complex vision-language reasoning via iterative self-improvement. arXiv preprint arXiv:2503.17352, 2025. — [arXiv:2503.17352](https://arxiv.org/abs/2503.17352) find
  13. Ricardo Dominguez-Olmedo, Vedant Nanda, Rediet Abebe, Stefan Bechtold, Christoph Engel, Jens Frankenreiter, Krishna Gummadi, Moritz Hardt, and Michael Livermore. Lawma: The power of specialization for legal tasks. In International Conference on Learning Representations (ICLR), 2025. find
  14. Hugging Face. Open r1: A fully open reproduction of deepseek-r1, January 2025. URL https: //github.com/huggingface/open-r1. find
  15. Alex Fang, Hadi Pouransari, Matt Jordan, Alexander Toshev, Vaishaal Shankar, Ludwig Schmidt, and Tom Gunter. Datasets, documents, and repetitions: The practicalities of unequal data quality, 2025. URL https://arxiv.org/abs/2503.07879. — [arXiv:2503.07879](https://arxiv.org/abs/2503.07879) ↗
  16. 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. The Pile: An 800gb dataset of diverse text for language modeling. arXiv preprint arXiv:2101.00027, 2020. find
  17. Gemini-Team, Rohan Anil, Sebastian Borgeaud, Jean-Baptiste Alayrac, Jiahui Yu, Radu Soricut, Johan Schalkwyk, Andrew M Dai, Anja Hauth, Katie Millican, et al. Gemini: a family of highly capable multimodal models. arXiv preprint arXiv:2312.11805, 2023. find our card
  18. Daya Guo, Dejian Yang, Haowei Zhang, Junxiao Song, Ruoyu Zhang, Runxin Xu, Qihao Zhu, Shirong Ma, Peiyi Wang, Xiao Bi, et al. Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning. arXiv preprint arXiv:2501.12948, 2025. find our card
  19. Jujie He, Jiacai Liu, Chris Yuhao Liu, Rui Yan, Chaojie Wang, Peng Cheng, Xiaoyu Zhang, Fuxiang Zhang, Jiacheng Xu, Wei Shen, Siyuan Li, Liang Zeng, Tianwen Wei, Cheng Cheng, Bo An, Yang Liu, and Yahui Zhou. Skywork open reasoner 1 technical report. arXiv preprint arXiv:2505.22312, 2025a. find
  20. Zhiwei He, Tian Liang, Jiahao Xu, Qiuzhi Liu, Xingyu Chen, Yue Wang, Linfeng Song, Dian Yu, Zhenwen Liang, Wenxuan Wang, Zhuosheng Zhang, Rui Wang, Zhaopeng Tu, Haitao Mi, and Dong Yu. Deepmath-103k: A large-scale, challenging, decontaminated, and verifiable mathematical dataset for advancing reason… find
  21. Dan Hendrycks, Steven Basart, Saurav Kadavath, Mantas Mazeika, Akul Arora, Ethan Guo, Collin Burns, Samir Puranik, Horace He, Dawn Song, and Jacob Steinhardt. Measuring coding challenge competence with apps. NeurIPS, 2021a. find our card
  22. Dan Hendrycks, Collin Burns, Saurav Kadavath, Akul Arora, Steven Basart, Eric Tang, Dawn Song, and Jacob Steinhardt. Measuring mathematical problem solving with the math dataset. NeurIPS, 2021b. find our card
  23. Naman Jain, King Han, Alex Gu, Wen-Ding Li, Fanjia Yan, Tianjun Zhang, Sida Wang, Armando Solar-Lezama, Koushik Sen, and Ion Stoica. Livecodebench: Holistic and contamination free evaluation of large language models for code. arXiv preprint arXiv:2403.07974, 2024. find our card
  24. Armand Joulin, Edouard Grave, Piotr Bojanowski, and Tomas Mikolov. Bag of tricks for efficient text classification. In Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics: Volume 2, Short Papers, pp. 427–431. Association for Computational Lingu… find
  25. Nathan Lambert, Jacob Morrison, Valentina Pyatkin, Shengyi Huang, Hamish Ivison, Faeze Brahman, Lester James V Miranda, Alisa Liu, Nouha Dziri, Shane Lyu, et al. Tulu 3: Pushing frontiers in open language model post-training. arXiv preprint arXiv:2411.15124, 2024. find our card
  26. Katherine Lee, Daphne Ippolito, Andrew Nystrom, Chiyuan Zhang, Douglas Eck, Chris Callison-Burch, and Nicholas Carlini. Deduplicating training data makes language models better. In Smaranda Muresan, Preslav Nakov, and Aline Villavicencio (eds.), Proceedings of the 60th Annual Meeting of the Associat… find
  27. Guohao Li, Hasan Abed Al Kader Hammoud, Hani Itani, Dmitrii Khizbullin, and Bernard Ghanem. Camel: Communicative agents for "mind" exploration of large scale language model society, 2023a. find
  28. Jeffrey Li, Alex Fang, Georgios Smyrnis, Maor Ivgi, Matt Jordan, Samir Gadre, Hritik Bansal, Etash Guha, Sedrick Keh, Kushal Arora, Saurabh Garg, Rui Xin, Niklas Muennighoff, Reinhard Heckel, Jean Mercat, Mayee Chen, Suchin Gururangan, Mitchell Wortsman, Alon Albalak, Yonatan Bitton, Marianna Nezhur… find our card
  29. Jia LI, Edward Beeching, Lewis Tunstall, Ben Lipkin, Roman Soletskyi, Shengyi Costa Huang, Kashif Rasul, Longhui Yu, Albert Jiang, Ziju Shen, Zihan Qin, Bin Dong, Li Zhou, Yann Fleureau, Guillaume Lample, and Stanislas Polu. Numinamath. [https://huggingface.co/AI-M O/NuminaMath-1.5](https://github.c… find
  30. Rongao Li, Jie Fu, Bo-Wen Zhang, Tao Huang, Zhihong Sun, Chen Lyu, Guang Liu, Zhi Jin, and Ge Li. Taco: Topics in algorithmic code generation dataset. arXiv preprint arXiv:2312.14852, 2023b. find
  31. Yujia Li, David Choi, Junyoung Chung, Nate Kushman, Julian Schrittwieser, Rémi Leblond, Tom Eccles, James Keeling, Felix Gimeno, Agustin Dal Lago, Thomas Hubert, Peter Choy, Cyprien de Masson d’Autume, Igor Babuschkin, Xinyun Chen, Po-Sen Huang, Johannes Welbl, Sven Gowal, Alexey Cherepanov, James M… find our card
  32. Zhenghao Lin, Zhibin Gou, Yeyun Gong, Xiao Liu, Yelong Shen, Ruochen Xu, Chen Lin, Yujiu Yang, Jian Jiao, Nan Duan, et al. Rho-1: Not all tokens are what you need. arXiv preprint arXiv:2404.07965, 2024. find
  33. Zhengzhong Liu, Bowen Tan, Hongyi Wang, Willie Neiswanger, Tianhua Tao, Haonan Li, Fajri Koto, Yuqi Wang, Suqi Sun, Omkar Pangarkar, Richard Fan, Yi Gu, Victor Miller, Liqun Ma, Liping Tang, Nikhil Ranjan, Yonghao Zhuang, Guowei He, Renxi Wang, Mingkai Deng, Robin Algayres, Yuanzhi Li, Zhiqiang Shen… find
  34. MAA. Amc 2023 problems, 2023. URL https://artofproblemsolving.com/wiki/i ndex.php/2023\_AMC\_12A\_Problems. Accessed: 2025-05-11. ↗
  35. MAA. Aime 2024 problems, 2024. URL https://artofproblemsolving.com/wiki/i ndex.php/2024\_AIME\_I\_Problems. Accessed: 2025-05-11. ↗
  36. MAA. Aime 2025 problems, 2025. URL https://artofproblemsolving.com/wiki/i ndex.php/2025\_AIME\_I\_Problems. Accessed: 2025-05-11. ↗
  37. Mantas Mazeika, Long Phan, Xuwang Yin, Andy Zou, Zifan Wang, Norman Mu, Elham Sakhaee, Nathaniel Li, Steven Basart, Bo Li, et al. Harmbench: A standardized evaluation framework for automated red teaming and robust refusal. arXiv preprint arXiv:2402.04249, 2024. find our card
  38. Ivan Moshkov, Darragh Hanley, Ivan Sorokin, Shubham Toshniwal, Christof Henkel, Benedikt Schifferer, Wei Du, and Igor Gitman. Aimo-2 winning solution: Building state-of-the-art mathematical reasoning models with openmathreasoning dataset. arXiv preprint arXiv:2504.16891, 2025. find
  39. Niklas Muennighoff, Zitong Yang, Weijia Shi, Xiang Lisa Li, Li Fei-Fei, Hannaneh Hajishirzi, Luke Zettlemoyer, Percy Liang, Emmanuel Candès, and Tatsunori Hashimoto. s1: Simple test-time scaling, 2025. URL https://arxiv.org/abs/2501.19393. ↗ our card
  40. Marianna Nezhurina, Lucia Cipolina-Kun, Mehdi Cherti, and Jenia Jitsev. Alice in wonderland: Simple tasks showing complete reasoning breakdown in state-of-the-art large language models. arXiv preprint arXiv:2406.02061, 2024. find
  41. NovaSky-Team. Think less, achieve more: Cut reasoning costs by 50 https://novasky-ai.github.io/posts/reduce-overthinking, 2025a. Accessed: 2025-01-23. ↗
  42. NovaSky-Team. Sky-t1: Train your own o1 preview model within 450. ai.github.io/posts/sky-t1, 2025b. Accessed: 2025-01-09. https://novasky- find
  43. OpenAI. Openai o3 and o4-mini system card, 2024. URL https://cdn.openai.com/pdf /2221c875-02dc-4789-800b-e7758f3722c1/o3-and-o4-mini-system-car d.pdf. Accessed: 2025-05-11. ↗
  44. OpenThoughts-Team. Open Thoughts. https://open-thoughts.ai, February 2025. ↗
  45. Guilherme Penedo, Hynek Kydlíček, Loubna Ben allal, Anton Lozhkov, Margaret Mitchell, Colin Raffel, Leandro Von Werra, and Thomas Wolf. The fineweb datasets: Decanting the web for the finest text data at scale. In The Thirty-eight Conference on Neural Information Processing Systems Datasets and Ben… find our card
  46. Guilherme Penedo, Hynek Kydlíček, Loubna Ben allal, Anton Lozhkov, Margaret Mitchell, Colin Raffel, Leandro Von Werra, and Thomas Wolf. The fineweb datasets: Decanting the web for the finest text data at scale. In The Thirty-eight Conference on Neural Information Processing Systems Datasets and Ben… find our card
  47. Guilherme Penedo, Anton Lozhkov, Hynek Kydlíček, Loubna Ben Allal, Edward Beeching, Agustín Piqueres Lajarín, Quentin Gallouédec, Nathan Habib, Lewis Tunstall, and Leandro von Werra. Codeforces. https://huggingface.co/datasets/open-r1/codeforces, 2025. ↗
  48. Long Phan, Alice Gatti, Ziwen Han, Nathaniel Li, Josephina Hu, Hugh Zhang, Chen Bo Calvin Zhang, Mohamed Shaaban, John Ling, and Sean Shi et. al. Humanity’s last exam, 2025. URL https://arxiv.org/abs/2501.14249. ↗ our card
  49. Xiangyu Qi, Yi Zeng, Tinghao Xie, Pin-Yu Chen, Ruoxi Jia, Prateek Mittal, and Peter Henderson. Fine-tuning aligned language models compromises safety, even when users do not intend to!, 2023. URL https://arxiv.org/abs/2310.03693. ↗ our card
  50. Shanghaoran Quan, Jiaxi Yang, Bowen Yu, Bo Zheng, Dayiheng Liu, An Yang, Xuancheng Ren, Bofei Gao, Yibo Miao, Yunlong Feng, et al. Codeelo: Benchmarking competition-level code generation of llms with human-comparable elo ratings. arXiv preprint arXiv:2501.01257, 2025. find
  51. Qwen-Team. Qwq-32b: Embracing the power of reinforcement learning, March 2025. URL https://qwenlm.github.io/blog/qwq-32b/. ↗
  52. Qwen2.5-Team. Qwen2.5 technical report. arXiv preprint arXiv:2412.15115, 2024. find
  53. Negin Raoof, Etash Kumar Guha, Ryan Marten, Jean Mercat, Eric Frankel, Sedrick Keh, Hritik Bansal, Georgios Smyrnis, Marianna Nezhurina, Trung Vu, Zayne Rea Sprague, Mike A Merrill, Liangyu Chen, Caroline Choi, Zaid Khan, Sachin Grover, Benjamin Feuer, Ashima Suvarna, Shiye Su, Wanjia Zhao, Kartik S… find
  54. David Rein, Betty Li Hou, Asa Cooper Stickland, Jackson Petty, Richard Yuanzhe Pang, Julien Dirani, Julian Michael, and Samuel R. Bowman. GPQA: A graduate-level google-proof q&a benchmark. In First Conference on Language Modeling, 2024. URL https://openreview.net/for um?id=Ti67584b98. ↗ our card
  55. Paul Röttger, Hannah Rose Kirk, Bertie Vidgen, Giuseppe Attanasio, Federico Bianchi, and Dirk Hovy. Xstest: A test suite for identifying exaggerated safety behaviours in large language models. arXiv preprint arXiv:2308.01263, 2023. find our card
  56. Kashun Shum, Yuzhen Huang, Hongjian Zou, Ding Qi, Yixuan Liao, Xiaoxin Chen, Qian Liu, and Junxian He. Predictive data selection: The data that predicts is the data that teaches. arXiv preprint arXiv:2503.00808, 2025. — [arXiv:2503.00808](https://arxiv.org/abs/2503.00808) find
  57. Luca Soldaini, Rodney Kinney, Akshita Bhagia, Dustin Schwenk, David Atkinson, Russell Authur, Ben Bogin, Khyathi Chandu, Jennifer Dumas, Yanai Elazar, Valentin Hofmann, Ananya Jha, Sachin Kumar, Li Lucy, Xinxi Lyu, Nathan Lambert, Ian Magnusson, Jacob Morrison, Niklas Muennighoff, Aakanksha Naik, Cr… find
  58. Dan Su, Kezhi Kong, Ying Lin, Joseph Jennings, Brandon Norick, Markus Kliegl, Mostofa Patwary, Mohammad Shoeybi, and Bryan Catanzaro. Nemotron-cc: Transforming common crawl into a refined long-horizon pretraining dataset, 2024. URL https://arxiv.org/abs/2412.0 2595. ↗
  59. Yang Sui, Yu-Neng Chuang, Guanchu Wang, Jiamu Zhang, Tianyi Zhang, Jiayi Yuan, Hongyi Liu, Andrew Wen, Shaochen Zhong, Hanjie Chen, et al. Stop overthinking: A survey on efficient reasoning for large language models. arXiv preprint arXiv:2503.16419, 2025. find our card
  60. Shubham Toshniwal, Wei Du, Ivan Moshkov, Branislav Kisacanin, Alexan Ayrapetyan, and Igor Gitman. Openmathinstruct-2: Accelerating ai for math with massive open-source instruction data. arXiv preprint arXiv:2410.01560, 2024. find
  61. Zijun Wang, Haoqin Tu, Yuhan Wang, Juncheng Wu, Jieru Mei, Brian R Bartoldson, Bhavya Kailkhura, and Cihang Xie. Star-1: Safer alignment of reasoning llms with 1k data. arXiv preprint arXiv:2504.01903, 2025. — [arXiv:2504.01903](https://arxiv.org/abs/2504.01903) find
  62. Alexander Wettig, Kyle Lo, Sewon Min, Hannaneh Hajishirzi, Danqi Chen, and Luca Soldaini. Organize the web: Constructing domains enhances pre-training data curation. arXiv preprint arXiv:2502.10341, 2025. find
  63. Yixin Ye, Zhen Huang, Yang Xiao, Ethan Chern, Shijie Xia, and Pengfei Liu. Limo: Less is more for reasoning, 2025. URL https://arxiv.org/abs/2502.03387. ↗ our card
  64. Weizhe Yuan, Jane Yu, Song Jiang, Karthik Padthe, Yang Li, Dong Wang, Ilia Kulikov, Kyunghyun Cho, Yuandong Tian, Jason E Weston, and Xian Li. Naturalreasoning: Reasoning in the wild with 2.8m challenging questions, 2025. URL https://arxiv.org/abs/2502.13124. ↗
  65. Xiang Yue, Tuney Zheng, Ge Zhang, and Wenhu Chen. Mammoth2: Scaling instructions from the web. Advances in Neural Information Processing Systems, 2024. find our card
  66. Yifan Zhang, Yifan Luo, Yang Yuan, and Andrew Chi-Chih Yao. Autonomous data selection with language models for mathematical texts. arXiv preprint arXiv:2402.07625, 2024. — [arXiv:2402.07625](https://arxiv.org/abs/2402.07625) find