---
title: "Satori — paper by Shen et al. (2025) — summary and ideas"
description: "The authors introduce Chain-of-Action-Thought (COAT), in which special meta-action tokens signal… 100 references, cited by 13 papers here."
authors: ["Maohao Shen", "Guangtao Zeng", "Zhenting Qi", "Zhang-Wei Hong", "Zhenfang Chen", "Wei Lu", "Gregory Wornell", "Subhro Das", "David Cox", "Chuang Gan"]
year: 2025
venue: "arXiv:2502.02508"
source: "https://arxiv.org/abs/2502.02508"
doi: "10.48550/arxiv.2502.02508"
arxiv: "2502.02508"
licence: "CC BY-NC-ND 4.0 (arXiv)"
licence_url: "http://creativecommons.org/licenses/by-nc-nd/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/shen-2025-satori-reinforcement-learning-chain-action/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/shen-2025-satori-reinforcement-learning-chain-action/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2025](https://md.research.elamur.ai/free/en/papers/2025/page/9/) / Satori: Reinforcement Learning with Chain-of-Action-Thought Enhances LLM Reasoning via Autoregressive Search

# Satori: Reinforcement Learning with Chain-of-Action-Thought Enhances LLM Reasoning via Autoregressive Search

**Authors:** Maohao Shen, [Guangtao Zeng](https://md.research.elamur.ai/free/en/authors/guangtao-zeng/), [Zhenting Qi](https://md.research.elamur.ai/free/en/authors/zhenting-qi/), [Zhang-Wei Hong](https://md.research.elamur.ai/free/en/authors/zhang-wei-hong/), [Zhenfang Chen](https://md.research.elamur.ai/free/en/authors/zhenfang-chen/), [Wei Lu](https://md.research.elamur.ai/free/en/authors/wei-lu/), [Gregory Wornell](https://md.research.elamur.ai/free/en/authors/gregory-w-wornell/), Subhro Das, [David Cox](https://md.research.elamur.ai/free/en/authors/david-cox/) and [Chuang Gan](https://md.research.elamur.ai/free/en/authors/chuang-gan/) · **Year:** 2025 · **Venue:** arXiv:2502.02508 · **Licence:** [CC BY-NC-ND 4.0 (arXiv)](http://creativecommons.org/licenses/by-nc-nd/4.0/)

**Links:** [Publisher's page](https://arxiv.org/abs/2502.02508) · [PDF](https://arxiv.org/pdf/2502.02508) · [DOI 10.48550/arxiv.2502.02508](https://doi.org/10.48550/arxiv.2502.02508) · [arXiv 2502.02508](https://arxiv.org/abs/2502.02508) · [OpenAlex](https://openalex.org/W4407186893)

**Citations:** 0 citations (OpenAlex, 2026-10-14) · [13](https://md.research.elamur.ai/free/en/papers/shen-2025-satori-reinforcement-learning-chain-action/cited-by/) in our library, of the 149,439 papers whose bibliographies we read

## Summary
Most ways of boosting LLM reasoning at inference time rely on heavy sampling steered by a separate verifier or reward model. This raises deployment costs and keeps the search ability outside the model itself. The paper asks whether a single LLM can learn to search on its own: checking its steps, catching mistakes and switching strategies within one generated response. The authors introduce Chain-of-Action-Thought (COAT), in which special meta-action tokens signal whether the model should continue, reflect or explore an alternative path.

Training has two stages:
- **Format tuning.** A small format-tuning phase on roughly 10K demonstration trajectories, built with a generator–critic pipeline, teaches the COAT format.
- **Reinforcement learning.** Large-scale RL with "Restart and Explore" (RAE) then lets the model improve itself, using only final-answer labels as supervision.

The resulting 7B model, Satori, is built on open-source base models and math data. It achieves state-of-the-art results on math benchmarks and transfers strongly to tasks beyond its training domain. All of its code, data and models have been made fully public, positioning it as a transparent, small-scale counterpart to DeepSeek-R1.

## Key points
- **Venue and authors:** Satori (Maohao Shen et al.; MIT, SUTD, Harvard, MIT-IBM Watson AI Lab, UMass Amherst) appeared at ICML 2025. Materials are available at satori-reasoning.github.io.
- **COAT tokens:** COAT generalizes Chain-of-Thought by adding meta-action tokens such as `<|explore|>`. Standard CoT corresponds to the special case where every step is a plain continuation.
- **FT vs. RL ablation:** On Qwen-2.5-Math-7B, format tuning scaled up to 300K samples fails to match 10K FT plus large-scale RL on math benchmarks. It also does notably worse out of domain.
- **Adaptive compute:** After RL, Satori-Qwen produces longer responses for harder MATH difficulty levels and gains accuracy over Satori-Qwen-FT.
- **Distillation to weaker models:** Training Llama-3.1-8B and Granite-3.1-8B on 240K samples generated by Satori-Qwen-7B beats format-tuning them on an equal amount of FT data. This points to a low-cost route for strengthening weaker base models.

*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

- [Massachusetts Institute of Technology](https://md.research.elamur.ai/free/en/universities/massachusetts-institute-of-technology/) — University, Cambridge, United States · 274 works · 82,013 citations
- [IBM](https://md.research.elamur.ai/free/en/companies/ibm/) — Company, Armonk, United States · 163 works · 73,139 citations
- [University of Massachusetts Amherst](https://md.research.elamur.ai/free/en/universities/university-of-massachusetts-amherst/) — University, Amherst Center, United States · 122 works · 69,574 citations
- [Harvard University](https://md.research.elamur.ai/free/en/universities/harvard-university/) — University, Cambridge, United States · 137 works · 45,850 citations
- [Singapore University of Technology and Design](https://md.research.elamur.ai/free/en/universities/singapore-university-of-technology-and-design/) — University, Singapore, Singapore · 56 works · 11,583 citations

📚 Shelves (5): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 19,503 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 18,805 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 17,959 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 17,789 works · [👁️ Computer Vision and Pattern Recognition](https://md.research.elamur.ai/free/en/shelves/computer-science/computer-vision-and-pattern-recognition/) · 3,621 works

🧭 Topics (2): [🧠 Language models & pretraining](https://md.research.elamur.ai/free/en/topics/language-models/) · 4,145 works · [🧮 Reasoning](https://md.research.elamur.ai/free/en/topics/reasoning/) · 1,948 works

## Cite

Cite as: Maohao Shen, Guangtao Zeng, Zhenting Qi, Zhang-Wei Hong, Zhenfang Chen, Wei Lu, Gregory Wornell, Subhro Das, David Cox and Chuang Gan. 2025. Satori: Reinforcement Learning with Chain-of-Action-Thought Enhances LLM Reasoning via Autoregressive Search. arXiv:2502.02508.

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

```bibtex
@misc{shen2025satori,
  title = {{Satori: Reinforcement Learning with Chain-of-Action-Thought Enhances LLM Reasoning via Autoregressive Search}},
  author = {Maohao Shen, Guangtao Zeng, Zhenting Qi, Zhang-Wei Hong, Zhenfang Chen, Wei Lu, Gregory Wornell, Subhro Das, David Cox and Chuang Gan},
  year = {2025},
  doi = {10.48550/arxiv.2502.02508},
  eprint = {2502.02508},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.48550/arxiv.2502.02508}
}
```

- [← previous Safety Alignment via Constrained Knowledge Unlearning](https://md.research.elamur.ai/free/en/papers/shi-2025-safety-alignment-constrained-knowledge-unlearning/)
- [next → SciMaster: Towards General-Purpose Scientific AI Agents, Part I. X-Master as Foundation: Can We Lead on Humanity's Last Exam?](https://md.research.elamur.ai/free/en/papers/chai-2025-scimaster-towards-general-purpose-scientific/)
- [↑ back to papers of 2025](https://md.research.elamur.ai/free/en/papers/2025/page/9/)

## References

100 references, 65 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. K. Cobbe, V. Kosaraju, M. Bavarian, M. Chen, H. Jun, L. Kaiser, M. Plappert, J. Tworek, J. Hilton, R. Nakano et al., “[Training verifiers to solve math word problems](https://md.research.elamur.ai/free/en/papers/cobbe-2021-training-verifiers-solve-math-word/),” arXiv preprint arXiv:2110.14168, 2021. 1, 24, 36
2. D. Hendrycks, C. Burns, S. Kadavath, A. Arora, S. Basart, E. Tang, D. Song, and J. Steinhardt, “[Measuring mathematical problem solving with the MATH dataset](https://md.research.elamur.ai/free/en/papers/hendrycks-2021-measuring-mathematical-problem-solving-math/),” in Thirty-fifth Conference on Neural Information Processing Systems Datasets and Benchmarks Track (Round 2), 2021. 1, 2, 24, 36
3. M. Chen, J. Tworek, H. Jun, Q. Yuan, H. P. D. O. Pinto, J. Kaplan, H. Edwards, Y. Burda, N. Joseph, G. Brockman et al., “[Evaluating large language models trained on code](https://md.research.elamur.ai/free/en/papers/chen-2021-evaluating-large-language-models-trained/),” arXiv preprint arXiv:2107.03374, 2021. 1, 24
4. T. Y. Zhuo, M. C. Vu, J. Chim, H. Hu, W. Yu, R. Widyasari, I. N. B. Yusuf, H. Zhan, J. He, I. Paul et al., “[Bigcodebench: Benchmarking code generation with diverse function calls and complex instructions](https://md.research.elamur.ai/free/en/papers/zhuo-2024-bigcodebench-benchmarking-code-generation-diverse/),” arXiv preprint arXiv:2406.15877, 2024. 1, 24
5. S. Han, H. Schoelkopf, Y. Zhao, Z. Qi, M. Riddell et al., “[FOLIO: Natural language reasoning with first-order logic](https://md.research.elamur.ai/free/en/papers/han-2022-folio-natural-language-reasoning-first/),” in Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, Y. Al-Onaizan, M. Bansal, and Y.-N. Chen, Eds. Miami, Florida, USA: Association for Computa…
6. J. Liu, L. Cui, H. Liu, D. Huang, Y. Wang, and Y. Zhang, “[LogiQA: A challenge dataset for machine reading comprehension with logical reasoning](https://md.research.elamur.ai/free/en/papers/liu-2020-logiqa-challenge-dataset-machine-reading/),” in Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence, IJCAI 2020, C. Bessiere, Ed. ijcai.org, 2020, pp. 3622–3628.…
7. J. Wei, X. Wang, D. Schuurmans, M. Bosma, F. Xia, E. Chi, Q. V. Le, D. Zhou et al., “[Chain-of-thought prompting elicits reasoning in large language models](https://md.research.elamur.ai/free/en/papers/wei-2022-chain-thought-prompting-elicits-reasoning/),” Advances in neural information processing systems, vol. 35, pp. 24 824–24 837, 2022. 1, 3, 24
8. X. Yue, X. Qu, G. Zhang, Y. Fu, W. Huang, H. Sun, Y. Su, and W. Chen, “[MAmmoTH: Building math generalist models through hybrid instruction tuning](https://md.research.elamur.ai/free/en/papers/yue-2023-mammoth-building-math-generalist-models/),” in The Twelfth International Conference on Learning Representations, ICLR 2024, Vienna, Austria, May 7-11, 2024.
9. L. Yu, W. Jiang, H. Shi, J. Yu, Z. Liu, Y. Zhang, J. T. Kwok, Z. Li, A. Weller, and W. Liu, “[MetaMath: Bootstrap your own mathematical questions for large language models](https://md.research.elamur.ai/free/en/papers/yu-2023-metamath-bootstrap-your-own-mathematical/),” in The Twelfth International Conference on Learning Representations, ICLR 2024, Vienna, Austria, May 7-11, 2024.
10. S. Toshniwal, W. Du, I. Moshkov, B. Kisacanin, A. Ayrapetyan, and I. Gitman, “[Openmathinstruct-2: Accelerating AI for math with massive open-source instruction data](https://md.research.elamur.ai/free/en/papers/toshniwal-2024-openmathinstruct-2-accelerating-ai-math/),” arXiv preprint arXiv:2410.01560, 2024. 1, 2, 24, 27
11. Y. Ding, X. Shi, X. Liang, J. Li, Q. Zhu, and M. Zhang, “Unleashing reasoning capability of llms via scalable question synthesis from scratch,” arXiv preprint arXiv:2410.18693, 2024. 1, 2, 24 — [arXiv:2410.18693](https://arxiv.org/abs/2410.18693) [find](https://scholar.google.com/scholar?q=Y%2E+Ding%2C+X%2E+Shi%2C+X%2E+Liang%2C+J%2E+Li%2C+Q%2E+Zhu%2C+and+M%2E+Zhang%2C+%E2%80%9CUnleashing+reasoning+capability+of+llms+via+scalable+question+synthesis+from+scratch%2C%E2%80%9D+arXiv+preprint+arXiv%3A2410%2E18693%2C+2024%2E+1%2C+2%2C+24+%E2%80%94+%5BarXiv)
12. X. Wang, J. Wei, D. Schuurmans, Q. V. Le, E. H. Chi, S. Narang, A. Chowdhery, and D. Zhou, “[Self-consistency improves chain of thought reasoning in language models](https://md.research.elamur.ai/free/en/papers/wang-2022-self-consistency-improves-chain-thought/),” in The Eleventh International Conference on Learning Representations, 2023. 1
13. S. Yao, D. Yu, J. Zhao, I. Shafran, T. Griffiths, Y. Cao, and K. Narasimhan, “Tree of thoughts: Deliberate problem solving with large language models,” Advances in Neural Information Processing Systems, vol. 36, 2024. 1, 2, 24 [find](https://scholar.google.com/scholar?q=S%2E+Yao%2C+D%2E+Yu%2C+J%2E+Zhao%2C+I%2E+Shafran%2C+T%2E+Griffiths%2C+Y%2E+Cao%2C+and+K%2E+Narasimhan%2C+%E2%80%9CTree+of+thoughts%3A+Deliberate+problem+solving+with+large+language+models%2C%E2%80%9D+Advances+in+Neural+Information+Processing+System)
14. Z. Wan, X. Feng, M. Wen, S. M. McAleer, Y. Wen, W. Zhang, and J. Wang, “Alphazero-like tree-search can guide large language model decoding and training,” in Forty-first International Conference on Machine Learning, 2024. 1 [find](https://scholar.google.com/scholar?q=Z%2E+Wan%2C+X%2E+Feng%2C+M%2E+Wen%2C+S%2E+M%2E+McAleer%2C+Y%2E+Wen%2C+W%2E+Zhang%2C+and+J%2E+Wang%2C+%E2%80%9CAlphazero%2Dlike+tree%2Dsearch+can+guide+large+language+model+decoding+and+training%2C%E2%80%9D+in+Forty%2Dfirst+International+Conference+on+Mac)
15. Z. Sun, L. Yu, Y. Shen, W. Liu, Y. Yang, S. Welleck, and C. Gan, “[Easy-to-hard generalization: Scalable alignment beyond human supervision](https://md.research.elamur.ai/free/en/papers/sun-2024-easy-hard-generalization-scalable-alignment/),” in The Thirty-eighth Annual Conference on Neural Information Processing Systems, 2024. 1, 2, 24
16. P. Wang, L. Li, Z. Shao, R. Xu, D. Dai, Y. Li, D. Chen, Y. Wu, and Z. Sui, “[Math-Shepherd: Verify and reinforce LLMs step-by-step without human annotations](https://md.research.elamur.ai/free/en/papers/wang-2024-math-shepherd-verify-reinforce-llms/),” in Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), L.-W. Ku, A. Martins, and…
17. OpenAI, “Learning to reason with llms,” 2024, accessed: 2024-12-18. [Online]. Available: [https://openai.com/index/learning-to-reason-with-llms/](https://openai.com/index/learning-to-reason-with-llms/) 2 [↗](https://openai.com/index/learning-to-reason-with-llms/)
18. Z. Huang, H. Zou, X. Li, Y. Liu, Y. Zheng, E. Chern, S. Xia, Y. Qin, W. Yuan, and P. Liu, “O1 replication journey–part 2: Surpassing o1-preview through simple distillation, big progress or bitter lesson?” arXiv preprint arXiv:2411.[16489, 2024](https://md.research.elamur.ai/free/en/papers/huang-2024-o1-replication-journey-part-2/). 2, 24
19. Y. Zhao, H. Yin, B. Zeng, H. Wang, T. Shi, C. Lyu, L. Wang, W. Luo, and K. Zhang, “[Marco-o1: Towards open reasoning models for open-ended solutions](https://md.research.elamur.ai/free/en/papers/zhao-2024-marco-o1-towards-open-reasoning/),” arXiv preprint arXiv:2411.14405, 2024. 2, 24
20. Y. Min, Z. Chen, J. Jiang, J. Chen, J. Deng, Y. Hu, Y. Tang, J. Wang, X. Cheng, H. Song et al., “[Imitate, explore, and self-improve: A reproduction report on slow-thinking reasoning systems](https://md.research.elamur.ai/free/en/papers/min-2024-imitate-explore-self-improve-reproduction/),” arXiv preprint arXiv:2412.09413, 2024. 2, 24
21. D. Guo, D. Yang, H. Zhang, J. Song, R. Zhang, R. Xu, Q. Zhu, S. Ma, P. Wang, X. Bi et al., “[DeepSeek-R1: Incentivizing reasoning capability in llms via reinforcement learning](https://md.research.elamur.ai/free/en/papers/deepseek-ai-2025-deepseek-r1-incentivizing-reasoning-capability/),” arXiv preprint arXiv:2501.12948, 2025. 2
22. E. Zelikman, Y. Wu, J. Mu, and N. D. Goodman, “[STaR: Bootstrapping reasoning with reasoning](https://md.research.elamur.ai/free/en/papers/zelikman-2022-star-bootstrapping-reasoning-reasoning/),” in Advances in Neural Information Processing Systems 35: Annual Conference on Neural Information Processing Systems 2022, NeurIPS 2022, New Orleans, LA, USA, November 28 - December 9, 2022, S. Koyejo, S. Mo…
23. E. Zelikman, G. R. Harik, Y. Shao, V. Jayasiri, N. Haber, and N. Goodman, “[Quiet-STaR: Language models can teach themselves to think before speaking](https://md.research.elamur.ai/free/en/papers/al-2024-quiet-star-language-models-teach/),” in First Conference on Language Modeling, 2024. 2, 24
24. A. Singh, J. D. Co-Reyes, R. Agarwal, A. Anand, P. Patil, X. Garcia, P. J. Liu, J. Harrison et al., “[Beyond human data: Scaling self-training for problem-solving with language models](https://md.research.elamur.ai/free/en/papers/singh-2023-human-data-scaling-self-training/),” Trans. Mach. Learn. Res., vol. 2024, 2024. 2, 24
25. X. Zhang, C. Du, T. Pang, Q. Liu, W. Gao, and M. Lin, “Chain of preference optimization: Improving chain-of-thought reasoning in LLMs,” in The Thirty-eighth Annual Conference on Neural Information Processing Systems, 2024. 2, 24 [find](https://scholar.google.com/scholar?q=X%2E+Zhang%2C+C%2E+Du%2C+T%2E+Pang%2C+Q%2E+Liu%2C+W%2E+Gao%2C+and+M%2E+Lin%2C+%E2%80%9CChain+of+preference+optimization%3A+Improving+chain%2Dof%2Dthought+reasoning+in+LLMs%2C%E2%80%9D+in+The+Thirty%2Deighth+Annual+Conference+on+Neural+Information+Pro)
26. J. Schulman, F. Wolski, P. Dhariwal, A. Radford, and O. Klimov, “Proximal policy optimization algorithms,” arXiv preprint arXiv:1707.06347, 2017. 2, 24, 35 [find](https://scholar.google.com/scholar?q=J%2E+Schulman%2C+F%2E+Wolski%2C+P%2E+Dhariwal%2C+A%2E+Radford%2C+and+O%2E+Klimov%2C+%E2%80%9CProximal+policy+optimization+algorithms%2C%E2%80%9D+arXiv+preprint+arXiv%3A1707%2E06347%2C+2017%2E+2%2C+24%2C+35)
27. L. Ouyang, J. Wu, X. Jiang, D. Almeida, C. L. Wainwright, P. Mishkin, C. Zhang, S. Agarwal, K. Slama, A. Ray, J. Schulman, J. Hilton, F. Kelton, L. Miller, M. Simens, A. Askell, P. Welinder, P. F. Christiano, J. Leike, and R. Lowe, “[Training language models to follow instructions with human feedback](https://md.research.elamur.ai/free/en/papers/ouyang-2022-training-language-models-follow-instructions/)…
28. L. Yuan, W. Li, H. Chen, G. Cui, N. Ding, K. Zhang, B. Zhou, Z. Liu, and H. Peng, “[Free process rewards without process labels](https://md.research.elamur.ai/free/en/papers/yuan-2024-free-process-rewards-process-labels/),” arXiv preprint arXiv:2412.01981, 2024. 2, 24
29. N. Shinn, F. Cassano, A. Gopinath, K. Narasimhan, and S. Yao, “[Reflexion: Language agents with verbal reinforcement learning](https://md.research.elamur.ai/free/en/papers/shinn-2023-reflexion-language-agents-verbal-reinforcement/),” Advances in Neural Information Processing Systems, vol. 36, 2024. 2, 24
30. S. Hao, Y. Gu, H. Ma, J. J. Hong, Z. Wang, D. Z. Wang, and Z. Hu, “Reasoning with language model is planning with world model,” in Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, EMNLP 2023, Singapore, December 6-10, 2023, H. Bouamor, J. Pino, and K. Bali, Eds… [find](https://scholar.google.com/scholar?q=S%2E+Hao%2C+Y%2E+Gu%2C+H%2E+Ma%2C+J%2E+J%2E+Hong%2C+Z%2E+Wang%2C+D%2E+Z%2E+Wang%2C+and+Z%2E+Hu%2C+%E2%80%9CReasoning+with+language+model+is+planning+with+world+model%2C%E2%80%9D+in+Proceedings+of+the+2023+Conference+on+Empirical+Methods+in+Natural+La)
31. Z. Qi, M. Ma, J. Xu, L. L. Zhang, F. Yang, and M. Yang, “[Mutual reasoning makes smaller llms stronger problem-solvers](https://md.research.elamur.ai/free/en/papers/qi-2024-mutual-reasoning-makes-smaller-llms/),” arXiv preprint arXiv:2408.06195, 2024. 2, 24, 27
32. Y. Zhang, M. Khalifa, L. Logeswaran, J. Kim, M. Lee, H. Lee, and L. Wang, “Small language models need strong verifiers to self-correct reasoning,” in Findings of the Association for Computational Linguistics, ACL 2024, Bangkok, Thailand and virtual meeting, August 11-16, 2024, L. Ku, A. Martins, and… [find](https://scholar.google.com/scholar?q=Y%2E+Zhang%2C+M%2E+Khalifa%2C+L%2E+Logeswaran%2C+J%2E+Kim%2C+M%2E+Lee%2C+H%2E+Lee%2C+and+L%2E+Wang%2C+%E2%80%9CSmall+language+models+need+strong+verifiers+to+self%2Dcorrect+reasoning%2C%E2%80%9D+in+Findings+of+the+Association+for+Computational+Ling)
33. R. Kamoi, Y. Zhang, N. Zhang, J. Han, and R. Zhang, “[When can llms actually correct their own mistakes? a critical survey of self-correction of llms](https://md.research.elamur.ai/free/en/papers/kamoi-2024-when-can-llms-actually-correct/),” Transactions of the Association for Computational Linguistics, vol. 12, pp. 1417–1440, 2024. 2, 24
34. W. Saunders, C. Yeh, J. Wu, S. Bills, L. Ouyang, J. Ward, and J. Leike, “[Self-critiquing models for assisting human evaluators](https://md.research.elamur.ai/free/en/papers/saunders-2022-self-critiquing-models-assisting-human/),” arXiv preprint arXiv:2206.05802, 2022. 2, 24
35. A. Kumar, V. Zhuang, R. Agarwal, Y. Su, J. D. Co-Reyes, A. Singh, K. Baumli, S. Iqbal, C. Bishop, R. Roelofs et al., “[Training language models to self-correct via reinforcement learning](https://md.research.elamur.ai/free/en/papers/kumar-2024-training-language-models-self-correct/),” arXiv preprint arXiv:2409.12917, 2024. 2, 24
36. Y. Qu, T. Zhang, N. Garg, and A. Kumar, “[Recursive introspection: Teaching language model agents how to self-improve](https://md.research.elamur.ai/free/en/papers/qu-2024-recursive-introspection-teaching-language-model/),” in The Thirty-eighth Annual Conference on Neural Information Processing Systems, 2024. 2, 24
37. A. Havrilla, S. C. Raparthy, C. Nalmpantis, J. Dwivedi-Yu, M. Zhuravinskyi, E. Hambro, and R. Raileanu, “[GLoRe: When, where, and how to improve LLM reasoning via global and local refinements](https://md.research.elamur.ai/free/en/papers/havrilla-2024-glore-where-improve-llm-reasoning/),” in Forty-first International Conference on Machine Learning, ICML 2024, Vienna, Austria, July 21-27, 2024.
38. Z. Xi, D. Yang, J. Huang, J. Tang, G. Li, Y. Ding, W. He, B. Hong, S. Do, W. Zhan et al., “Enhancing llm reasoning via critique models with test-time and training-time supervision,” arXiv preprint arXiv:2411.16579, 2024. 2, 24 — [arXiv:2411.16579](https://arxiv.org/abs/2411.16579) [find](https://scholar.google.com/scholar?q=Z%2E+Xi%2C+D%2E+Yang%2C+J%2E+Huang%2C+J%2E+Tang%2C+G%2E+Li%2C+Y%2E+Ding%2C+W%2E+He%2C+B%2E+Hong%2C+S%2E+Do%2C+W%2E+Zhan+et+al%2E%2C+%E2%80%9CEnhancing+llm+reasoning+via+critique+models+with+test%2Dtime+and+training%2Dtime+supervision%2C%E2%80%9D+arXiv+preprint+arX)
39. A. Setlur, C. Nagpal, A. Fisch, X. Geng, J. Eisenstein, R. Agarwal, A. Agarwal, J. Berant, and A. Kumar, “[Rewarding progress: Scaling automated process verifiers for llm reasoning](https://md.research.elamur.ai/free/en/papers/setlur-2024-rewarding-progress-scaling-automated-process/),” arXiv preprint arXiv:2410.08146, 2024. 2, 24
40. L. Zhang, A. Hosseini, H. Bansal, M. Kazemi, A. Kumar, and R. Agarwal, “[Generative verifiers: Reward modeling as next-token prediction](https://md.research.elamur.ai/free/en/papers/zhang-2024-generative-verifiers-reward-modeling-as/),” in The 4th Workshop on Mathematical Reasoning and AI at NeurIPS’24, 2024. 2, 24
41. X. Guan, L. L. Zhang, Y. Liu, N. Shang, Y. Sun, Y. Zhu, F. Yang, and M. Yang, “[rstar-math: Small llms can master math reasoning with self-evolved deep thinking](https://md.research.elamur.ai/free/en/papers/guan-2025-rstar-math-small-llms-can/),” arXiv preprint arXiv:2501.04519, 2025. 2, 24
42. D. Zhang, S. Zhoubian, Z. Hu, Y. Yue, Y. Dong, and J. Tang, “[ReST-MCTS*: LLM self-training via process reward guided tree search](https://md.research.elamur.ai/free/en/papers/zhang-2024-rest-mcts-llm-self-training/),” in The Thirty-eighth Annual Conference on Neural Information Processing Systems, 2024. 2, 24
43. K. Gandhi, D. H. J. Lee, G. Grand, M. Liu, W. Cheng, A. Sharma, and N. Goodman, “[Stream of search (SoS): Learning to search in language](https://md.research.elamur.ai/free/en/papers/gandhi-2024-stream-search-sos-learning-search/),” in First Conference on Language Modeling, 2024. 2, 24
44. L. P. Kaelbling, M. L. Littman, and A. W. Moore, “Reinforcement learning: A survey,” Journal of artificial intelligence research, vol. 4, pp. 237–285, 1996. 2 [find](https://scholar.google.com/scholar?q=L%2E+P%2E+Kaelbling%2C+M%2E+L%2E+Littman%2C+and+A%2E+W%2E+Moore%2C+%E2%80%9CReinforcement+learning%3A+A+survey%2C%E2%80%9D+Journal+of+artificial+intelligence+research%2C+vol%2E+4%2C+pp%2E+237%E2%80%93285%2C+1996%2E+2)
45. M. Bellemare, S. Srinivasan, G. Ostrovski, T. Schaul, D. Saxton, and R. Munos, “Unifying count-based exploration and intrinsic motivation,” Advances in neural information processing systems, vol. 29, 2016. 3 [find](https://scholar.google.com/scholar?q=M%2E+Bellemare%2C+S%2E+Srinivasan%2C+G%2E+Ostrovski%2C+T%2E+Schaul%2C+D%2E+Saxton%2C+and+R%2E+Munos%2C+%E2%80%9CUnifying+count%2Dbased+exploration+and+intrinsic+motivation%2C%E2%80%9D+Advances+in+neural+information+processing+systems%2C+vol%2E+29%2C+)
46. A. Ecoffet, J. Huizinga, J. Lehman, K. O. Stanley, and J. Clune, “Go-explore: a new approach for hard-exploration problems,” arXiv preprint arXiv:1901.10995, 2019. 4 [find](https://scholar.google.com/scholar?q=A%2E+Ecoffet%2C+J%2E+Huizinga%2C+J%2E+Lehman%2C+K%2E+O%2E+Stanley%2C+and+J%2E+Clune%2C+%E2%80%9CGo%2Dexplore%3A+a+new+approach+for+hard%2Dexploration+problems%2C%E2%80%9D+arXiv+preprint+arXiv%3A1901%2E10995%2C+2019%2E+4)
47. A. Hussein, M. M. Gaber, E. Elyan, and C. Jayne, “Imitation learning: A survey of learning methods,” ACM Computing Surveys (CSUR), vol. 50, no. 2, pp. 1–35, 2017. 4 [find](https://scholar.google.com/scholar?q=A%2E+Hussein%2C+M%2E+M%2E+Gaber%2C+E%2E+Elyan%2C+and+C%2E+Jayne%2C+%E2%80%9CImitation+learning%3A+A+survey+of+learning+methods%2C%E2%80%9D+ACM+Computing+Surveys+%28CSUR%29%2C+vol%2E+50%2C+no%2E+2%2C+pp%2E+1%E2%80%9335%2C+2017%2E+4)
48. S. Ross and D. Bagnell, “Efficient reductions for imitation learning,” in Proceedings of the thirteenth international conference on artificial intelligence and statistics. JMLR Workshop and Conference Proceedings, 2010, pp. 661– 668. 4 [find](https://scholar.google.com/scholar?q=S%2E+Ross+and+D%2E+Bagnell%2C+%E2%80%9CEfficient+reductions+for+imitation+learning%2C%E2%80%9D+in+Proceedings+of+the+thirteenth+international+conference+on+artificial+intelligence+and+statistics%2E+JMLR+Workshop+and+Conference)
49. S. Ross, G. Gordon, and D. Bagnell, “A reduction of imitation learning and structured prediction to no-regret online learning,” in Proceedings of the fourteenth international conference on artificial intelligence and statistics. JMLR Workshop and Conference Proceedings, 2011, pp. 627–635. 4 [find](https://scholar.google.com/scholar?q=S%2E+Ross%2C+G%2E+Gordon%2C+and+D%2E+Bagnell%2C+%E2%80%9CA+reduction+of+imitation+learning+and+structured+prediction+to+no%2Dregret+online+learning%2C%E2%80%9D+in+Proceedings+of+the+fourteenth+international+conference+on+artificial+)
50. J. Ho and S. Ermon, “Generative adversarial imitation learning,” Advances in neural information processing systems, vol. 29, 2016. 4 [find](https://scholar.google.com/scholar?q=J%2E+Ho+and+S%2E+Ermon%2C+%E2%80%9CGenerative+adversarial+imitation+learning%2C%E2%80%9D+Advances+in+neural+information+processing+systems%2C+vol%2E+29%2C+2016%2E+4)
51. J. Schulman, F. Wolski, P. Dhariwal, A. Radford, and O. Klimov, “Proximal policy optimization algorithms,” arXiv preprint arXiv:1707.06347, 2017. 4 [find](https://scholar.google.com/scholar?q=J%2E+Schulman%2C+F%2E+Wolski%2C+P%2E+Dhariwal%2C+A%2E+Radford%2C+and+O%2E+Klimov%2C+%E2%80%9CProximal+policy+optimization+algorithms%2C%E2%80%9D+arXiv+preprint+arXiv%3A1707%2E06347%2C+2017%2E+4)
52. S. Kakade and J. Langford, “Approximately optimal approximate reinforcement learning,” in Proceedings of the nineteenth international conference on machine learning, 2002, pp. 267–274. 5 [find](https://scholar.google.com/scholar?q=S%2E+Kakade+and+J%2E+Langford%2C+%E2%80%9CApproximately+optimal+approximate+reinforcement+learning%2C%E2%80%9D+in+Proceedings+of+the+nineteenth+international+conference+on+machine+learning%2C+2002%2C+pp%2E+267%E2%80%93274%2E+5)
53. R. Agarwal, M. Schwarzer, P. S. Castro, A. C. Courville, and M. Bellemare, “Reincarnating reinforcement learning: Reusing prior computation to accelerate progress,” Advances in neural information processing systems, vol. 35, pp. 28 955–28 971, 2022. 5 [find](https://scholar.google.com/scholar?q=R%2E+Agarwal%2C+M%2E+Schwarzer%2C+P%2E+S%2E+Castro%2C+A%2E+C%2E+Courville%2C+and+M%2E+Bellemare%2C+%E2%80%9CReincarnating+reinforcement+learning%3A+Reusing+prior+computation+to+accelerate+progress%2C%E2%80%9D+Advances+in+neural+information+proc)
54. S. Schmitt, J. J. Hudson, A. Zidek, S. Osindero, C. Doersch, W. M. Czarnecki, J. Z. Leibo, H. Kuttler, A. Zisserman, K. Simonyan et al., “Kickstarting deep reinforcement learning,” arXiv preprint arXiv:1803.03835, 2018. 5 [find](https://scholar.google.com/scholar?q=S%2E+Schmitt%2C+J%2E+J%2E+Hudson%2C+A%2E+Zidek%2C+S%2E+Osindero%2C+C%2E+Doersch%2C+W%2E+M%2E+Czarnecki%2C+J%2E+Z%2E+Leibo%2C+H%2E+Kuttler%2C+A%2E+Zisserman%2C+K%2E+Simonyan+et+al%2E%2C+%E2%80%9CKickstarting+deep+reinforcement+learning%2C%E2%80%9D+arXiv+preprint+arXi)
55. H. Lightman, V. Kosaraju, Y. Burda, H. Edwards, B. Baker, T. Lee, J. Leike, J. Schulman, I. Sutskever, and K. Cobbe, “Let’s verify step by step,” arXiv preprint arXiv:2305.[20050, 2023](https://md.research.elamur.ai/free/en/papers/lightman-2023-let-s-verify-step-step/). 6, 36
56. M. Kazemi, Q. Yuan, D. Bhatia, N. Kim, X. Xu, V. Imbrasaite, and D. Ramachandran, “Boardgameqa: A dataset for natural language reasoning with contradictory information,” Advances in Neural Information Processing Systems, vol. 36, 2024. 6, 36 [find](https://scholar.google.com/scholar?q=M%2E+Kazemi%2C+Q%2E+Yuan%2C+D%2E+Bhatia%2C+N%2E+Kim%2C+X%2E+Xu%2C+V%2E+Imbrasaite%2C+and+D%2E+Ramachandran%2C+%E2%80%9CBoardgameqa%3A+A+dataset+for+natural+language+reasoning+with+contradictory+information%2C%E2%80%9D+Advances+in+Neural+Information)
57. A. Gu, B. Rozière, H. Leather, A. Solar-Lezama, G. Synnaeve, and S. I. Wang, “[Cruxeval: A benchmark for code reasoning, understanding and execution](https://md.research.elamur.ai/free/en/papers/gu-2024-cruxeval-benchmark-code-reasoning-understanding/),” arXiv preprint arXiv:2401.03065, 2024. 6, 24, 36
58. M. Geva, D. Khashabi, E. Segal, T. Khot, D. Roth, and J. Berant, “Did aristotle use a laptop? a question answering benchmark with implicit reasoning strategies,” Transactions of the Association for Computational Linguistics, vol. 9, pp. 346–361, 2021. 6, 24, 36 [find](https://scholar.google.com/scholar?q=M%2E+Geva%2C+D%2E+Khashabi%2C+E%2E+Segal%2C+T%2E+Khot%2C+D%2E+Roth%2C+and+J%2E+Berant%2C+%E2%80%9CDid+aristotle+use+a+laptop%3F+a+question+answering+benchmark+with+implicit+reasoning+strategies%2C%E2%80%9D+Transactions+of+the+Association+for+Co)
59. X. Wu, J. Yang, L. Chai, G. Zhang, J. Liu, X. Du, D. Liang, D. Shu, X. Cheng, T. Sun et al., “Tablebench: A comprehensive and complex benchmark for table question answering,” arXiv preprint arXiv:2408.09174, 2024. 6, 24, 36 — [arXiv:2408.09174](https://arxiv.org/abs/2408.09174) [find](https://scholar.google.com/scholar?q=X%2E+Wu%2C+J%2E+Yang%2C+L%2E+Chai%2C+G%2E+Zhang%2C+J%2E+Liu%2C+X%2E+Du%2C+D%2E+Liang%2C+D%2E+Shu%2C+X%2E+Cheng%2C+T%2E+Sun+et+al%2E%2C+%E2%80%9CTablebench%3A+A+comprehensive+and+complex+benchmark+for+table+question+answering%2C%E2%80%9D+arXiv+preprint+arXiv%3A2408)
60. Y. Wang, X. Ma, G. Zhang, Y. Ni, A. Chandra, S. Guo, W. Ren, A. Arulraj, X. He, Z. Jiang et al., “[Mmlu-pro: A more robust and challenging multi-task language understanding benchmark](https://md.research.elamur.ai/free/en/papers/wang-2024-mmlu-pro-more-robust-challenging/),” arXiv preprint arXiv:2406.01574, 2024. 6, 24, 36
61. A. Yang, B. Zhang, B. Hui, B. Gao, B. Yu, C. Li, D. Liu, J. Tu, J. Zhou, J. Lin et al., “[Qwen2.5-math technical report: Toward mathematical expert model via self-improvement](https://md.research.elamur.ai/free/en/papers/yang-2024-qwen2-5-math-technical-report/),” arXiv preprint arXiv:2409.12122, 2024. 6
62. E. Glazer, E. Erdil, T. Besiroglu, D. Chicharro, E. Chen, A. Gunning, C. F. Olsson, J.-S. Denain, A. Ho, E. d. O. Santos et al., “[Frontiermath: A benchmark for evaluating advanced mathematical reasoning in ai](https://md.research.elamur.ai/free/en/papers/glazer-2024-frontiermath-benchmark-evaluating-advanced-mathematical/),” arXiv preprint arXiv:2411.04872, 2024. 24
63. I. Mirzadeh, K. Alizadeh, H. Shahrokhi, O. Tuzel, S. Bengio, and M. Farajtabar, “[Gsm-symbolic: Understanding the limitations of mathematical reasoning in large language models](https://md.research.elamur.ai/free/en/papers/mirzadeh-2024-gsm-symbolic-understanding-limitations-mathematical/),” arXiv preprint arXiv:2410.05229, 2024. 24
64. J. Austin, A. Odena, M. Nye, M. Bosma, H. Michalewski, D. Dohan, E. Jiang, C. Cai, M. Terry, Q. Le et al., “[Program synthesis with large language models](https://md.research.elamur.ai/free/en/papers/austin-2021-program-synthesis-large-language-models/),” arXiv preprint arXiv:2108.07732, 2021. 24
65. J. Dai, J. Lu, Y. Feng, D. Huang, G. Zeng, R. Ruan, M. Cheng, H. Tan, and Z. Guo, “Mhpp: Exploring the capabilities and limitations of language models beyond basic code generation,” arXiv preprint arXiv:2405.11430, 2024. 24 — [arXiv:2405.11430](https://arxiv.org/abs/2405.11430) [find](https://scholar.google.com/scholar?q=J%2E+Dai%2C+J%2E+Lu%2C+Y%2E+Feng%2C+D%2E+Huang%2C+G%2E+Zeng%2C+R%2E+Ruan%2C+M%2E+Cheng%2C+H%2E+Tan%2C+and+Z%2E+Guo%2C+%E2%80%9CMhpp%3A+Exploring+the+capabilities+and+limitations+of+language+models+beyond+basic+code+generation%2C%E2%80%9D+arXiv+preprint+arX)
66. C. E. Jimenez, J. Yang, A. Wettig, S. Yao, K. Pei, O. Press, and K. R. Narasimhan, “[SWE-bench: Can language models resolve real-world github issues?](https://md.research.elamur.ai/free/en/papers/jimenez-2024-swe-bench-can-language-models/)” in The Twelfth International Conference on Learning Representations, ICLR 2024, Vienna, Austria, May 7-11, 2024.
67. C. Liu, J. Shen, H. Xin, Z. Liu, Y. Yuan, H. Wang, W. Ju, C. Zheng, Y. Yin, L. Li et al., “[Fimo: A challenge formal dataset for automated theorem proving](https://md.research.elamur.ai/free/en/papers/liu-2023-fimo-challenge-formal-dataset-automated/),” arXiv preprint arXiv:2309.04295, 2023. 24
68. O. Tafjord, B. Dalvi, and P. Clark, “[ProofWriter: Generating implications, proofs, and abductive statements over natural language](https://md.research.elamur.ai/free/en/papers/tafjord-2021-proofwriter-generating-implications-proofs-abductive/),” in Findings of the Association for Computational Linguistics: ACL/IJCNLP 2021, Online Event, August 1-6, 2021, ser. Findings of ACL, C. Zong, F. Xia, W. Li, and R. Navi…
69. A. Talmor, J. Herzig, N. Lourie, and J. Berant, “[Commonsenseqa: A question answering challenge targeting commonsense knowledge](https://md.research.elamur.ai/free/en/papers/talmor-2019-commonsenseqa-question-answering-challenge-targeting/),” arXiv preprint arXiv:1811.00937, 2018. 24
70. P. Veličković, A. P. Badia, D. Budden, R. Pascanu, A. Banino, M. Dashevskiy, R. Hadsell, and C. Blundell, “The CLRS algorithmic reasoning benchmark,” in International Conference on Machine Learning. PMLR, 2022, pp. 22 084–22 102. 24 [find](https://scholar.google.com/scholar?q=P%2E+Velic%CC%8Ckovic%CC%81%2C+A%2E+P%2E+Badia%2C+D%2E+Budden%2C+R%2E+Pascanu%2C+A%2E+Banino%2C+M%2E+Dashevskiy%2C+R%2E+Hadsell%2C+and+C%2E+Blundell%2C+%E2%80%9CThe+CLRS+algorithmic+reasoning+benchmark%2C%E2%80%9D+in+International+Conference+on+Machine+Learning%2E)
71. L. Markeeva, S. McLeish, B. Ibarz, W. Bounsi, O. Kozlova, A. Vitvitskyi, C. Blundell, T. Goldstein, A. Schwarzschild, and P. Veličković, “The CLRS-Text algorithmic reasoning language benchmark,” arXiv preprint arXiv:2406.04229, 2024. 24 [find](https://scholar.google.com/scholar?q=L%2E+Markeeva%2C+S%2E+McLeish%2C+B%2E+Ibarz%2C+W%2E+Bounsi%2C+O%2E+Kozlova%2C+A%2E+Vitvitskyi%2C+C%2E+Blundell%2C+T%2E+Goldstein%2C+A%2E+Schwarzschild%2C+and+P%2E+Velic%CC%8Ckovic%CC%81%2C+%E2%80%9CThe+CLRS%2DText+algorithmic+reasoning+language+benchmark%2C%E2%80%9D+arX)
72. Z. Qi, H. Luo, X. Huang, Z. Zhao, Y. Jiang, X. Fan, H. Lakkaraju, and J. Glass, “Quantifying generalization complexity for large language models,” arXiv preprint arXiv:2410.01769, 2024. 24 — [arXiv:2410.01769](https://arxiv.org/abs/2410.01769) [find](https://scholar.google.com/scholar?q=Z%2E+Qi%2C+H%2E+Luo%2C+X%2E+Huang%2C+Z%2E+Zhao%2C+Y%2E+Jiang%2C+X%2E+Fan%2C+H%2E+Lakkaraju%2C+and+J%2E+Glass%2C+%E2%80%9CQuantifying+generalization+complexity+for+large+language+models%2C%E2%80%9D+arXiv+preprint+arXiv%3A2410%2E01769%2C+2024%2E+24+%E2%80%94+%5BarXiv%3A24)
73. L. Fan, W. Hua, L. Li, H. Ling, and Y. Zhang, “NPHard-Eval: Dynamic benchmark on reasoning ability of large language models via complexity classes,” in Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), ACL 2024, Bangkok, Thailand, August… [find](https://scholar.google.com/scholar?q=L%2E+Fan%2C+W%2E+Hua%2C+L%2E+Li%2C+H%2E+Ling%2C+and+Y%2E+Zhang%2C+%E2%80%9CNPHard%2DEval%3A+Dynamic+benchmark+on+reasoning+ability+of+large+language+models+via+complexity+classes%2C%E2%80%9D+in+Proceedings+of+the+62nd+Annual+Meeting+of+the+As)
74. P. Lu, L. Qiu, K.-W. Chang, Y. N. Wu, S.-C. Zhu, T. Rajpurohit, P. Clark, and A. Kalyan, “[Dynamic prompt learning via policy gradient for semi-structured mathematical reasoning](https://md.research.elamur.ai/free/en/papers/lu-2022-dynamic-prompt-learning-policy-gradient/),” arXiv preprint arXiv:2209.14610, 2022. 24
75. X. Wang, Z. Hu, P. Lu, Y. Zhu, J. Zhang, S. Subramaniam, A. R. Loomba, S. Zhang, Y. Sun, and W. Wang, “[SciBench: Evaluating college-level scientific problem-solving abilities of large language models](https://md.research.elamur.ai/free/en/papers/wang-2023-scibench-evaluating-college-level-scientific/),” in Forty-first International Conference on Machine Learning, ICML 2024, Vienna, Austria, July 21-2…
76. D. Rein, B. L. Hou, A. C. Stickland, J. Petty, R. Y. Pang, J. Dirani, J. Michael, and S. R. Bowman, “[GPQA: A graduate-level google-proof q&a benchmark](https://md.research.elamur.ai/free/en/papers/rein-2023-gpqa-graduate-level-google-proof/),” in First Conference on Language Modeling, 2024. 24
77. D. Hendrycks, C. Burns, S. Basart, A. Zou, M. Mazeika, D. Song, and J. Steinhardt, “[Measuring massive multitask language understanding](https://md.research.elamur.ai/free/en/papers/hendrycks-2020-measuring-massive-multitask-language-understanding/),” in 9th International Conference on Learning Representations, ICLR 2021, Virtual Event, Austria, May 3-7, 2021.
78. A. Srivastava, A. Rastogi, A. Rao, A. A. M. Shoeb, A. Abid, A. Fisch, A. R. Brown et al., “[Beyond the imitation game: Quantifying and extrapolating the capabilities of language models](https://md.research.elamur.ai/free/en/papers/srivastava-2022-imitation-game-quantifying-extrapolating-capabilities/),” Trans. Mach. Learn. Res., vol. 2023, 2023. 24
79. P. Liang, R. Bommasani, T. Lee, D. Tsipras, D. Soylu, M. Yasunaga et al., “Holistic evaluation of language models,” Trans. Mach. Learn. Res., vol. 2023, 2023. 24 [find](https://scholar.google.com/scholar?q=P%2E+Liang%2C+R%2E+Bommasani%2C+T%2E+Lee%2C+D%2E+Tsipras%2C+D%2E+Soylu%2C+M%2E+Yasunaga+et+al%2E%2C+%E2%80%9CHolistic+evaluation+of+language+models%2C%E2%80%9D+Trans%2E+Mach%2E+Learn%2E+Res%2E%2C+vol%2E+2023%2C+2023%2E+24)
80. L. Phan, A. Gatti, Z. Han, N. Li, J. Hu, H. Zhang, S. Shi, M. Choi, A. Agrawal, A. Chopra, A. Khoja, R. Kim, J. Hausenloy et al., “Humanity’s last exam,” 2025. [Online](https://md.research.elamur.ai/free/en/papers/phan-2025-humanity-s-last-exam/). Available: https://arxiv.org/abs/2501.14249 24 [↗](https://arxiv.org/abs/2501.14249)
81. W. Saunders, C. Yeh, J. Wu, S. Bills, L. Ouyang, J. Ward, and J. Leike, “[Self-critiquing models for assisting human evaluators](https://md.research.elamur.ai/free/en/papers/saunders-2022-self-critiquing-models-assisting-human/),” arXiv preprint arXiv:2206.05802, 2022. 24
82. S. Toshniwal, I. Moshkov, S. Narenthiran, D. Gitman, F. Jia, and I. Gitman, “[OpenMathInstruct-1: A 1.8 million math instruction tuning dataset](https://md.research.elamur.ai/free/en/papers/toshniwal-2024-openmathinstruct-1-1-8-million/),” in The Thirty-eight Conference on Neural Information Processing Systems Datasets and Benchmarks Track, 2024. 24
83. H. Luo, Q. Sun, C. Xu, P. Zhao, J. Lou, C. Tao, X. Geng, Q. Lin, S. Chen, and D. Zhang, “[Wizardmath: Empowering mathematical reasoning for large language models via reinforced evol-instruct](https://md.research.elamur.ai/free/en/papers/luo-2023-wizardmath-empowering-mathematical-reasoning-large/),” arXiv preprint arXiv:2308.09583, 2023. 24
84. M. Abdin, J. Aneja, H. Behl, S. Bubeck, R. Eldan, S. Gunasekar, M. Harrison, R. J. Hewett, M. Javaheripi, P. Kauffmann et al., “[Phi-4 technical report](https://md.research.elamur.ai/free/en/papers/abdin-2024-phi-4-technical-report/),” arXiv preprint arXiv:2412.08905, 2024. 24
85. I. Shumailov, Z. Shumaylov, Y. Zhao, Y. Gal, N. Papernot, and R. Anderson, “[The curse of recursion: Training on generated data makes models forget](https://md.research.elamur.ai/free/en/papers/shumailov-2023-curse-recursion-training-generated-data/),” arXiv preprint arXiv:2305.17493, 2023. 24
86. T. Wu, X. Li, and P. Liu, “[Progress or regress? self-improvement reversal in post-training](https://md.research.elamur.ai/free/en/papers/wu-2024-progress-regress-self-improvement-reversal/),” arXiv preprint arXiv:2407.05013, 2024. 24
87. M. Besta, N. Blach, A. Kubicek, R. Gerstenberger, M. Podstawski, L. Gianinazzi, J. Gajda, T. Lehmann, H. Niewiadomski, P. Nyczyk et al., “Graph of thoughts: Solving elaborate problems with large language models,” in Proceedings of the AAAI Conference on Artificial Intelligence, vol. 38, no. 16, 2024… [find](https://scholar.google.com/scholar?q=M%2E+Besta%2C+N%2E+Blach%2C+A%2E+Kubicek%2C+R%2E+Gerstenberger%2C+M%2E+Podstawski%2C+L%2E+Gianinazzi%2C+J%2E+Gajda%2C+T%2E+Lehmann%2C+H%2E+Niewiadomski%2C+P%2E+Nyczyk+et+al%2E%2C+%E2%80%9CGraph+of+thoughts%3A+Solving+elaborate+problems+with+large+langu)
88. A. Madaan, N. Tandon, P. Gupta, S. Hallinan, L. Gao, S. Wiegreffe, U. Alon, N. Dziri, S. Prabhumoye, Y. Yang et al., “Self-refine: Iterative refinement with self-feedback,” Advances in Neural Information Processing Systems, vol. 36, 2024. 24 [find](https://scholar.google.com/scholar?q=A%2E+Madaan%2C+N%2E+Tandon%2C+P%2E+Gupta%2C+S%2E+Hallinan%2C+L%2E+Gao%2C+S%2E+Wiegreffe%2C+U%2E+Alon%2C+N%2E+Dziri%2C+S%2E+Prabhumoye%2C+Y%2E+Yang+et+al%2E%2C+%E2%80%9CSelf%2Drefine%3A+Iterative+refinement+with+self%2Dfeedback%2C%E2%80%9D+Advances+in+Neural+Informat)
89. L. Yang, Z. Yu, T. Zhang, S. Cao, M. Xu, W. Zhang, J. E. Gonzalez, and B. CUI, “[Buffer of Thoughts: Thought-augmented reasoning with large language models](https://md.research.elamur.ai/free/en/papers/yang-2024-buffer-thoughts-thought-augmented-reasoning/),” in The Thirty-eighth Annual Conference on Neural Information Processing Systems, 2024. 24
90. J. Huang, X. Chen, S. Mishra, H. S. Zheng, A. W. Yu, X. Song, and D. Zhou, “[Large language models cannot self-correct reasoning yet](https://md.research.elamur.ai/free/en/papers/huang-2023-large-language-models-cannot-self/),” in The Twelfth International Conference on Learning Representations, ICLR 2024, Vienna, Austria, May 7-11, 2024.
91. S. Welleck, X. Lu, P. West, F. Brahman, T. Shen, D. Khashabi, and Y. Choi, “[Generating sequences by learning to self-correct](https://md.research.elamur.ai/free/en/papers/welleck-2022-generating-sequences-learning-self-correct/),” in The Eleventh International Conference on Learning Representations, ICLR 2023, Kigali, Rwanda, May 1-5, 2023.
92. D. Paul, M. Ismayilzada, M. Peyrard, B. Borges, A. Bosselut, R. West, and B. Faltings, “[REFINER: reasoning feedback on intermediate representations](https://md.research.elamur.ai/free/en/papers/paul-2023-refiner-reasoning-feedback-intermediate-representations/),” in Proceedings of the 18th Conference of the European Chapter of the Association for Computational Linguistics, EACL 2024 - Volume 1: Long Papers, St.…
93. D. Zhang, J. Wu, J. Lei, T. Che, J. Li, T. Xie, X. Huang, S. Zhang, M. Pavone, Y. Li et al., “[Llama-berry: Pairwise optimization for o1-like olympiad-level mathematical reasoning](https://md.research.elamur.ai/free/en/papers/zhang-2024-llama-berry-pairwise-optimization-o1/),” arXiv preprint arXiv:2410.02884, 2024. 24
94. J. LI, E. Beeching, L. Tunstall, B. Lipkin, R. Soletskyi, S. C. Huang, K. Rasul, L. Yu, A. Jiang, Z. Shen, Z. Qin, B. Dong, L. Zhou, Y. Fleureau, G. Lample, and S. Polu, “NuminaMath,” [https://huggingface.co/AI-MO/NuminaMath-CoT](https://github.com/project-numina/aimo-progress-prize/blob/ main/rep… [find](https://scholar.google.com/scholar?q=J%2E+LI%2C+E%2E+Beeching%2C+L%2E+Tunstall%2C+B%2E+Lipkin%2C+R%2E+Soletskyi%2C+S%2E+C%2E+Huang%2C+K%2E+Rasul%2C+L%2E+Yu%2C+A%2E+Jiang%2C+Z%2E+Shen%2C+Z%2E+Qin%2C+B%2E+Dong%2C+L%2E+Zhou%2C+Y%2E+Fleureau%2C+G%2E+Lample%2C+and+S%2E+Polu%2C+%E2%80%9CNuminaMath%2C%E2%80%9D+%5Bhttps%3A%2F%2Fhugging)
95. Qwen, “QwQ: Reflect deeply on the boundaries of the unknown,” November 2024. [Online]. Available: [https://qwenlm.github.io/blog/qwq-32b-preview/](https://qwenlm.github.io/blog/qwq-32b-preview/) 27 [↗](https://qwenlm.github.io/blog/qwq-32b-preview/)
96. D. Kocetkov, R. Li, L. B. allal, J. LI, C. Mou, Y. Jernite, M. Mitchell, C. M. Ferrandis, S. Hughes, T. Wolf, D. Bahdanau, L. V. Werra, and H. de Vries, “The Stack: 3 TB of permissively licensed source code,” Transactions on Machine Learning Research, 2023. 27 [find](https://scholar.google.com/scholar?q=D%2E+Kocetkov%2C+R%2E+Li%2C+L%2E+B%2E+allal%2C+J%2E+LI%2C+C%2E+Mou%2C+Y%2E+Jernite%2C+M%2E+Mitchell%2C+C%2E+M%2E+Ferrandis%2C+S%2E+Hughes%2C+T%2E+Wolf%2C+D%2E+Bahdanau%2C+L%2E+V%2E+Werra%2C+and+H%2E+de+Vries%2C+%E2%80%9CThe+Stack%3A+3+TB+of+permissively+licensed+sourc)
97. Y. Zheng, R. Zhang, J. Zhang, Y. Ye, and Z. Luo, “[LlamaFactory: Unified efficient fine-tuning of 100+ language models](https://md.research.elamur.ai/free/en/papers/zheng-2024-llamafactory-unified-efficient-fine-tuning/),” in Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 3: System Demonstrations), Y. Cao, Y. Feng, and D. Xiong, Eds. Bangkok, Thailand:…
98. J. Hu, X. Wu, Z. Zhu, Xianyu, W. Wang, D. Zhang, and Y. Cao, “[OpenRLHF: An easy-to-use, scalable and high-performance rlhf framework](https://md.research.elamur.ai/free/en/papers/hu-2024-openrlhf-easy-use-scalable/),” arXiv preprint arXiv:2405.11143, 2024. 35
99. P. F. Christiano, J. Leike, T. Brown, M. Martic, S. Legg, and D. Amodei, “Deep reinforcement learning from human preferences,” Advances in neural information processing systems, vol. 30, 2017. 35 [find](https://scholar.google.com/scholar?q=P%2E+F%2E+Christiano%2C+J%2E+Leike%2C+T%2E+Brown%2C+M%2E+Martic%2C+S%2E+Legg%2C+and+D%2E+Amodei%2C+%E2%80%9CDeep+reinforcement+learning+from+human+preferences%2C%E2%80%9D+Advances+in+neural+information+processing+systems%2C+vol%2E+30%2C+2017%2E+35)
100. C. He, R. Luo, Y. Bai, S. Hu, Z. L. Thai, J. Shen, J. Hu, X. Han, Y. Huang, Y. Zhang et al., “[Olympiadbench: A challenging benchmark for promoting agi with olympiad-level bilingual multimodal scientific problems](https://md.research.elamur.ai/free/en/papers/he-2024-olympiadbench-challenging-benchmark-promoting-agi/),” arXiv preprint arXiv:2402.14008, 2024. 36

*Updated 3 November 2026*