---
title: "PILOT — paper by Zheng et al. (2026) — summary and ideas"
description: "Supervision comes from a Construct-and-Verify pipeline. 37 references. Among them: 0."
authors: ["Haoyu Zheng", "Yun Zhu", "Yuqian Yuan", "Bo Yuan", "Wenqiao Zhang", "Siliang Tang", "Jun Xiao"]
year: 2026
venue: "Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)"
source: "https://aclanthology.org/2026.acl-long.236/"
doi: "10.18653/v1/2026.acl-long.236"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/zheng-2026-pilot-planning-internalized-latent-optimization/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/zheng-2026-pilot-planning-internalized-latent-optimization/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/45/) / PILOT: Planning via Internalized Latent Optimization Trajectories for Large Language Models

# PILOT: Planning via Internalized Latent Optimization Trajectories for Large Language Models

**Authors:** Haoyu Zheng, [Yun Zhu](https://md.research.elamur.ai/free/en/authors/yun-zhu/), Yuqian Yuan, [Bo Yuan](https://md.research.elamur.ai/free/en/authors/bo-yuan/), Wenqiao Zhang, [Siliang Tang](https://md.research.elamur.ai/free/en/authors/siliang-tang/) and [Jun Xiao](https://md.research.elamur.ai/free/en/authors/jun-xiao/) · **Year:** 2026 · **Venue:** [Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)](https://md.research.elamur.ai/free/en/venues/acl/) ([ACL 2026](https://md.research.elamur.ai/free/en/venues/acl/2026/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2026.acl-long.236/) · [PDF](https://aclanthology.org/2026.acl-long.236.pdf) · [DOI 10.18653/v1/2026.acl-long.236](https://doi.org/10.18653/v1/2026.acl-long.236) · [OpenAlex](https://openalex.org/W7166838476)

**Citations:** 0 citations (OpenAlex, 2026-10-14)

## Summary
Compact LLMs often have enough domain knowledge for multi-step math and coding problems, but they lack a global strategy, so early mistakes snowball across long reasoning chains. The authors show that such models improve markedly when a stronger teacher supplies an explicit plan. Querying a teacher at inference time, however, is slow and not always possible. PILOT moves this guidance inside the model. A lightweight Hyper-Network reads the query and generates an instance-specific latent anchor vector. This vector is injected, via energy-aligned injection, into a frozen backbone at a chosen pivot layer, so no backbone weights change. Supervision comes from a Construct-and-Verify pipeline. Expert heuristic guidance is written for problems the base model fails. Hints that reveal the answer on their own are filtered out with a blind test. The target vector is the pivot-layer hidden state mean-pooled over the guidance tokens. On Qwen2.5 models, PILOT beats strong baselines on mathematical and coding benchmarks, gaining up to +8.9% on MATH500 with almost no added latency. This offers a cheap, non-invasive way to stabilize reasoning in small models.

## Key points
- **Latency:** On Qwen2.5-7B, total latency rises only about 0.2% (10,230 ms vs. 10,209 ms), compared with roughly +6.2% for Soft CoT.
- **Parameter cost:** PILOT adds 38.6M trainable parameters to Qwen2.5-7B (~0.5%) and 7.1M to Qwen2.5-1.5B (~0.46%).
- **Accuracy (Pass@1):**
  - Qwen2.5-1.5B reaches 52.08 on MATH and 56.34 on HumanEval.
  - Qwen2.5-7B reaches 75.24 on MATH and 77.44 on HumanEval.
- **Ablations:**
  - Math depends mostly on the Hyper-Network; a static variant drops to 47.72 on 1.5B.
  - Code depends heavily on energy alignment.
  - Training on unfiltered data lowers the 1.5B model to 50.85 (Math) and 52.20 (Code).
- **Injection depth:** The best pivot layer differs by domain, so it is tuned manually per domain, which the authors list as a limitation.
  - Math uses layer 26, which works as a late correction.
  - Code uses layer 20, which causes a sharp representation shift followed by recovery.

*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

- [Zhejiang University](https://md.research.elamur.ai/free/en/universities/zhejiang-university/) — University, Hangzhou, China · 224 works · 13,587 citations
- [Shanghai Artificial Intelligence Laboratory](https://md.research.elamur.ai/free/en/labs/shanghai-artificial-intelligence-laboratory/) — Research institute, Shanghai, China · 152 works · 13,166 citations

📚 Shelves (4): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 18,614 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 17,965 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 17,450 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 17,296 works

🧭 Topics (1): [🧠 Language models & pretraining](https://md.research.elamur.ai/free/en/topics/language-models/) · 3,894 works

## Cite

Cite as: Haoyu Zheng, Yun Zhu, Yuqian Yuan, Bo Yuan, Wenqiao Zhang, Siliang Tang and Jun Xiao. 2026. PILOT: Planning via Internalized Latent Optimization Trajectories for Large Language Models. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers).

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

```bibtex
@inproceedings{zheng2026pilot,
  title = {{PILOT: Planning via Internalized Latent Optimization Trajectories for Large Language Models}},
  author = {Haoyu Zheng, Yun Zhu, Yuqian Yuan, Bo Yuan, Wenqiao Zhang, Siliang Tang and Jun Xiao},
  booktitle = {Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2026},
  doi = {10.18653/v1/2026.acl-long.236},
  url = {https://doi.org/10.18653/v1/2026.acl-long.236}
}
```

- [← previous PII-VisBench: Evaluating Personally Identifiable Information Safety in Vision Language Models Along a Continuum of Visibility](https://md.research.elamur.ai/free/en/papers/shahariar-2026-pii-visbench-evaluating-personally-identifiable/)
- [next → Pioneering Bot Detection on Polish Reddit at the Comment Level](https://md.research.elamur.ai/free/en/papers/matyjaszek-2026-pioneering-bot-detection-polish-reddit/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/45/)

## References

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

1. Jacob Austin, Augustus Odena, Maxwell Nye, Maarten Bosma, Henryk Michalewski, David Dohan, Ellen Jiang, Carrie Cai, Michael Terry, Quoc Le, and 1 others. 2021. [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.
2. Mark Chen, Jerry Tworek, Heewoo Jun, Qiming Yuan, Henrique Ponde de Oliveira Pinto, Jared Kaplan, Harri Edwards, Yuri Burda, Nicholas Joseph, Greg Brockman, Alex Ray, Raul Puri, Gretchen Krueger, Michael Petrov, Heidy Khlaaf, Girish Sastry, Pamela Mishkin, Brooke Chan, Scott Gray, and 39 others. 202… — [Evaluating Large Language Models Trained on Code](https://md.research.elamur.ai/free/en/papers/chen-2021-evaluating-large-language-models-trained/)
3. Younwoo Choi, Muhammad Adil Asif, Ziwen Han, John Willes, and Rahul G Krishnan. 2025. Teaching llms how to learn with contextual fine-tuning. arXiv preprint arXiv:2503.09032. — [arXiv:2503.09032](https://arxiv.org/abs/2503.09032) [find](https://scholar.google.com/scholar?q=Younwoo+Choi%2C+Muhammad+Adil+Asif%2C+Ziwen+Han%2C+John+Willes%2C+and+Rahul+G+Krishnan%2E+2025%2E+Teaching+llms+how+to+learn+with+contextual+fine%2Dtuning%2E+arXiv+preprint+arXiv%3A2503%2E09032%2E+%E2%80%94+%5BarXiv%3A2503%2E09032%5D%28http)
4. Karl Cobbe, Vineet Kosaraju, Mohammad Bavarian, Mark Chen, Heewoo Jun, Lukasz Kaiser, Matthias Plappert, Jerry Tworek, Jacob Hilton, Reiichiro Nakano, Christopher Hesse, and John Schulman. 2021. [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.
5. Nouha Dziri, Ximing Lu, Melanie Sclar, Xiang Lorraine Li, Liwei Jiang, Bill Yuchen Lin, Sean Welleck, Peter West, Chandra Bhagavatula, Ronan Le Bras, and 1 others. 2023. [Faith and fate: Limits of transformers on compositionality](https://md.research.elamur.ai/free/en/papers/dziri-2023-faith-fate-limits-transformers-compositionality/). Advances in Neural Information Processing Systems, 36:70293–70332.
6. Sachin Goyal, Ziwei Ji, Ankit Singh Rawat, Aditya Krishna Menon, Sanjiv Kumar, and Vaishnavh Nagarajan. 2023. Think before you speak: Training language models with pause tokens. arXiv preprint arXiv:2310.02226. [find](https://scholar.google.com/scholar?q=Sachin+Goyal%2C+Ziwei+Ji%2C+Ankit+Singh+Rawat%2C+Aditya+Krishna+Menon%2C+Sanjiv+Kumar%2C+and+Vaishnavh+Nagarajan%2E+2023%2E+Think+before+you+speak%3A+Training+language+models+with+pause+tokens%2E+arXiv+preprint+arXiv%3A2)
7. Aaron Grattafiori, Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Abhishek Kadian, Ahmad Al-Dahle, Aiesha Letman, Akhil Mathur, Alan Schelten, Alex Vaughan, and 1 others. 2024. [The llama 3 herd of models](https://md.research.elamur.ai/free/en/papers/grattafiori-2024-llama-3-herd-models/). arXiv preprint arXiv:2407.21783.
8. David Ha, Andrew Dai, and Quoc V Le. 2016. Hypernetworks. arXiv preprint arXiv:1609.09106. [find](https://scholar.google.com/scholar?q=David+Ha%2C+Andrew+Dai%2C+and+Quoc+V+Le%2E+2016%2E+Hypernetworks%2E+arXiv+preprint+arXiv%3A1609%2E09106%2E)
9. Shibo Hao, Sainbayar Sukhbaatar, DiJia Su, Xian Li, Zhiting Hu, Jason Weston, and Yuandong Tian. 2024. Training large language models to reason in a continuous latent space. arXiv preprint arXiv:2412.06769. [find](https://scholar.google.com/scholar?q=Shibo+Hao%2C+Sainbayar+Sukhbaatar%2C+DiJia+Su%2C+Xian+Li%2C+Zhiting+Hu%2C+Jason+Weston%2C+and+Yuandong+Tian%2E+2024%2E+Training+large+language+models+to+reason+in+a+continuous+latent+space%2E+arXiv+preprint+arXiv%3A2412%2E)
10. 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](https://md.research.elamur.ai/free/en/papers/hendrycks-2021-measuring-mathematical-problem-solving-math/). NeurIPS.
11. Edward J Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, Weizhu Chen, and 1 others. 2022. [Lora: Low-rank adaptation of large language models](https://md.research.elamur.ai/free/en/papers/hu-2021-lora-low-rank-adaptation-large/). ICLR, 1(2):3.
12. Ceyhun Efe Kayan and Li Zhang. 2025. Prototype-based dynamic steering for large language models. arXiv preprint arXiv:2510.05498. — [arXiv:2510.05498](https://arxiv.org/abs/2510.05498) [find](https://scholar.google.com/scholar?q=Ceyhun+Efe+Kayan+and+Li+Zhang%2E+2025%2E+Prototype%2Dbased+dynamic+steering+for+large+language+models%2E+arXiv+preprint+arXiv%3A2510%2E05498%2E+%E2%80%94+%5BarXiv%3A2510%2E05498%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F2510%2E05498%29)
13. James Kirkpatrick, Razvan Pascanu, Neil Rabinowitz, Joel Veness, Guillaume Desjardins, Andrei A Rusu, Kieran Milan, John Quan, Tiago Ramalho, Agnieszka Grabska-Barwinska, and 1 others. 2017. Overcoming catastrophic forgetting in neural networks. Proceedings of the national academy of sciences, 114(1… [find](https://scholar.google.com/scholar?q=James+Kirkpatrick%2C+Razvan+Pascanu%2C+Neil+Rabinowitz%2C+Joel+Veness%2C+Guillaume+Desjardins%2C+Andrei+A+Rusu%2C+Kieran+Milan%2C+John+Quan%2C+Tiago+Ramalho%2C+Agnieszka+Grabska%2DBarwinska%2C+and+1+others%2E+2017%2E+Overcomin)
14. Xiang Lisa Li and Percy Liang. 2021. Prefix-tuning: Optimizing continuous prompts for generation. arXiv preprint arXiv:2101.00190. [find](https://scholar.google.com/scholar?q=Xiang+Lisa+Li+and+Percy+Liang%2E+2021%2E+Prefix%2Dtuning%3A+Optimizing+continuous+prompts+for+generation%2E+arXiv+preprint+arXiv%3A2101%2E00190%2E)
15. Yun Luo, Zhen Yang, Fandong Meng, Yafu Li, Jie Zhou, and Yue Zhang. 2025. An empirical study of catastrophic forgetting in large language models during continual fine-tuning. IEEE Transactions on Audio, Speech and Language Processing. [find](https://scholar.google.com/scholar?q=Yun+Luo%2C+Zhen+Yang%2C+Fandong+Meng%2C+Yafu+Li%2C+Jie+Zhou%2C+and+Yue+Zhang%2E+2025%2E+An+empirical+study+of+catastrophic+forgetting+in+large+language+models+during+continual+fine%2Dtuning%2E+IEEE+Transactions+on+Audi)
16. Nina Panickssery, Nick Gabrieli, Julian Schulz, Meg Tong, Evan Hubinger, and Alexander Matt Turner. 2024. Steering llama 2 via contrastive activation addition. Preprint, arXiv:2312.06681. [find](https://scholar.google.com/scholar?q=Nina+Panickssery%2C+Nick+Gabrieli%2C+Julian+Schulz%2C+Meg+Tong%2C+Evan+Hubinger%2C+and+Alexander+Matt+Turner%2E+2024%2E+Steering+llama+2+via+contrastive+activation+addition%2E+Preprint%2C+arXiv%3A2312%2E06681%2E)
17. DiJia Su, Hanlin Zhu, Yingchen Xu, Jiantao Jiao, Yuandong Tian, and Qinqing Zheng. 2025. Token assorted: Mixing latent and text tokens for improved language model reasoning. arXiv preprint arXiv:2502.03275. — [arXiv:2502.03275](https://arxiv.org/abs/2502.03275) [find](https://scholar.google.com/scholar?q=DiJia+Su%2C+Hanlin+Zhu%2C+Yingchen+Xu%2C+Jiantao+Jiao%2C+Yuandong+Tian%2C+and+Qinqing+Zheng%2E+2025%2E+Token+assorted%3A+Mixing+latent+and+text+tokens+for+improved+language+model+reasoning%2E+arXiv+preprint+arXiv%3A2502%2E)
18. Tianci Xue, Ziqi Wang, Zhenhailong Wang, Chi Han, Pengfei Yu, and Heng Ji. 2023. Rcot: Detecting and rectifying factual inconsistency in reasoning by reversing chain-of-thought. arXiv preprint arXiv:2305.11499. — [arXiv:2305.11499](https://arxiv.org/abs/2305.11499) [find](https://scholar.google.com/scholar?q=Tianci+Xue%2C+Ziqi+Wang%2C+Zhenhailong+Wang%2C+Chi+Han%2C+Pengfei+Yu%2C+and+Heng+Ji%2E+2023%2E+Rcot%3A+Detecting+and+rectifying+factual+inconsistency+in+reasoning+by+reversing+chain%2Dof%2Dthought%2E+arXiv+preprint+arXiv%3A2)
19. Qi Sun, Edoardo Cetin, and Yujin Tang. 2025. Transformer-squared: Self-adaptive llms. arXiv preprint arXiv:2501.06252. — [arXiv:2501.06252](https://arxiv.org/abs/2501.06252) [find](https://scholar.google.com/scholar?q=Qi+Sun%2C+Edoardo+Cetin%2C+and+Yujin+Tang%2E+2025%2E+Transformer%2Dsquared%3A+Self%2Dadaptive+llms%2E+arXiv+preprint+arXiv%3A2501%2E06252%2E+%E2%80%94+%5BarXiv%3A2501%2E06252%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F2501%2E06252%29)
20. An Yang, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chengyuan Li, Dayiheng Liu, Fei Huang, Haoran Wei, Huan Lin, Jian Yang, Jianhong Tu, Jianwei Zhang, Jianxin Yang, Jiaxi Yang, Jingren Zhou, Junyang Lin, Kai Dang, and 22 others. 2024. Qwen2.5 technical report. arXiv preprint arXi… [find](https://scholar.google.com/scholar?q=An+Yang%2C+Baosong+Yang%2C+Beichen+Zhang%2C+Binyuan+Hui%2C+Bo+Zheng%2C+Bowen+Yu%2C+Chengyuan+Li%2C+Dayiheng+Liu%2C+Fei+Huang%2C+Haoran+Wei%2C+Huan+Lin%2C+Jian+Yang%2C+Jianhong+Tu%2C+Jianwei+Zhang%2C+Jianxin+Yang%2C+Jiaxi+Yang%2C+Jin)
21. Xinyu Tang, Xiaolei Wang, Zhihao Lv, Yingqian Min, Wayne Xin Zhao, Binbin Hu, Ziqi Liu, and Zhiqiang Zhang. 2025. [Unlocking general long chain-of-thought reasoning capabilities of large language models via representation engineering](https://md.research.elamur.ai/free/en/papers/tang-2025-unlocking-general-long-chain-thought/). arXiv preprint arXiv:2503.11314.
22. AI-MO Team. 2024. Aimo validation amc dataset. [https://huggingface.co/datasets/](https://huggingface.co/datasets/) AI-MO/aimo-validation-amc. Accessed: 2024-05-22. [↗](https://huggingface.co/datasets/)
23. Alexander Matt Turner, Lisa Thiergart, Gavin Leech, David Udell, Juan J Vazquez, Ulisse Mini, and Monte MacDiarmid. 2023. [Steering language models with activation engineering](https://md.research.elamur.ai/free/en/papers/turner-2023-steering-language-models-activation-engineering/). arXiv preprint arXiv:2308.10248.
24. Miles Turpin, Julian Michael, Ethan Perez, and Samuel Bowman. 2023. [Language models don’t always say what they think: Unfaithful explanations in chain-of-thought prompting](https://md.research.elamur.ai/free/en/papers/turpin-2023-language-models-don-t-always/). Advances in Neural Information Processing Systems, 36:74952–74965.
25. Constantin Venhoff, Iván Arcuschin, Philip Torr, Arthur Conmy, and Neel Nanda. 2025. Understanding reasoning in thinking language models via steering vectors. arXiv preprint arXiv:2506.18167. — [arXiv:2506.18167](https://arxiv.org/abs/2506.18167) [find](https://scholar.google.com/scholar?q=Constantin+Venhoff%2C+Iv%C3%A1n+Arcuschin%2C+Philip+Torr%2C+Arthur+Conmy%2C+and+Neel+Nanda%2E+2025%2E+Understanding+reasoning+in+thinking+language+models+via+steering+vectors%2E+arXiv+preprint+arXiv%3A2506%2E18167%2E+%E2%80%94+%5BarXiv)
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](https://md.research.elamur.ai/free/en/papers/wang-2022-self-consistency-improves-chain-thought/). arXiv preprint arXiv:2203.11171.
27. Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Fei Xia, Ed Chi, Quoc V Le, Denny Zhou, and 1 others. 2022. [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, 35:24824– 24837.
28. Zhengxuan Wu, Aryaman Arora, Zheng Wang, Atticus Geiger, Dan Jurafsky, Christopher D Manning, and Christopher Potts. 2024a. [Reft: Representation finetuning for language models](https://md.research.elamur.ai/free/en/papers/wu-2024-reft-representation-finetuning-language-models/). Advances in Neural Information Processing Systems, 37:63908–63962.
29. Zhenyu Wu, Meng Jiang, and Chao Shen. 2024b. Get an a in math: Progressive rectification prompting. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, pages 19288–19296. [find](https://scholar.google.com/scholar?q=Zhenyu+Wu%2C+Meng+Jiang%2C+and+Chao+Shen%2E+2024b%2E+Get+an+a+in+math%3A+Progressive+rectification+prompting%2E+In+Proceedings+of+the+AAAI+Conference+on+Artificial+Intelligence%2C+volume+38%2C+pages+19288%E2%80%9319296%2E)
30. Shunyu Yao, Dian Yu, Jeffrey Zhao, Izhak Shafran, Tom Griffiths, Yuan Cao, and Karthik Narasimhan. 2023. Tree of thoughts: Deliberate problem solving with large language models. Advances in neural information processing systems, 36:11809–11822. [find](https://scholar.google.com/scholar?q=Shunyu+Yao%2C+Dian+Yu%2C+Jeffrey+Zhao%2C+Izhak+Shafran%2C+Tom+Griffiths%2C+Yuan+Cao%2C+and+Karthik+Narasimhan%2E+2023%2E+Tree+of+thoughts%3A+Deliberate+problem+solving+with+large+language+models%2E+Advances+in+neural+inf)
31. Eric Zelikman, Georges Harik, Yijia Shao, Varuna Jayasiri, Nick Haber, and Noah D Goodman. 2024. [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/). arXiv preprint arXiv:2403.09629.
32. Dongxu Zhang, Ning Yang, Jihua Zhu, Jinnan Yang, Miao Xin, and Baoliang Tian. 2025. Ascot: An adaptive self-correction chain-of-thought method for late-stage fragility in llms. arXiv preprint arXiv:2508.05282. — [arXiv:2508.05282](https://arxiv.org/abs/2508.05282) [find](https://scholar.google.com/scholar?q=Dongxu+Zhang%2C+Ning+Yang%2C+Jihua+Zhu%2C+Jinnan+Yang%2C+Miao+Xin%2C+and+Baoliang+Tian%2E+2025%2E+Ascot%3A+An+adaptive+self%2Dcorrection+chain%2Dof%2Dthought+method+for+late%2Dstage+fragility+in+llms%2E+arXiv+preprint+arXiv%3A25)
33. Yu Zhao, Alessio Devoto, Giwon Hong, Xiaotang Du, Aryo Pradipta Gema, Hongru Wang, Xuanli He, Kam-Fai Wong, and Pasquale Minervini. 2025. Steering knowledge selection behaviours in llms via sae-based representation engineering. In Proceedings of the 2025 Conference of the Nations of the Americas Cha… [find](https://scholar.google.com/scholar?q=Yu+Zhao%2C+Alessio+Devoto%2C+Giwon+Hong%2C+Xiaotang+Du%2C+Aryo+Pradipta+Gema%2C+Hongru+Wang%2C+Xuanli+He%2C+Kam%2DFai+Wong%2C+and+Pasquale+Minervini%2E+2025%2E+Steering+knowledge+selection+behaviours+in+llms+via+sae%2Dbased+)
34. Yufa Zhou, Yixiao Wang, Xunjian Yin, Shuyan Zhou, and Anru R Zhang. 2025. The geometry of reasoning: Flowing logics in representation space. arXiv preprint arXiv:2510.09782. — [arXiv:2510.09782](https://arxiv.org/abs/2510.09782) [find](https://scholar.google.com/scholar?q=Yufa+Zhou%2C+Yixiao+Wang%2C+Xunjian+Yin%2C+Shuyan+Zhou%2C+and+Anru+R+Zhang%2E+2025%2E+The+geometry+of+reasoning%3A+Flowing+logics+in+representation+space%2E+arXiv+preprint+arXiv%3A2510%2E09782%2E+%E2%80%94+%5BarXiv%3A2510%2E09782%5D%28https)
35. Rui-Jie Zhu, Tianhao Peng, Tianhao Cheng, Xingwei Qu, Jinfa Huang, Dawei Zhu, Hao Wang, Kaiwen Xue, Xuanliang Zhang, Yong Shan, and 1 others. 2025. [A survey on latent reasoning](https://md.research.elamur.ai/free/en/papers/zhu-2025-survey-latent-reasoning/). arXiv preprint arXiv:2507.06203.
36. Andy Zou, Long Phan, Sarah Chen, James Campbell, Phillip Guo, Richard Ren, Alexander Pan, Xuwang Yin, Mantas Mazeika, Ann-Kathrin Dombrowski, and 1 others. 2023. [Representation engineering: A top-down approach to ai transparency](https://md.research.elamur.ai/free/en/papers/zou-2023-representation-engineering-top-down-approach/). arXiv preprint arXiv:2310.01405.
37. Yige Xu, Xu Guo, Zhiwei Zeng, and Chunyan Miao. 2025. [Softcot: Soft chain-of-thought for efficient reasoning with llms](https://md.research.elamur.ai/free/en/papers/xu-2025-softcot-soft-chain-thought-efficient/). arXiv preprint arXiv:2502.12134.

*Updated 2 November 2026*