---
title: "MEAV — paper by Shahriar et al. (2026) — summary and ideas"
description: "The authors propose MEAV, which treats alignment as model editing in the spirit of Task Arithmetic. 34 references."
authors: ["Sadat Shahriar", "Zheng Qi", "Nikolaos Pappas", "Srikanth Doss", "Kishaloy Halder", "Monica Sunkara", "Manuel Mager", "Yassine Benajiba"]
year: 2026
venue: "Findings of the Association for Computational Linguistics: ACL 2026"
source: "https://aclanthology.org/2026.findings-acl.2035/"
doi: "10.18653/v1/2026.findings-acl.2035"
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/shahriar-2026-meav-model-editing-alignment-vectors/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/shahriar-2026-meav-model-editing-alignment-vectors/)

[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/14/) / MEAV: Model Editing with Alignment Vectors for inference time LLM alignment in single and multidomain preference spectrum

# MEAV: Model Editing with Alignment Vectors for inference time LLM alignment in single and multidomain preference spectrum

**Authors:** Sadat Shahriar, [Zheng Qi](https://md.research.elamur.ai/free/en/authors/zheng-qi/), [Nikolaos Pappas](https://md.research.elamur.ai/free/en/authors/nikolaos-pappas/), Srikanth Doss, Kishaloy Halder, [Monica Sunkara](https://md.research.elamur.ai/free/en/authors/monica-sunkara/), [Manuel Mager](https://md.research.elamur.ai/free/en/authors/manuel-mager/) and [Yassine Benajiba](https://md.research.elamur.ai/free/en/authors/yassine-benajiba/) · **Year:** 2026 · **Venue:** [Findings of the Association for Computational Linguistics: ACL 2026](https://md.research.elamur.ai/free/en/venues/findings/) ([Findings of the Association for Computational Linguistics 2026](https://md.research.elamur.ai/free/en/venues/findings/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.findings-acl.2035/) · [PDF](https://aclanthology.org/2026.findings-acl.2035.pdf) · [DOI 10.18653/v1/2026.findings-acl.2035](https://doi.org/10.18653/v1/2026.findings-acl.2035) · [OpenAlex](https://openalex.org/W7166861016)

**Citations:** 1 citations (OpenAlex, 2026-10-14)

## Summary
The paper addresses the cost of adjusting LLM alignment after training. Methods like RLHF or DPO need retraining for every new preference setting, and many inference-time alternatives rely on a reward model at each step. The authors propose MEAV, which treats alignment as model editing in the spirit of Task Arithmetic. A model is first aligned with DPO using an IPO loss. The difference between its weights and those of the unaligned model is stored as an Alignment Vector (AV). At inference, the AV is added back to the base model, scaled by a coefficient λ. Positive values move responses toward expert-level answers, and negative values push the model toward refusing. To train and test this, the authors built a synthetic English dataset from personas taken from PersonaHub or produced by their CreatePersona pipeline. They used Claude-3-Sonnet to generate queries and three graded responses per query. The method offers cheap, continuous control over response proficiency and supports mixing preferences across several domains without joint retraining.

## Key points
- The dataset has about 38k queries in the Medical, Financial and Legal domains. Each query comes with responses at three proficiency levels: expert, generic and avoidant. Human evaluation was used to check response-level quality.
- Combining per-domain AVs with separate coefficients (α, β, γ) produces mixed behaviors, such as expert legal answers alongside avoidance of medical questions. Compared with joint training, this cuts the number of training runs by a factor of p^D/D and is roughly 12x faster. The authors also report about half the inference cost of a prompt-engineering approach.
- AVs transfer to a safety-aligned version of the same model. When λ ranges from −1 to +1 in the medical domain, safety preference accuracy changes by only 0.11, while medical expert preference accuracy rises by 0.81.
- The main experiments use Mistral-7B-Instruct, and similar behavior appears on Qwen-2.5-3B-Instruct. On Qwen, medical expert accuracy reaches 0.91 at one λ setting, and avoidant accuracy reaches 0.68–0.77 at strongly negative λ.
- The authors list several limitations:
  - The method only works between models with the same architecture and size.
  - Multidomain coefficients require an expensive grid search.
  - Factual correctness of the synthetic responses was not validated at scale.

*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

- [Universidad Iberoamericana](https://md.research.elamur.ai/free/en/universities/universidad-iberoamericana/) — University, Mexico City, Mexico · 2 works · 257 citations

📚 Shelves (4): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 16,667 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 16,194 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 16,039 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 15,891 works

🧭 Topics (1): [🧠 Language models & pretraining](https://md.research.elamur.ai/free/en/topics/language-models/) · 3,320 works

## Cite

Cite as: Sadat Shahriar, Zheng Qi, Nikolaos Pappas, Srikanth Doss, Kishaloy Halder, Monica Sunkara, Manuel Mager and Yassine Benajiba. 2026. MEAV: Model Editing with Alignment Vectors for inference time LLM alignment in single and multidomain preference spectrum. Findings of the Association for Computational Linguistics: ACL 2026.

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

```bibtex
@inproceedings{shahriar2026meav,
  title = {{MEAV: Model Editing with Alignment Vectors for inference time LLM alignment in single and multidomain preference spectrum}},
  author = {Sadat Shahriar, Zheng Qi, Nikolaos Pappas, Srikanth Doss, Kishaloy Halder, Monica Sunkara, Manuel Mager and Yassine Benajiba},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.2035},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.2035}
}
```

- [← previous Measuring Linguistic Competence of LLMs on Indigenous Languages of the Americas](https://md.research.elamur.ai/free/en/papers/vasselli-2026-measuring-linguistic-competence-llms-indigenous/)
- [next → Mechanistic Interpretability of Text-to-Image Diffusion Models via Cross-Attention Interventions](https://md.research.elamur.ai/free/en/papers/maliha-2026-mechanistic-interpretability-text-image-diffusion/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/14/)

## References

34 references, 15 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Anthropic. 2024. Introducing the next generation of claude: The claude 3 family. Accessed: 2024-09-10. [find](https://scholar.google.com/scholar?q=Anthropic%2E+2024%2E+Introducing+the+next+generation+of+claude%3A+The+claude+3+family%2E+Accessed%3A+2024%2D09%2D10%2E)
2. Mohammad Gheshlaghi Azar, Zhaohan Daniel Guo, Bilal Piot, Remi Munos, Mark Rowland, Michal Valko, and Daniele Calandriello. 2024. [A general theoretical paradigm to understand learning from human preferences](https://md.research.elamur.ai/free/en/papers/azar-2023-general-theoretical-paradigm-understand-learning/). pages 4447–4455. PMLR.
3. Jinze Bai, Shuai Bai, Yunfei Chu, Zeyu Cui, Kai Dang, Xiaodong Deng, Yang Fan, Wenbin Ge, Yu Han, Fei Huang, et al. 2023. [Qwen technical report](https://md.research.elamur.ai/free/en/papers/bai-2023-qwen-technical-report/). arXiv preprint arXiv:2309.16609.
4. Yuntao Bai, Andy Jones, Kamal Ndousse, Amanda Askell, Anna Chen, Nova DasSarma, Dawn Drain, Stanislav Fort, Deep Ganguli, Tom Henighan, et al. 2022. Training a helpful and harmless assistant with reinforcement learning from human feedback. arXiv preprint arXiv:2204.05862. [find](https://scholar.google.com/scholar?q=Yuntao+Bai%2C+Andy+Jones%2C+Kamal+Ndousse%2C+Amanda+Askell%2C+Anna+Chen%2C+Nova+DasSarma%2C+Dawn+Drain%2C+Stanislav+Fort%2C+Deep+Ganguli%2C+Tom+Henighan%2C+et+al%2E+2022%2E+Training+a+helpful+and+harmless+assistant+with+rein)
5. Xin Chan, Xiaoyang Wang, Dian Yu, Haitao Mi, and Dong Yu. 2024. [Scaling synthetic data creation with 1,000,000,000 personas](https://md.research.elamur.ai/free/en/papers/ge-2024-scaling-synthetic-data-creation-1/). arXiv preprint arXiv:2406.20094.
6. Jacob Cohen. 1960. A coefficient of agreement for nominal scales. Educational and psychological measurement, 20(1):37–46. [find](https://scholar.google.com/scholar?q=Jacob+Cohen%2E+1960%2E+A+coefficient+of+agreement+for+nominal+scales%2E+Educational+and+psychological+measurement%2C+20%281%29%3A37%E2%80%9346%2E)
7. Yi Dong, Zhilin Wang, Makesh Narsimhan Sreedhar, Xianchao Wu, and Oleksii Kuchaiev. 2023. Steerlm: Attribute conditioned sft as an (user-steerable) alternative to rlhf. arXiv preprint arXiv:2310.05344. [find](https://scholar.google.com/scholar?q=Yi+Dong%2C+Zhilin+Wang%2C+Makesh+Narsimhan+Sreedhar%2C+Xianchao+Wu%2C+and+Oleksii+Kuchaiev%2E+2023%2E+Steerlm%3A+Attribute+conditioned+sft+as+an+%28user%2Dsteerable%29+alternative+to+rlhf%2E+arXiv+preprint+arXiv%3A2310%2E05344)
8. Lei Gao, Yue Niu, Tingting Tang, Salman Avestimehr, and Murali Annavaram. 2024. Ethos: Rectifying language models in orthogonal parameter space. arXiv preprint arXiv:2403.08994. [find](https://scholar.google.com/scholar?q=Lei+Gao%2C+Yue+Niu%2C+Tingting+Tang%2C+Salman+Avestimehr%2C+and+Murali+Annavaram%2E+2024%2E+Ethos%3A+Rectifying+language+models+in+orthogonal+parameter+space%2E+arXiv+preprint+arXiv%3A2403%2E08994%2E)
9. Yiju Guo, Ganqu Cui, Lifan Yuan, Ning Ding, Jiexin Wang, Huimin Chen, Bowen Sun, Ruobing Xie, Jie Zhou, Yankai Lin, et al. 2024. Controllable preference optimization: Toward controllable multi-objective alignment. arXiv preprint arXiv:2402.19085. [find](https://scholar.google.com/scholar?q=Yiju+Guo%2C+Ganqu+Cui%2C+Lifan+Yuan%2C+Ning+Ding%2C+Jiexin+Wang%2C+Huimin+Chen%2C+Bowen+Sun%2C+Ruobing+Xie%2C+Jie+Zhou%2C+Yankai+Lin%2C+et+al%2E+2024%2E+Controllable+preference+optimization%3A+Toward+controllable+multi%2Dobjecti)
10. James Y Huang, Sailik Sengupta, Daniele Bonadiman, Yi-an Lai, Arshit Gupta, Nikolaos Pappas, Saab Mansour, Katrin Kirchoff, and Dan Roth. 2024. [Deal: Decoding-time alignment for large language models](https://md.research.elamur.ai/free/en/papers/huang-2025-deal-decoding-time-alignment-large/). arXiv preprint arXiv:2402.06147.
11. Gabriel Ilharco, Marco Tulio Ribeiro, Mitchell Wortsman, Ludwig Schmidt, Hannaneh Hajishirzi, and Ali Farhadi. 2023. [Editing models with task arithmetic](https://md.research.elamur.ai/free/en/papers/ilharco-2022-editing-models-task-arithmetic/). In The Eleventh International Conference on Learning Representations.
12. Joel Jang, Seungone Kim, Bill Yuchen Lin, Yizhong Wang, Jack Hessel, Luke Zettlemoyer, Hannaneh Hajishirzi, Yejin Choi, and Prithviraj Ammanabrolu. 2023. Personalized soups: Personalized large language model alignment via post-hoc parameter merging. arXiv preprint arXiv:2310.11564. [find](https://scholar.google.com/scholar?q=Joel+Jang%2C+Seungone+Kim%2C+Bill+Yuchen+Lin%2C+Yizhong+Wang%2C+Jack+Hessel%2C+Luke+Zettlemoyer%2C+Hannaneh+Hajishirzi%2C+Yejin+Choi%2C+and+Prithviraj+Ammanabrolu%2E+2023%2E+Personalized+soups%3A+Personalized+large+languag)
13. Jiaming Ji, Mickel Liu, Josef Dai, Xuehai Pan, Chi Zhang, Ce Bian, Boyuan Chen, Ruiyang Sun, Yizhou Wang, and Yaodong Yang. 2024. [Beavertails: Towards improved safety alignment of llm via a human-preference dataset](https://md.research.elamur.ai/free/en/papers/ji-2023-beavertails-towards-improved-safety-alignment/). Advances in Neural Information Processing Systems, 36.
14. Albert Q Jiang, Alexandre Sablayrolles, Arthur Mensch, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Florian Bressand, Gianna Lengyel, Guillaume Lample, Lucile Saulnier, et al. 2023. Mistral 7b. arXiv preprint arXiv:2310.06825. [find](https://scholar.google.com/scholar?q=Albert+Q+Jiang%2C+Alexandre+Sablayrolles%2C+Arthur+Mensch%2C+Chris+Bamford%2C+Devendra+Singh+Chaplot%2C+Diego+de+las+Casas%2C+Florian+Bressand%2C+Gianna+Lengyel%2C+Guillaume+Lample%2C+Lucile+Saulnier%2C+et+al%2E+2023%2E+Mist)
15. Kenneth Li, Oam Patel, Fernanda Viégas, Hanspeter Pfister, and Martin Wattenberg. 2024. [Inference-time intervention: Eliciting truthful answers from a language model](https://md.research.elamur.ai/free/en/papers/li-2023-inference-time-intervention-eliciting-truthful/). Advances in Neural Information Processing Systems, 36.
16. Lei Li, Yongfeng Zhang, and Li Chen. 2023. Prompt distillation for efficient llm-based recommendation. In Proceedings of the 32nd ACM International Conference on Information and Knowledge Management, pages 1348–1357. [find](https://scholar.google.com/scholar?q=Lei+Li%2C+Yongfeng+Zhang%2C+and+Li+Chen%2E+2023%2E+Prompt+distillation+for+efficient+llm%2Dbased+recommendation%2E+In+Proceedings+of+the+32nd+ACM+International+Conference+on+Information+and+Knowledge+Management%2C+)
17. Tianlin Liu, Shangmin Guo, Leonardo Bianco, Daniele Calandriello, Quentin Berthet, Felipe Llinares, Jessica Hoffmann, Lucas Dixon, Michal Valko, and Mathieu Blondel. 2024. Decoding-time realignment of language models. arXiv preprint arXiv:2402.02992. — [arXiv:2402.02992](https://arxiv.org/abs/2402.02992) [find](https://scholar.google.com/scholar?q=Tianlin+Liu%2C+Shangmin+Guo%2C+Leonardo+Bianco%2C+Daniele+Calandriello%2C+Quentin+Berthet%2C+Felipe+Llinares%2C+Jessica+Hoffmann%2C+Lucas+Dixon%2C+Michal+Valko%2C+and+Mathieu+Blondel%2E+2024%2E+Decoding%2Dtime+realignment+of)
18. Bertalan Meskó. 2023. Prompt engineering as an important emerging skill for medical professionals: tutorial. Journal of medical Internet research, 25:e50638. [find](https://scholar.google.com/scholar?q=Bertalan+Mesk%C3%B3%2E+2023%2E+Prompt+engineering+as+an+important+emerging+skill+for+medical+professionals%3A+tutorial%2E+Journal+of+medical+Internet+research%2C+25%3Ae50638%2E)
19. OpenAI. 2025a. [Gpt-4o system card](https://md.research.elamur.ai/free/en/papers/openai-2024-gpt-4o-system-card/). Accessed: 2025-03-28.
20. OpenAI. 2025b. o3-mini system card. Accessed: 2025-03-28. [find](https://scholar.google.com/scholar?q=OpenAI%2E+2025b%2E+o3%2Dmini+system+card%2E+Accessed%3A+2025%2D03%2D28%2E)
21. Jonas Oppenlaender, Rhema Linder, and Johanna Silvennoinen. 2023. Prompting ai art: An investigation into the creative skill of prompt engineering. arXiv preprint arXiv:2303.13534. — [arXiv:2303.13534](https://arxiv.org/abs/2303.13534) [find](https://scholar.google.com/scholar?q=Jonas+Oppenlaender%2C+Rhema+Linder%2C+and+Johanna+Silvennoinen%2E+2023%2E+Prompting+ai+art%3A+An+investigation+into+the+creative+skill+of+prompt+engineering%2E+arXiv+preprint+arXiv%3A2303%2E13534%2E+%E2%80%94+%5BarXiv%3A2303%2E13534)
22. Long Ouyang, Jeffrey Wu, Xu Jiang, Diogo Almeida, Carroll Wainwright, Pamela Mishkin, Chong Zhang, Sandhini Agarwal, Katarina Slama, Alex Ray, et al. 2022. [Training language models to follow instructions with human feedback](https://md.research.elamur.ai/free/en/papers/ouyang-2022-training-language-models-follow-instructions/). Advances in neural information processing systems, 35:27730–27744.
23. Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, Ilya Sutskever, et al. 2019. Language models are unsupervised multitask learners. OpenAI blog, 1(8):9. [find](https://scholar.google.com/scholar?q=Alec+Radford%2C+Jeffrey+Wu%2C+Rewon+Child%2C+David+Luan%2C+Dario+Amodei%2C+Ilya+Sutskever%2C+et+al%2E+2019%2E+Language+models+are+unsupervised+multitask+learners%2E+OpenAI+blog%2C+1%288%29%3A9%2E)
24. Rafael Rafailov, Archit Sharma, Eric Mitchell, Christopher D Manning, Stefano Ermon, and Chelsea Finn. 2024. [Direct preference optimization: Your language model is secretly a reward model](https://md.research.elamur.ai/free/en/papers/rafailov-2023-direct-preference-optimization-language-model/). Advances in Neural Information Processing Systems, 36.
25. Alexandre Rame, Guillaume Couairon, Corentin Dancette, Jean-Baptiste Gaya, Mustafa Shukor, Laure Soulier, and Matthieu Cord. 2023. Rewarded soups: towards pareto-optimal alignment by interpolating weights fine-tuned on diverse rewards. Advances in Neural Information Processing Systems, 36:71095–7113… [find](https://scholar.google.com/scholar?q=Alexandre+Rame%2C+Guillaume+Couairon%2C+Corentin+Dancette%2C+Jean%2DBaptiste+Gaya%2C+Mustafa+Shukor%2C+Laure+Soulier%2C+and+Matthieu+Cord%2E+2023%2E+Rewarded+soups%3A+towards+pareto%2Doptimal+alignment+by+interpolating+wei)
26. Pranab Sahoo, Ayush Kumar Singh, Sriparna Saha, Vinija Jain, Samrat Mondal, and Aman Chadha. 2024. [A systematic survey of prompt engineering in large language models: Techniques and applications](https://md.research.elamur.ai/free/en/papers/sahoo-2024-systematic-survey-prompt-engineering-large/). arXiv preprint arXiv:2402.07927.
27. Nisan Stiennon, Long Ouyang, Jeffrey Wu, Daniel Ziegler, Ryan Lowe, Chelsea Voss, Alec Radford, Dario Amodei, and Paul F Christiano. 2020. Learning to summarize with human feedback. Advances in Neural Information Processing Systems, 33:3008– 3021. [find](https://scholar.google.com/scholar?q=Nisan+Stiennon%2C+Long+Ouyang%2C+Jeffrey+Wu%2C+Daniel+Ziegler%2C+Ryan+Lowe%2C+Chelsea+Voss%2C+Alec+Radford%2C+Dario+Amodei%2C+and+Paul+F+Christiano%2E+2020%2E+Learning+to+summarize+with+human+feedback%2E+Advances+in+Neural)
28. Hugo Touvron, Thibaut Lavril, Gautier Izacard, Xavier Martinet, Marie-Anne Lachaux, Timothée Lacroix, Baptiste Rozière, Naman Goyal, Eric Hambro, Faisal Azhar, et al. 2023. [Llama: Open and efficient foundation language models](https://md.research.elamur.ai/free/en/papers/touvron-2023-llama-open-efficient-foundation-language/). arXiv preprint arXiv:2302.13971.
29. Vectara. 2025. Hallucination evaluation leaderboard. Hugging Face Spaces. Accessed: 2025-02-15. [find](https://scholar.google.com/scholar?q=Vectara%2E+2025%2E+Hallucination+evaluation+leaderboard%2E+Hugging+Face+Spaces%2E+Accessed%3A+2025%2D02%2D15%2E)
30. Pengyu Wang, Dong Zhang, Linyang Li, Chenkun Tan, Xinghao Wang, Ke Ren, Botian Jiang, and Xipeng Qiu. 2024. Inferaligner: Inference-time alignment for harmlessness through cross-model guidance. arXiv preprint arXiv:2401.11206. [find](https://scholar.google.com/scholar?q=Pengyu+Wang%2C+Dong+Zhang%2C+Linyang+Li%2C+Chenkun+Tan%2C+Xinghao+Wang%2C+Ke+Ren%2C+Botian+Jiang%2C+and+Xipeng+Qiu%2E+2024%2E+Inferaligner%3A+Inference%2Dtime+alignment+for+harmlessness+through+cross%2Dmodel+guidance%2E+arXiv+)
31. Prateek Yadav, Derek Tam, Leshem Choshen, Colin A Raffel, and Mohit Bansal. 2024. [Ties-merging: Resolving interference when merging models](https://md.research.elamur.ai/free/en/papers/yadav-2023-ties-merging-resolving-interference-merging/). Advances in Neural Information Processing Systems, 36.
32. Rui Yang, Xiaoman Pan, Feng Luo, Shuang Qiu, Han Zhong, Dong Yu, and Jianshu Chen. 2024. Rewards-in-context: Multi-objective alignment of foundation models with dynamic preference adjustment. arXiv preprint arXiv:2402.10207. [find](https://scholar.google.com/scholar?q=Rui+Yang%2C+Xiaoman+Pan%2C+Feng+Luo%2C+Shuang+Qiu%2C+Han+Zhong%2C+Dong+Yu%2C+and+Jianshu+Chen%2E+2024%2E+Rewards%2Din%2Dcontext%3A+Multi%2Dobjective+alignment+of+foundation+models+with+dynamic+preference+adjustment%2E+arXiv+pr)
33. Le Yu, Bowen Yu, Haiyang Yu, Fei Huang, and Yongbin Li. 2024. [Language models are super mario: Absorbing abilities from homologous models as a free lunch](https://md.research.elamur.ai/free/en/papers/yu-2023-language-models-super-mario-absorbing/). In Forty-first International Conference on Machine Learning.
34. Lianmin Zheng, Wei-Lin Chiang, Ying Sheng, Siyuan Zhuang, Zhanghao Wu, Yonghao Zhuang, Zi Lin, Zhuohan Li, Dacheng Li, Eric Xing, et al. 2024. [Judging llm-as-a-judge with mt-bench and chatbot arena](https://md.research.elamur.ai/free/en/papers/zheng-2023-judging-llm-judge-mt-bench/). Advances in Neural Information Processing Systems, 36.

*Updated 1 November 2026*