Elamur Research⌕
☰
home / papers / 2026 / Beyond Static Personas: Situational Personality Steering for Large Language Models

Beyond Static Personas: Situational Personality Steering for Large Language Models

Paper · 2026 · Zesheng Wei, Mengxiang Li, Zilei Wang et al.

Summary

Most LLM personalization methods treat personality as a fixed trait, even though psychology (e.g., CAPS theory) holds that behavior depends on both the person and the situation. This paper first checks whether LLMs show this kind of situation dependence internally. The authors locate "persona neurons" in the FFN layers of Llama-3-8B-Instruct and Qwen3-8B using PersonalityBench prompts sorted into 30 UltraChat-style topics. They find that topic-driven shifts in these neurons are as large as differences between Big-Five domains, and that similar topics produce similar activation patterns. Based on this, they propose IRIS, a training-free Identify-Retrieve-Steer framework. It builds a bank of topic-specific persona neurons, retrieves the ones relevant to a new question through similarity scores, and steers activations with similarity-based weights. Tested on PersonalityBench and the new SPBench benchmark, IRIS beats the strongest baselines, handles unseen and complex situations, and works across architectures. This offers a cheap and interpretable route to context-sensitive personas for role-play and simulation.

Key points

  • Personality is modeled with the Big-Five framework (O, C, E, A, N), each domain having positive and negative aspects. Neuron identification uses PersonalityBench questions split into M = 30 topics following the UltraChat taxonomy.
  • In the PCA analysis, the largest distance between topics within one domain is comparable to the distance between separate personality domains. Per-layer persona-neuron proportions differ by up to 2.23% (layer 16), and middle layers hold the most persona neurons.
  • IRIS has three stages: situational persona neuron identification, situation-aware retrieval via cosine similarity with softmax weighting, and similarity-weighted steering, which stimulates target-aspect neurons and deactivates opposing ones.
  • SPBench is a new synthetic, LLM-generated situational personality benchmark that contains no personal data. Models evaluated include Llama-3-8B-Instruct, Qwen3-8B and gemma-3-12b-it, with RoBERTa-large used for verification.
  • Reported limitations:
    • Steering can hurt instruction following; for example, low Openness may lead to refusals.
    • The fixed 30-topic taxonomy restricts how finely situations can be distinguished.

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

🎓University of Science and Technology of ChinaHefei · China108 works · 112,698 citations🎓Singapore Management UniversitySingapore · Singapore27 works · 2,382 citations

 Fields, topics and concepts

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 works

🧭 Topics 1Research subjects and languages the work is about.Language models & pretraining · 3,320 works

Cite

Cite as Zesheng Wei, Mengxiang Li, Zilei Wang and Yang Deng. 2026. Beyond Static Personas: Situational Personality Steering for Large Language Models. Findings of the Association for Computational Linguistics: ACL 2026.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.findings-acl.958

@inproceedings{wei2026beyond,
  title = {{Beyond Static Personas: Situational Personality Steering for Large Language Models}},
  author = {Zesheng Wei, Mengxiang Li, Zilei Wang and Yang Deng},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.958},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.958}
}

References

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

  1. Abien Fred Agarap. 2019. Deep learning using rectified linear units (relu). Preprint, arXiv:1803.08375. — [arXiv:1803.08375](https://arxiv.org/abs/1803.08375) find
  2. Özlem Ayduk and Anett Gyurak. 2008. Applying the cognitive-affective processing systems approach to conceptualizing rejection sensitivity. Social and personality psychology compass, 2(5):2016–2033. find
  3. Mengyu Bu, Shaolei Zhang, Zhongjun He, Hua Wu, and Yang Feng. 2025. AlignX: Advancing multilingual large language models with multilingual representation alignment. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, pages 6471–6500, Suzhou, China. Association… find
  4. Nuo Chen, Yang Deng, and Jia Li. 2024a. The oscars of AI theater: A survey on role-playing with language models. CoRR, abs/2407.11484. — [arXiv:2407.11484](https://arxiv.org/abs/2407.11484) find
  5. Runjin Chen, Andy Arditi, Henry Sleight, Owain Evans, and Jack Lindsey. 2025. Persona vectors: Monitoring and controlling character traits in language models. Preprint, arXiv:2507.21509. — [arXiv:2507.21509](https://arxiv.org/abs/2507.21509) find
  6. Tiantian Chen, Jiaqi Lu, Ying Shen, and Lin Zhang. 2026. Es-memeval: Benchmarking conversational agents on personalized long-term emotional support. In Proceedings of the ACM Web Conference 2026, WWW 2026, Dubai, United Arab Emirates, originally scheduled for April 13-17, 2026, rescheduled for June… find
  7. Weize Chen, Yusheng Su, Jingwei Zuo, Cheng Yang, Chenfei Yuan, Chi-Min Chan, Heyang Yu, Yaxi Lu, Yi-Hsin Hung, Chen Qian, Yujia Qin, Xin Cong, Ruobing Xie, Zhiyuan Liu, Maosong Sun, and Jie Zhou. 2024b. Agentverse: Facilitating multi-agent collaboration and exploring emergent behaviors. In Internati…
  8. 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. Preprint, arXiv:2110.14168.
  9. Damai Dai, Li Dong, Yaru Hao, Zhifang Sui, Baobao Chang, and Furu Wei. 2022. Knowledge neurons in pretrained transformers. Preprint, arXiv:2104.08696. find
  10. Pedro Henrique Luz de Araujo, Michael A. Hedderich, Ali Modarressi, Hinrich Schuetze, and Benjamin Roth. 2025. Persistent personas? role-playing, instruction following, and safety in extended interactions. Preprint, arXiv:2512.12775. find
  11. Gordon Willard Allport. 1937. Personality: A psychological interpretation. find
  12. Jia Deng, Tianyi Tang, Yanbin Yin, Wenhao Yang, Wayne Xin Zhao, and Ji-Rong Wen. 2024a. Neuron-based personality trait induction in large language models. Preprint, arXiv:2410.12327. 1 https://huggingface.co/meta-llama/Meta-Llama-3-8B/blob/main/LICENSE 2 https://huggingface.co/Qwen/Qwen3-8B/blob/mai… — [arXiv:2410.12327](https://arxiv.org/abs/2410.12327) ↗
  13. Yang Deng, Lizi Liao, Zhonghua Zheng, Grace Hui Yang, and Tat-Seng Chua. 2024b. Towards human-centered proactive conversational agents. In Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR 2024, Washington DC, USA, July 14-18, 2024… find
  14. Ning Ding, Yulin Chen, Bokai Xu, Yujia Qin, Shengding Hu, Zhiyuan Liu, Maosong Sun, and Bowen Zhou. 2023. Enhancing chat language models by scaling high-quality instructional conversations. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 3029–3051, Si…
  15. Stefan Elfwing, Eiji Uchibe, and Kenji Doya. 2017. Sigmoid-weighted linear units for neural network function approximation in reinforcement learning. Preprint, arXiv:1702.03118. — [arXiv:1702.03118](https://arxiv.org/abs/1702.03118) find
  16. Lewis R Goldberg. 2013. An alternative “description of personality”: The big-five factor structure. In Personality and personality disorders, pages 34–47. Routledge. find
  17. Aaron Grattafiori, Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Abhishek Kadian, Ahmad Al-Dahle, Aiesha Letman, Akhil Mathur, Alan Schelten, Alex Vaughan, Amy Yang, Angela Fan, Anirudh Goyal, Anthony Hartshorn, Aobo Yang, Archi Mitra, Archie Sravankumar, Artem Korenev, Arthur Hinsvark, and 542 o… — The Llama 3 Herd of Models
  18. Justyna Gromada, Alicja Kasicka, Ewa Komkowska, Lukasz Krajewski, Natalia Krawczyk, Morgan Veyret, Bartosz Przybyl, Lina Maria Rojas-Barahona, and Michal Szczerbak. 2025. Evaluating conversational agents with persona-driven user simulations based on large language models: A sales bot case study. In… find
  19. Gunmay Handa, Zekun Wu, Adriano Koshiyama, and Philip Treleaven. 2025. Personality as a probe for llm evaluation: Method trade-offs and downstream effects. Preprint, arXiv:2509.04794. — [arXiv:2509.04794](https://arxiv.org/abs/2509.04794) find
  20. Dan Hendrycks and Kevin Gimpel. 2023. Gaussian error linear units (gelus). Preprint, arXiv:1606.08415. find
  21. Edward J. Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. 2021. Lora: Low-rank adaptation of large language models. Preprint, arXiv:2106.09685.
  22. Aaron Hurst, Adam Lerer, Adam P Goucher, Adam Perelman, Aditya Ramesh, Aidan Clark, AJ Ostrow, Akila Welihinda, Alan Hayes, Alec Radford, and 1 others. 2024. Gpt-4o system card. arXiv preprint arXiv:2410.21276.
  23. Guangyuan Jiang, Manjie Xu, Song-Chun Zhu, Wenjuan Han, Chi Zhang, and Yixin Zhu. 2023. Evaluating and inducing personality in pre-trained language models. In Proceedings of the 37th International Conference on Neural Information Processing Systems, NIPS ’23, Red Hook, NY, USA. Curran Associates Inc… find
  24. Oliver P John, Eileen M Donahue, and Robert L Kentle. 1991. Big five inventory. Journal of personality and social psychology. find
  25. Tianjie Ju, Zhenyu Shao, Bowen Wang, Yujia Chen, Zhuosheng Zhang, Hao Fei, Mong-Li Lee, Wynne Hsu, Sufeng Duan, and Gongshen Liu. 2025. Probing then editing response personality of large language models. arXiv preprint arXiv:2504.10227. — [arXiv:2504.10227](https://arxiv.org/abs/2504.10227) find
  26. Kurt Lewin. 2013. Principles of topological psychology. Read Books Ltd. find
  27. Cheng Li, Jindong Wang, Yixuan Zhang, Kaijie Zhu, Wenxin Hou, Jianxun Lian, Fang Luo, Qiang Yang, and Xing Xie. 2023. Large language models understand and can be enhanced by emotional stimuli. Preprint, arXiv:2307.11760.
  28. Wenkai Li, Jiarui Liu, Andy Liu, Xuhui Zhou, Mona T. Diab, and Maarten Sap. 2025. BIG5-CHAT: Shaping LLM personalities through training on human-grounded data. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 20434– 20471, Vien…
  29. Xingxuan Li, Yutong Li, Lin Qiu, Shafiq Joty, and Lidong Bing. 2024a. Evaluating psychological safety of large language models. Preprint, arXiv:2212.10529. find
  30. Xinyu Li, Zachary C. Lipton, and Liu Leqi. 2024b. Personalized language modeling from personalized human feedback. CoRR, abs/2402.05133. — [arXiv:2402.05133](https://arxiv.org/abs/2402.05133) find
  31. Danni Liu and Jan Niehues. 2025. Middle-layer representation alignment for cross-lingual transfer in fine-tuned LLMs. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 15979– 15996, Vienna, Austria. Association for Computational…
  32. Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, and Veselin Stoyanov. 2019. Roberta: A robustly optimized bert pretraining approach. Preprint, arXiv:1907.11692.
  33. Shayne Longpre, Le Hou, Tu Vu, Albert Webson, Hyung Won Chung, Yi Tay, Denny Zhou, Quoc V. Le, Barret Zoph, Jason Wei, and Adam Roberts. 2023. The flan collection: Designing data and methods for effective instruction tuning. Preprint, arXiv:2301.13688.
  34. Robert R McCrae and Oliver P John. 1992. An introduction to the five-factor model and its applications. Journal of personality, 60(2):175–215. find
  35. Walter Mischel. 1968. Personality and assessment. find
  36. Walter Mischel and Philip K Peake. 1982. Beyond déjà vu in the search for cross-situational consistency. Psychological review, 89(6):730. find
  37. Walter Mischel and Yuichi Shoda. 1995. A cognitive-affective system theory of personality: reconceptualizing situations, dispositions, dynamics, and invariance in personality structure. Psychological review, 102(2):246. find
  38. Walter Mischel, Yuichi Shoda, and Rodolfo Mendoza-Denton. 2002. Situation-behavior profiles as a locus of consistency in personality. Current Directions in Psychological Science, 11(2):50–54. find
  39. Jisoo Mok, Ik-hwan Kim, Sangkwon Park, and Sungroh Yoon. 2025. Exploring the potential of LLMs as personalized assistants: Dataset, evaluation, and analysis. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 10212–10239, Vienna,… find
  40. Vinod Nair and Geoffrey E. Hinton. 2010. Rectified linear units improve restricted boltzmann machines. In Proceedings of the 27th International Conference on International Conference on Machine Learning, ICML’10, page 807–814, Madison, WI, USA. Omnipress. find
  41. Theodore Mead Newcomb. 1929. The consistency of certain extrovert-introvert behavior patterns in 51 problem boys. 382. Teachers College, Columbia University. find
  42. Chris Richardson, Yao Zhang, Kellen Gillespie, Sudipta Kar, Arshdeep Singh, Zeynab Raeesy, Omar Zia Khan, and Abhinav Sethy. 2023. Integrating summarization and retrieval for enhanced personalization via large language models. Preprint, arXiv:2310.20081. — [arXiv:2310.20081](https://arxiv.org/abs/2310.20081) find
  43. Alireza Salemi, Surya Kallumadi, and Hamed Zamani. 2024. Optimization methods for personalizing large language models through retrieval augmentation. Preprint, arXiv:2404.05970. — [arXiv:2404.05970](https://arxiv.org/abs/2404.05970) find
  44. Noam Shazeer. 2020. Glu variants improve transformer. Preprint, arXiv:2002.05202. find
  45. Oscar Skean, Md Rifat Arefin, Dan Zhao, Niket Nikul Patel, Jalal Naghiyev, Yann LeCun, and Ravid Shwartz-Ziv. 2025. Layer by layer: Uncovering hidden representations in language models. In Forty-second International Conference on Machine Learning. find
  46. Marton Szep, Daniel Rueckert, Rüdiger von Eisenhart-Rothe, and Florian Hinterwimmer. 2025. Fine-tuning large language models with limited data: A survey and practical guide. Preprint, arXiv:2411.09539.
  47. Alon Talmor, Jonathan Herzig, Nicholas Lourie, and Jonathan Berant. 2019. CommonsenseQA: A question answering challenge targeting commonsense knowledge. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies,…
  48. Fiona Anting Tan, Gerard Christopher Yeo, Kokil Jaidka, Fanyou Wu, Weijie Xu, Vinija Jain, Aman Chadha, Yang Liu, and See-Kiong Ng. 2024a. Phantom: Persona-based prompting has an effect on theory-of-mind reasoning in large language models. Preprint, arXiv:2403.02246.
  49. Zhaoxuan Tan, Qingkai Zeng, Yijun Tian, Zheyuan Liu, Bing Yin, and Meng Jiang. 2024b. Democratizing large language models via personalized parameter-efficient fine-tuning. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pages 6476–6491, Miami, Florida, USA.…
  50. Gemma Team, Aishwarya Kamath, Johan Ferret, Shreya Pathak, Nino Vieillard, Ramona Merhej, Sarah Perrin, Tatiana Matejovicova, Alexandre Ramé, Morgane Rivière, Louis Rouillard, Thomas Mesnard, Geoffrey Cideron, Jean bastien Grill, Sabela Ramos, Edouard Yvinec, Michelle Casbon, Etienne Pot, Ivo Penche… find
  51. Gemma Team, Thomas Mesnard, Cassidy Hardin, Robert Dadashi, Surya Bhupatiraju, Shreya Pathak, Laurent Sifre, Morgane Rivière, Mihir Sanjay Kale, Juliette Love, Pouya Tafti, Léonard Hussenot, Pier Giuseppe Sessa, Aakanksha Chowdhery, Adam Roberts, Aditya Barua, Alex Botev, Alex Castro-Ros, Ambrose Sl… find
  52. Yu-Min Tseng, Yu-Chao Huang, Teng-Yun Hsiao, Wei-Lin Chen, Chao-Wei Huang, Yu Meng, and Yun-Nung Chen. 2024. Two tales of persona in LLMs: A survey of role-playing and personalization. In Findings of the Association for Computational Linguistics: EMNLP 2024, pages 16612–16631, Miami, Florida, USA. A…
  53. Ernest C. Tupes and Raymond E. Christal. 1992. Recurrent personality factors based on trait ratings. Journal of personality, 60 2:225–51. find
  54. Alexander Matt Turner, Lisa Thiergart, Gavin Leech, David Udell, Juan J. Vazquez, Ulisse Mini, and Monte MacDiarmid. 2024. Steering language models with activation engineering. Preprint, arXiv:2308.10248.
  55. Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Łukasz Kaiser, and Illia Polosukhin. 2017. Attention is all you need. In Proceedings of the 31st International Conference on Neural Information Processing Systems, NIPS’17, page 6000–6010, Red Hook, NY, USA. Cur…
  56. Lei Wang, Jianxun Lian, Yi Huang, Yanqi Dai, Haoxuan Li, Xu Chen, Xing Xie, and Ji-Rong Wen. 2025a. CharacterBox: Evaluating the role-playing capabilities of LLMs in text-based virtual worlds. In Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computa… find
  57. Xiaoyang Wang, Hongming Zhang, Tao Ge, Wenhao Yu, Dian Yu, and Dong Yu. 2025b. Opencharacter: Training customizable role-playing llms with large-scale synthetic personas. Preprint, arXiv:2501.15427. — [arXiv:2501.15427](https://arxiv.org/abs/2501.15427) find
  58. Xuan Zhang, Wenxuan Zhang, Anxu Wang, See-Kiong Ng, and Yang Deng. 2025. Masim: Multilingual agent-based simulation for social science. CoRR, abs/2512.07195. — [arXiv:2512.07195](https://arxiv.org/abs/2512.07195) find
  59. Xuhui Zhou, Hao Zhu, Leena Mathur, Ruohong Zhang, Haofei Yu, Zhengyang Qi, Louis-Philippe Morency, Yonatan Bisk, Daniel Fried, Graham Neubig, and Maarten Sap. 2024. SOTOPIA: interactive evaluation for social intelligence in language agents. In The Twelfth International Conference on Learning Represe…
  60. Minjun Zhu, Yixuan Weng, Linyi Yang, and Yue Zhang. 2025. Personality alignment of large language models. In The Thirteenth International Conference on Learning Representations, ICLR 2025, Singapore, April 24-28, 2025. OpenReview.net.
  61. Xintao Wang, Heng Wang, Yifei Zhang, Xinfeng Yuan, Rui Xu, Jen tse Huang, Siyu Yuan, Haoran Guo, Jiangjie Chen, Shuchang Zhou, Wei Wang, and Yanghua Xiao. 2025c. Coser: Coordinating llm-based persona simulation of established roles. Preprint, arXiv:2502.09082. — [arXiv:2502.09082](https://arxiv.org/abs/2502.09082) find
  62. Zixiao Wang, Duzhen Zhang, Ishita Agarwal, Shen Gao, Le Song, and Xiuying Chen. 2025d. Beyond profile: From surface-level facts to deep persona simulation in LLMs. In Findings of the Association for Computational Linguistics: ACL 2025, pages 21239–21257, Vienna, Austria. Association for Computationa… find
  63. Shenghan Wu, Yimo Zhu, Wynne Hsu, Mong-Li Lee, and Yang Deng. 2025. From personas to talks: Revisiting the impact of personas on llm-synthesized emotional support conversations. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, EMNLP 2025, Suzhou, China, Nove… find
  64. An Yang, Anfeng Li, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chang Gao, Chengen Huang, Chenxu Lv, Chujie Zheng, Dayiheng Liu, Fan Zhou, Fei Huang, Feng Hu, Hao Ge, Haoran Wei, Huan Lin, Jialong Tang, and 41 others. 2025. Qwen3 technical report. Preprint, arXiv:2505.09388.
  65. Tong Zhang, Chen Huang, Yang Deng, Hongru Liang, Jia Liu, Zujie Wen, Wenqiang Lei, and Tat-Seng Chua. 2024. Strength lies in differences! improving strategy planning for non-collaborative dialogues via diversified user simulation. In Proceedings of the 2024 Conference on Empirical Methods in Natural… find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/wei-2026-beyond-static-personas-situational-personality/