Elamur Research⌕
☰
home / papers / 2026 / Beyond Prompt: Fine-grained Simulation of Cognitively Impaired Standardized Patients via Stochastic Steering

Beyond Prompt: Fine-grained Simulation of Cognitively Impaired Standardized Patients via Stochastic Steering

Paper · 2026 · Weikang Zhang, Zimo Zhu, Zhichuan Yang et al.

Summary

Training clinical staff to talk with people who have cognitive impairment (e.g., MCI or Alzheimer's) usually depends on human actors playing Standardized Patients (SPs). Actors are costly and struggle to reproduce how deficits vary across cognitive domains and severity levels. Prompt-based LLM patients scale better, but prompts are too coarse to control these differences. The authors propose StsPatient, which steers a model's hidden states instead.

First, an LLM synthesizes contrastive data for each cognitive domain in two forms: impaired-versus-healthy instructions and impaired-versus-healthy responses. A steering vector is taken as the normalized mean difference of activations at the layer where the two classes separate best. Second, Stochastic Token Modulation (STM) sets severity by changing how likely the vector is to be injected at each token, rather than scaling its magnitude. The idea borrows from probabilistic synaptic release, and it avoids the instability that coefficient scaling usually causes.

StsPatient beats prompt-based and steering baselines on clinical authenticity and severity control. This makes realistic, graded patient simulation more practical for clinical training.

Key points

  • Five cognitive domains are modeled: Memory, Attention, Processing Speed, Reasoning and Social Cognition. Severity is a continuous value s ∈ [0, 1], applied through a per-token Bernoulli gate.
  • The method reports an average gain of +11.23% across all metrics over baselines, and +18.54% over the strongest baseline in severity controllability.
  • Steering vectors come only from synthetic therapist–patient dialogues. Publicly available, de-identified MTSamples transcriptions serve only as templates for patient profiles, and no raw MTSamples text is released.
  • In the safety audit, keeping severity below 0.8 produced no flagged outputs. An extra LLM-based filter caught all harmful content in samples with s > 0.5.
  • Limitations: the simulation is text-only, with no prosody, speech rate or visual cues. It mainly targets single domains, and combining several deficits is examined only preliminarily.

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

🎓National University of SingaporeSingapore · Singapore207 works · 39,576 citations🎓Xi'an Jiaotong UniversityXi'an · China42 works · 36,431 citations🎓Sichuan UniversityChengdu · China20 works · 1,293 citations🏛️Ministry of Education of the People's Republic of ChinaBeijing · China4 works · 117 citations🔬Institute of Data ScienceSingapore · Singapore3 works · 4 citations

 Fields, topics and concepts

📚 Shelves 5The field of science this work belongs to (OpenAlex fields).Computer Science · 18,614 worksArtificial Intelligence · 17,965 worksMedicine · 275 worksHealth Informatics · 214 worksPhysiology · 35 works

Cite

Cite as Weikang Zhang, Zimo Zhu, Zhichuan Yang, Chen Huang, Wenqiang Lei and See-Kiong Ng. 2026. Beyond Prompt: Fine-grained Simulation of Cognitively Impaired Standardized Patients via Stochastic Steering. 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.1401

@inproceedings{zhang2026beyond,
  title = {{Beyond Prompt: Fine-grained Simulation of Cognitively Impaired Standardized Patients via Stochastic Steering}},
  author = {Weikang Zhang, Zimo Zhu, Zhichuan Yang, Chen Huang, Wenqiang Lei and See-Kiong Ng},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.1401},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.1401}
}

References

63 references, 11 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Aydan Akkurt Yalcinturk and Melike Dissiz. 2024. The effect of standardized patient simulation method on getting nursing students to gain care providing skill for the individual diagnosed with dementia: A randomized controlled trial. Educational Gerontology, 50(3):229–239. find
  2. Howard S Barrows. 1993. An overview of the uses of standardized patients for teaching and evaluating clinical skills. aamc. Academic medicine, 68(6):443– 51. find
  3. Lukasz Bartoszcze, Sarthak Munshi, Bryan Sukidi, Jennifer Yen, Zejia Yang, David Williams-King, Linh Le, Kosi Asuzu, and Carsten Maple. 2025. Representation engineering for large-language models: Survey and research challenges. arXiv preprint arXiv:2502.17601. — [arXiv:2502.17601](https://arxiv.org/abs/2502.17601) find
  4. Daniela Beltrami, Laura Calzà, Gloria Gagliardi, Enrico Ghidoni, Norina Marcello, Rema Rossini Favretti, and Fabio Tamburini. 2016. Automatic identification of mild cognitive impairment through the analysis of Italian spontaneous speech productions. In Proceedings of the Tenth International Conferen… find
  5. Anna Bodonhelyi, Christian Stegemann-Philipps, Alessandra Sonanini, Lea Herschbach, Marton Szep, Anne Herrmann-Werner, Teresa Festl-Wietek, Enkelejda Kasneci, and Friederike Holderried. 2025. Modeling challenging patient interactions: Llms for medical communication training. Preprint, arXiv:2503.222… — [arXiv:2503.22250](https://arxiv.org/abs/2503.22250) find
  6. Christopher R Bowie and Philip D Harvey. 2005. Cognition in schizophrenia: impairments, determinants, and functional importance. Psychiatric Clinics, 28(3):613–633. find
  7. Tiago Branco and Kevin Staras. 2009. The probability of neurotransmitter release: variability and feedback control at single synapses. Nature Reviews Neuroscience, 10(5):373–383. find
  8. Rosa Burgos, Irene Bretón, Emanuele Cereda, Jean Claude Desport, Rainer Dziewas, Laurence Genton, Filomena Gomes, Pierre Jésus, Andreas Leis-chker, Maurizio Muscaritoli, and 1 others. 2018. Espen guideline clinical nutrition in neurology. Clinical nutrition, 37(1):354–396. find
  9. Nuo Chen, Hongguang Li, Jianhui Chang, Juhua Huang, Baoyuan Wang, and Jia Li. 2025a. Compress to impress: Unleashing the potential of compressive memory in real-world long-term conversations. In Proceedings of the 31st International Conference on Computational Linguistics, pages 755–773. find
  10. Nuo Chen, Yan Wang, Yang Deng, and Jia Li. 2024. The oscars of ai theater: A survey on role-playing with language models. arXiv preprint arXiv:2407.11484. — [arXiv:2407.11484](https://arxiv.org/abs/2407.11484) find
  11. Nuo Chen, Yan Wang, Haiyun Jiang, Deng Cai, Yuhan Li, Ziyang Chen, Longyue Wang, and Jia Li. 2023a. Large language models meet harry potter: A dataset for aligning dialogue agents with characters. In Findings of the association for computational linguistics: EMNLP 2023, pages 8506–8520.
  12. Runjin Chen, Andy Arditi, Henry Sleight, Owain Evans, and Jack Lindsey. 2025b. Persona vectors: Monitoring and controlling character traits in language models. arXiv preprint arXiv:2507.21509. — [arXiv:2507.21509](https://arxiv.org/abs/2507.21509) find
  13. Siyuan Chen, Mengyue Wu, Kenny Q Zhu, Kunyao Lan, Zhiling Zhang, and Lyuchun Cui. 2023b. Llm-empowered chatbots for psychiatrist and patient simulation: application and evaluation. arXiv preprint arXiv:2305.13614. — [arXiv:2305.13614](https://arxiv.org/abs/2305.13614) find
  14. Valerie T Cotter, Danetta H Sloan, Daniel L Scer-pella, Kelly M Smith, Martha Abshire Saylor, and Jennifer L Wolff. 2025. Feasibility of using simulation to evaluate implementation fidelity in an advance care planning pragmatic trial. American Journal of Hospice and Palliative Medicine®, page 104990… find
  15. Department of Social Work and Social Administration, The University of Hong Kong. 2024. Brain live: Cognitive impairment and rehabilitation materials. Accessed: 2025-12-11. find
  16. Zhuoyun Du, LujieZheng LujieZheng, Renjun Hu, Yuyang Xu, Xiawei Li, Ying Sun, Wei Chen, Jian Wu, Haolei Cai, and Haochao Ying. 2025a. LLMs can simulate standardized patients via agent coevolution. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: L…
  17. Zhuoyun Du, LujieZheng LujieZheng, Renjun Hu, Yuyang Xu, Xiawei Li, Ying Sun, Wei Chen, Jian Wu, Haolei Cai, and Haochao Ying. 2025b. Llms can simulate standardized patients via agent coevolution. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: L…
  18. Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Abhishek Kadian, Ahmad Al-Dahle, Aiesha Letman, Akhil Mathur, Alan Schelten, Amy Yang, Angela Fan, and 1 others. 2024. The llama 3 herd of models. arXiv preprint arXiv:2407.21783.
  19. Chukwuka Elendu, Dependable C Amaechi, Alexander U Okatta, Emmanuel C Amaechi, Tochi C Elendu, Chiamaka P Ezeh, and Ijeoma D Elendu. 2024. The impact of simulation-based training in medical education: A review. Medicine, 103(27):e38813. find
  20. Katrina E Forbes-McKay and Annalena Venneri. 2005. Detecting subtle spontaneous language decline in early alzheimer’s disease with a picture description task. Neurological sciences, 26(4):243–254. find
  21. Meiko Fukuda, Ryota Nishimura, Maina Umezawa, Kazumasa Yamamoto, Yurie Iribe, and Norihide Kitaoka. 2022. Elderly conversational speech corpus with cognitive impairment test and pilot dementia detection experiment using acoustic characteristics of speech in Japanese dialects. In Proceedings of the T… find
  22. Asma Ghandeharioun, Avi Caciularu, Adam Pearce, Lucas Dixon, and Mor Geva. 2024. Patchscopes: A unifying framework for inspecting hidden representations of language models. arXiv preprint arXiv:2401.06102.
  23. Dimitris Gkoumas, Adam Tsakalidis, and Maria Liakata. 2023. A digital language coherence marker for monitoring dementia. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 16021–16034, Singapore. Association for Computational Linguistics. find
  24. Kai He, Yucheng Huang, Wenqing Wang, Delong Ran, Dongming Sheng, Junxuan Huang, Qika Lin, Jiaxing Xu, Wenqiang Liu, and Mengling Feng. 2025. Crab: A novel configurable role-playing LLM with assessing benchmark. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistic… find
  25. Chen Huang, Peixin Qin, Yang Deng, Wenqiang Lei, Jiancheng Lv, and Tat-Seng Chua. 2024. Concept – an evaluation protocol on conversational recommender systems with system-centric and user-centric factors. Preprint, arXiv:2404.03304. — [arXiv:2404.03304](https://arxiv.org/abs/2404.03304) find
  26. Youcheng Huang, Chen Huang, Duanyu Feng, Wenqiang Lei, and Jiancheng Lv. 2025. Cross-model transferability among large language models on the platonic representations of concepts. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), page… find
  27. Jiyue Jiang, Sheng Wang, Qintong Li, Lingpeng Kong, and Chuan Wu. 2023. A cognitive stimulation dialogue system with multi-source knowledge fusion for elders with cognitive impairment. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers),… find
  28. René S Kahn and Richard SE Keefe. 2013. Schizophrenia is a cognitive illness: time for a change in focus. JAMA psychiatry, 70(10). find
  29. Ruhul Amin Khalil, Kashif Ahmad, and Hazrat Ali. 2025. Redefining elderly care with agentic ai: Challenges and opportunities. arXiv preprint arXiv:2507.14912. — [arXiv:2507.14912](https://arxiv.org/abs/2507.14912) find
  30. Jihyun Lee, Hyungsin Kim, Kwan Hoon Kim, Daeun Jung, Tanisha Jowsey, and Craig S Webster. 2020. Effective virtual patient simulators for medical communication training: a systematic review. Medical education, 54(9):786–795. find
  31. Keyeun Lee, Seolhee Lee, Esther Hehsun Kim, Yena Ko, Jinsu Eun, Dahee Kim, Hyewon Cho, Haiyi Zhu, Robert E. Kraut, Eunyoung E. Suh, Eun-mee Kim, and Hajin Lim. 2025. Adaptive-VP: A framework for LLM-based virtual patients that adapts to trainees’ dialogue to facilitate nurse communication training.… find
  32. Suyeon Lee, Sunghwan Kim, Minju Kim, Dongjin Kang, Dongil Yang, Harim Kim, Minseok Kang, Dayi Jung, Min Hee Kim, Seungbeen Lee, Kyong-Mee Chung, Youngjae Yu, Dongha Lee, and Jinyoung Yeo. 2024. Cactus: Towards psychological counseling conversations using cognitive behavioral theory. In Findings of t… find
  33. Yanzeng Li, Cheng Zeng, Jinchao Zhang, Jie Zhou, and Lei Zou. 2025. Meddit: A knowledge-controlled diffusion transformer framework for dynamic medical image generation in virtual simulated patient. In Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence, IJCAI-2… find
  34. Yanzeng Li, Cheng Zeng, Jialun Zhong, Ruoyu Zhang, Minhao Zhang, and Lei Zou. 2024. Leveraging large language model as simulated patients for clinical education. arXiv preprint arXiv:2404.13066. — [arXiv:2404.13066](https://arxiv.org/abs/2404.13066) find
  35. Yusheng Liao, Yutong Meng, Yuhao Wang, Hongcheng Liu, Yanfeng Wang, and Yu Wang. 2024. Automatic interactive evaluation for large language models with state aware patient simulator. arXiv preprint arXiv:2403.08495. — [arXiv:2403.08495](https://arxiv.org/abs/2403.08495) find
  36. Maria R Lima, Amy O’Connell, Feiyang Zhou, Alethea Nagahara, Avni Hulyalkar, Anura Deshpande, Jesse Thomason, Ravi Vaidyanathan, and Maja Matarić. 2025. Promoting cognitive health in elder care with large language model-powered socially assistive robots. In Proceedings of the 2025 CHI Conference on… find
  37. Ryan Louie, Ananjan Nandi, William Fang, Cheng Chang, Emma Brunskill, and Diyi Yang. 2024. Roleplay-doh: Enabling domain-experts to create llm-simulated patients via eliciting and adhering to principles. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pages…
  38. Keming Lu, Bowen Yu, Chang Zhou, and Jingren Zhou. 2024. Large language models are superpositions of all characters: Attaining arbitrary role-play via self-alignment. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 7828–7840.
  39. Lauren McCollum and Jason Karlawish. 2020. Cognitive impairment evaluation and management. The Medical Clinics of North America, 104(5):807. find
  40. Paul Roux, Yujiro Okuya, Cristina Morel, Mariane Soulès, Hugo Bottemanne, Eric Brunet-Gouet, Solène Frileux, Christine Passerieux, Nadia Younes, and Jean Claude Martin. 2025. Effectiveness of a web-based virtual simulation to train nursing students in suicide risk assessment: Randomized controlled i… find
  41. Robert A McCutcheon, Richard SE Keefe, and Philip K McGuire. 2023. Cognitive impairment in schizophrenia: aetiology, pathophysiology, and treatment. Molecular psychiatry, 28(5):1902–1918. find
  42. Haoyu Song, Yan Wang, Weinan Zhang, Xiaojiang Liu, and Ting Liu. 2020. Generate, delete and rewrite: A three-stage framework for improving persona consistency of dialogue generation. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 5821–5831. find
  43. Hongbin Na, Yining Hua, Zimu Wang, Tao Shen, Beibei Yu, Lilin Wang, Wei Wang, John Torous, and Ling Chen. 2025. A survey of large language models in psychotherapy: Current landscape and future directions. In Findings of the Association for Computational Linguistics: ACL 2025, pages 7362–7376, Vienna… find
  44. Daniel Tan, David Chanin, Aengus Lynch, Brooks Paige, Dimitrios Kanoulas, Adrià Garriga-Alonso, and Robert Kirk. 2024. Analysing the generalisation and reliability of steering vectors. In Proceedings of the 38th International Conference on Neural Information Processing Systems, NIPS ’24, Red Hook, N… find
  45. Ruihao Pan, Ziming Liu, Fengpei Yuan, Maryam Zare, Xiaopeng Zhao, and Rebecca Jane Passonneau. 2022. A database of multimodal data to construct a simulated dialogue partner with varying degrees of cognitive health. In Proceedings of the RaPID Workshop - Resources and ProcessIng of linguistic, parali… find
  46. Joon Sung Park, Joseph O’Brien, Carrie Jun Cai, Meredith Ringel Morris, Percy Liang, and Michael S Bernstein. 2023. Generative agents: Interactive simulacra of human behavior. In Proceedings of the 36th annual acm symposium on user interface software and technology, pages 1–22. find
  47. Daniele PotertĂŹ, Andrea Seveso, and Fabio Mercorio. 2025. Designing role vectors to improve llm inference behaviour. arXiv preprint arXiv:2502.12055. — [arXiv:2502.12055](https://arxiv.org/abs/2502.12055) find
  48. Huachuan Qiu and Zhenzhong Lan. 2024. Interactive agents: Simulating counselor-client psychological counseling via role-playing llm-to-llm interactions. Preprint, arXiv:2408.15787.
  49. Daniel Reichenpfader and Kerstin Denecke. 2024. Simulating diverse patient populations using patient vignettes and large language models. In Proceedings of the First Workshop on Patient-Oriented Language Processing (CL4Health)@ LREC-COLING 2024, pages 20–25. find
  50. Nina Rimsky, Nick Gabrieli, Julian Schulz, Meg Tong, Evan Hubinger, and Alexander Turner. 2024. Steering llama 2 via contrastive activation addition. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 15504–15522. find
  51. Quan Hung Tran, Svetha Venkatesh, Hung Le, and 1 others. 2025. Dynamic steering with episodic memory for large language models. In Findings of the Association for Computational Linguistics: ACL 2025, pages 13731–13749. find
  52. Shaun Varrecchia, Carol Maritz, Colleen Maher, and Megan Strauss. 2020. Managing older adults with cognitive impairment: An interprofessional, standardized patient approach. Innovation in Aging, 4(Supplement_1):10–10. find
  53. Veronika Vincze, Gábor Gosztolya, László Tóth, Ildikó Hoffmann, Gréta Szatlóczki, Zoltán Bánréti, Magdolna Pákáski, and János Kálmán. 2016. Detecting mild cognitive impairment by exploiting linguistic information from transcripts. In Proceedings of the 54th Annual Meeting of the Association for Comp… find
  54. Rohit Voleti, Julie M Liss, and Visar Berisha. 2019. A review of automated speech and language features for assessment of cognitive and thought disorders. IEEE journal of selected topics in signal processing, 14(2):282–298. find
  55. Anyi Wang, Dong Shu, Yifan Wang, Yunpu Ma, and Mengnan Du. 2025. Improving llm reasoning through interpretable role-playing steering. arXiv preprint arXiv:2506.07335. find
  56. Jiashuo Wang, Yang Xiao, Yanran Li, Changhe Song, Chunpu Xu, Chenhao Tan, and Wenjie Li. 2024a. Towards a client-centered assessment of llm therapists by client simulation. Preprint, arXiv:2406.12266. — [arXiv:2406.12266](https://arxiv.org/abs/2406.12266) find
  57. Noah Wang, Z.y. Peng, Haoran Que, Jiaheng Liu, Wangchunshu Zhou, Yuhan Wu, Hongcheng Guo, Ruitong Gan, Zehao Ni, Jian Yang, Man Zhang, Zhaoxiang Zhang, Wanli Ouyang, Ke Xu, Wenhao Huang, Jie Fu, and Junran Peng. 2024b. RoleLLM: Benchmarking, eliciting, and enhancing role-playing abilities of large l…
  58. Ruiyi Wang, Stephanie Milani, Jamie C. Chiu, Jiayin Zhi, Shaun M. Eack, Travis Labrum, Samuel M Murphy, Nev Jones, Kate V Hardy, Hong Shen, Fei Fang, and Zhiyu Chen. 2024c. PATIENT-ψ: Using large language models to simulate patients for training mental health professionals. In Proceedings of the 202… find
  59. Lidewij A Wind, Jan Van Dalen, Arno MM Muijtjens, and Jan-Joost Rethans. 2004. Assessing simulated patients in an educational setting: the masp (maastricht assessment of simulated patients). Medical education, 38(1):39–44. find
  60. An Yang, Anfeng Li, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chang Gao, Chengen Huang, Chenxu Lv, and 1 others. 2025. Qwen3 technical report. arXiv preprint arXiv:2505.09388.
  61. Stav Yosef, Moreah Zisquit, Ben Cohen, Anat Klomek Brunstein, Kfir Bar, and Doron Friedman. 2024. Assessing motivational interviewing sessions with AI-generated patient simulations. In Proceedings of the 9th Workshop on Computational Linguistics and Clinical Psychology (CLPsych 2024), pages 1–11, St… find
  62. Runcong Zhao, Wenjia Zhang, Jiazheng Li, Lixing Zhu, Yanran Li, Yulan He, and Lin Gui. 2024. Narrativeplay: Interactive narrative understanding. In Proceedings of the 18th Conference of the European Chapter of the Association for Computational Linguistics: System Demonstrations, pages 82–93. find
  63. 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. arXiv preprint arXiv:2310.01405.

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/zhang-2026-beyond-prompt-fine-grained-simulation/