[1]CHAI Chunlei,GE Zhichao,YIN Min,et al.A survey of techniques for realizing personality expression in large language models[J].CAAI Transactions on Intelligent Systems,2026,21(2):321-336.[doi:10.11992/tis.202505031]
Copy

A survey of techniques for realizing personality expression in large language models

CAAI Transactions on Intelligent Systems[ISSN 1673-4785/CN 23-1538/TP] Volume: 21 Number of periods: 2026 2 Page number: 321-336 Column: 综述 Public date: 2026-04-30
Title:
A survey of techniques for realizing personality expression in large language models
Author(s):
CHAI Chunlei1; 2;  GE Zhichao1;  YIN Min23;  WANG Zheng1;  LIAN Boyi1
1. State Key Laboratory of CAD & CG, Zhejiang University, Hangzhou 310027, China;
2. Yangtze River Delta Smart Oasis Innovation Center, Zhejiang University, Jiaxing 314100, China;
3. School of Management, Zhejiang University, Hangzhou 310058, China
Keywords:
large language models; personality expression in large language models; personality assessment; personality generation; digital persona; natural language processing; role-playing; personality alignment
CLC:
TP18
DOI:
10.11992/tis.202505031
Abstract:
Personality expression in Large Language Models (LLMs) has emerged as a key direction in human–computer interaction research. Enabling machines to exhibit uniquely human-like expressiveness remains a significant challenge in the LLM domain. In recent years, the application of pre-training, fine-tuning, and agent-based collaboration techniques has matured to the point where LLMs can adopt role-specific personalities tailored to diverse scenario and task requirements. Whether through carefully curated training datasets at the input stage, parameter-level adaptations within the model, or external agents and workflows around the model, personality expression in LLMs can be achieved. This paper provides a comprehensive review of the current state and future trends of personality expression in LLMs. On one hand, it examines techniques for extracting and simulating personality traits; on the other, it explores methods for controlling and aligning model personality. By analyzing and summarizing these approaches, we discuss the developmental directions of personality-driven LLM research.
References:
[1] WEIZENBAUM J. ELIZA: a computer program for the study of natural language communication between man and machine[J]. Communications of the ACM, 1966, 9(1): 36-45.
[2] METCALF K, THEOBALD B J, WEINBERG G, et al. Mirroring to build trust in digital assistants[EB/OL]. (2019-04-02)[2025-04-16]. http://arxiv.org/abs/1904.01664.
[3] JAMALIAN N, CONSTANTINIDES M, JOGLEKAR S, et al. Our Nudges, Our Selves: tailoring mobile user engagement using personality[EB/OL]. (2023-07-24)[2025-04-16]. http://arxiv.org/abs/2307.13145.
[4] ZHU Jianfeng, JIN Ruoming, COIFMAN K G. Investigating large language models in inferring personality traits from user conversations[EB/OL]. (2025-01-13)[2025-04-16]. http://arxiv.org/abs/2501.07532.
[5] JIANG Hang, ZHANG Xianzhe, CHOI J D. Automatic text-based personality recognition on monologues and multiparty dialogues using attentive networks and contextual embeddings[EB/OL]. (2019-11-21)[2025-04-16]. http://arxiv.org/abs/1911.09304.
[6] DATTA A, CHAKRABORTY S, MUKHERJEE A. Personality detection and analysis using Twitter data[EB/OL]. (2023-09-11)[2025-04-16]. http://arxiv.org/abs/2309.05497.
[7] YANG Qi, FARSEEV A, FILCHENKOV A. Two-faced humans on Twitter and Facebook: harvesting social multimedia for human personality profiling[C]//Proceedings of the 2021 ACM Workshop on Intelligent Cross-Data Analysis and Retrieval. New York: ACM, 2021: 39-47.
[8] VU X S, FLEKOVA L, JIANG Lili, et al. Lexical-semantic resources: yet powerful resources for automatic personality classification[EB/OL]. (2017-11-27)[2025-04-16]. http://arxiv.org/abs/1711.09824.
[9] XING Yujie, FERN?NDEZ R. Automatic evaluation of neural personality-based chatbots[EB/OL]. (2018-09-30)[2025-04-16]. http://arxiv.org/abs/1810.00472.
[10] ZHANG Le, PENG Songyou, WINKLER S. PersEmoN: a deep network for joint analysis of apparent personality, emotion and their relationship[J]. IEEE transactions on affective computing, 2022, 13(1): 298-305.
[11] CAO Xubo, KOSINSKI M. Large language models know how the personality of public figures is perceived by the general public[J]. Scientific reports, 2024, 14: 6735.
[12] HABIB F, ALI Z, AZAM A, et al. Navigating pathways to automated personality prediction: a comparative study of small and medium language models[J]. Frontiers in big data, 2024, 7: 1387325.
[13] NGO A, KOCO? J. Integrating personalized and contextual information in fine-grained emotion recognition in text: a multi-source fusion approach with explainability[J]. Information fusion, 2025, 118: 102966.
[14] LIU Weiwei, HU Wenxuan, JING Wei, et al. Learning to model diverse driving behaviors in highly interactive autonomous driving scenarios with multi-agent reinforcement learning[EB/OL]. (2024-02-21)[2025-04-16]. http://arxiv.org/abs/2402.13481.
[15] PETERS H, MATZ S. Large language models can infer psychological dispositions of social media users[EB/OL]. (2024-07-05)[2025-04-16]. http://arxiv.org/abs/2309.08631.
[16] CHEN Yirong, FAN Weiquan, XING Xiaofen, et al. CPED: a large-scale Chinese personalized and emotional dialogue dataset for conversational AI[EB/OL]. (2022-05-29)[2025-04-16]. http://arxiv.org/abs/2205.14727.
[17] CAI Cong, LIANG Shan, LIU Xuefei, et al. MDPE: a multimodal deception dataset with personality and emotional characteristics[EB/OL]. (2024-07-17)[2025-04-16]. http://arxiv.org/abs/2407.12274.
[18] OCCHIPINTI D, TEKIROGLU S S, GUERINI M. PRODIGy: a profile-based dialogue generation dataset[EB/OL]. (2024-08-27)[2025-04-16]. http://arxiv.org/abs/2311.05195.
[19] KEUM B, SUN J, LEE W, et al. Persona-identified chatbot through small-scale modeling and data transformation[J]. Electronics, 2024, 13(8): 1409.
[20] SONI N, MATERO M, BALASUBRAMANIAN N, et al. Human language modeling[C]//Findings of the Association for Computational Linguistics: ACL 2022. Dublin: ACL, 2022: 622–636.
[21] JAIN N, WU Zekun, MUNOZ C, et al. From Text to Emoji: how peft-driven personality manipulation unleashes the emoji potential in LLMs[EB/OL]. (2025-02-25)[2025-04-16]. http://arxiv.org/abs/2409.10245.
[22] LU Zhenyi, WEI Wei, QU Xiaoye, et al. MIRACLE: towards personalized dialogue generation with latent-space multiple personal attribute control[EB/OL]. (2023-10-22)[2025-04-16]. http://arxiv.org/abs/2310.18342.
[23] LEE J Y, LEE K A, GAN W S. DLVGen: a dual latent variable approach to personalized dialogue generation[EB/OL]. (2021-11-22)[2025-04-16]. http://arxiv.org/abs/2111.11363.
[24] HUANG Yuxuan. Orca: enhancing role-playing abilities of large language models by integrating personality traits[EB/OL]. (2024-11-15)[2025-04-16]. http://arxiv.org/abs/2411.10006.
[25] TAO Meiling, LIANG Xuechen, SHI Tianyu, et al. RoleCraft-GLM: advancing personalized role-playing in large language models[EB/OL]. (2023-12-17)[2025-04-16]. http://arxiv.org/abs/2401.09432.
[26] SHI Haozhe, NIU Kun. Enhancing persona consistency with large language models[C]//Proceedings of the 2024 5th International Conference on Computing, Networks and Internet of Things. Tokyo: ACM, 2024: 210-215.
[27] LIU Jianzhi, GU Hexiang, ZHENG Tianyu, et al. Dynamic generation of personalities with large language models[EB/OL]. (2024-04-10)[2025-04-16]. http://arxiv.org/abs/2404.07084.
[28] VU H, NGUYEN H A, GANESAN A V, et al. PsychAdapter: adapting LLM Transformers to reflect traits, personality and mental health[EB/OL]. (2024-12-22)[2025-04-16]. http://arxiv.org/abs/2412.16882.
[29] GU Heng, DEGACHI C, GEN? U, et al. On the effectiveness of creating conversational agent personalities through prompting[EB/OL]. (2023-10-17)[2025-04-16]. http://arxiv.org/abs/2310.11182.
[30] RAMIREZ A, ALSALIHY M, AGGARWAL K, et al. Controlling personality style in dialogue with zero-shot prompt-based learning[EB/OL]. (2023-02-08)[2025-04-16]. http://arxiv.org/abs/2302.03848.
[31] HOUSLEY W, DAHL P. Membership categorisation, sociological description and role prompt engineering with ChatGPT[J]. Discourse & communication, 2024, 18(6): 848-858.
[32] CARON G, SRIVASTAVA S. Identifying and manipulating the personality traits of language models[EB/OL]. (2022-12-20)[2025-04-16]. http://arxiv.org/abs/2212.10276.
[33] JIANG Hang, ZHANG Xiajie, CAO Xubo, et al. PersonaLLM: investigating the ability of large language models to express personality traits[EB/OL]. (2023-05-04)[2025-04-16]. http://arxiv.org/abs/2305.02547.
[34] SHEN Chenglei, XIE Guofu, ZHANG Xiao, et al. On the decision-making abilities in role-playing using large language models[EB/OL]. (2024-02-29)[2025-04-16]. http://arxiv.org/abs/2402.18807.
[35] ZHAO Yilin, YUAN Xinbin, GAO Shanghua, et al. ChatAnything: facetime chat with LLM-enhanced personas[EB/OL]. (2023-11-12)[2025-04-16]. http://arxiv.org/abs/2311.06772.
[36] DUAN Yifan, TANG Yihong, BAI Xuefeng, et al. The power of personality: a human simulation perspective to investigate large language model agents[EB/OL]. (2025-02-28)[2025-04-16]. http://arxiv.org/abs/2502.20859.
[37] YANG Tao, ZHU Yuhua, QUAN Xiaojun, et al. PsyPlay: personality-infused role-playing conversational agents[EB/OL]. (2025-02-06)[2025-04-16]. http://arxiv.org/abs/2502.03821.
[38] JU H, ARAL S. Collaborating with AI Agents: field experiments on teamwork, productivity, and performance[EB/OL]. (2025-03-23)[2025-04-16]. http://arxiv.org/abs/2503.18238.
[39] MEDGYESY D, GALAS J, POL J van, et al. Existential Crisis: a social robot’s reason for being[EB/OL]. (2025-01-06)[2025-04-16]. http://arxiv.org/abs/2501.03376.
[40] HASAN M, OZEL C, POTTER S, et al. SAPIEN: affective virtual agents powered by large language models[C]//2023 11th International Conference on Affective Computing and Intelligent Interaction Workshops and Demos. Cambridge: IEEE, 2023: 1-3.
[41] SUN Yuqian, WANG Hanyi, CHAN P M, et al. Fictional worlds, real connections: developing community storytelling social chatbots through LLMs[EB/OL]. (2023-09-20)[2025-04-16]. http://arxiv.org/abs/2309.11478.
[42] ZHAO Runcong, ZHANG Wenjia, LI Jiazheng, et al. NarrativePlay: interactive narrative understanding[EB/OL]. (2023-10-02)[2025-04-16]. http://arxiv.org/abs/2310.01459.
[43] SCHAAFF K, REINIG C, SCHLIPPE T. Exploring ChatGPT’s empathic abilities[EB/OL]. (2023-08-07)[2025-04-16]. http://arxiv.org/abs/2308.03527.
[44] RUTINOWSKI J, FRANKE S, ENDENDYK J, et al. The self-perception and political biases of ChatGPT[J]. Human behavior and emerging technologies, 2024: 7115633.
[45] GRASSI L, RECCHIUTO C, SGORBISSA A. Enhancing LLM-based human-robot interaction with nuances for diversity awareness[EB/OL]. (2024-06-25)[2025-04-16]. http://arxiv.org/abs/2406.17531.
[46] HUANG Y J, HADFI R. How personality traits influence negotiation outcomes? A simulation based on large language models[EB/OL]. (2024-07-16)[2025-04-16]. http://arxiv.org/abs/2407.11549.
[47] BORMAN H, LEONTJEVA A, PIZZATO L, et al. Do LLM personas dream of bull markets? Comparing human and AI investment strategies through the lens of the five-factor model[EB/OL]. (2024-10-28)[2025-04-16]. http://arxiv.org/abs/2411.05801.
[48] TAKATA R, MASUMORI A, IKEGAMI T. Spontaneous emergence of agent individuality through social interactions in LLM-based communities[EB/OL]. (2024-11-05)[2025-04-16]. http://arxiv.org/abs/2411.03252.
[49] FRISCH I, GIULIANELLI M. LLM agents in interaction: measuring personality consistency and linguistic alignment in interacting populations of large language models[EB/OL]. (2024-02-05)[2025-04-16]. http://arxiv.org/abs/2402.02896.
[50] FISCHER K. Reflective linguistic programming (RLP): a stepping stone in socially-aware AGI (SocialAGI)[EB/OL]. (2023-05-22)[2025-04-16]. https://arxiv.org/abs/2305.12647.
[51] GUO Yaoqi, CHEN Zhenpeng, ZHANG J M, et al. Personality-guided code generation using large language models[EB/OL]. (2024-10-16)[2025-04-16]. http://arxiv.org/abs/2411.00006.
[52] PARK J S, ZOU C Q, SHAW A, et al. Generative agent simulations of 1, 000 people[EB/OL]. (2024-11-15)[2025-04-16]. http://arxiv.org/abs/2411.10109.
[53] REN Mingjun, XU Wentao. The impact of big five personality traits on AI agent decision-making in public spaces: a social simulation study[EB/OL]. (2025-01-15)[2025-04-16]. http://arxiv.org/abs/2503.15497.
[54] LUKIN S M, ANAND P, WALKER M, et al Argument strength is in the eye of the beholder: audience effects in persuasion[EB/OL]. (2017-08-30)[2025-04-16]. http://arxiv.org/abs/1708.09085.
[55] NOEVER D, HYAMS S. AI Text-to-Behavior: a study in steerability[EB/OL]. (2023-08-07)[2025-04-16]. http://arxiv.org/abs/2308.07326.
[56] LIOU H C, HSIEH H Y. Modeling friendship networks among agents with personality traits[EB/OL]. (2020-04-27)[2025-04-16]. http://arxiv.org/abs/2004.12901.
[57] ZHAN Baohua, HUANG Yongyi, CUI Wenyao, et al. Humanity in AI: detecting the personality of large language models[EB/OL]. (2024-10-11)[2025-04-16]. http://arxiv.org/abs/2410.08545.
[58] KRUIJSSEN J D, EMMONS N. Deterministic AI agent personality expression through standard psychological diagnostics[J]. Allora decentralized intelligence, 2025, 2: 15-39.
[59] KERZ E, QIAO Yu, ZANWAR S, et al. Pushing on personality detection from verbal behavior: a Transformer meets text contours of psycholinguistic features[EB/OL]. (2022-04-10)[2025-04-16]. http://arxiv.org/abs/2204.04629.
[60] JIANG Guangyuan, XU Manjie, ZHU Songchun, et al. Evaluating and inducing personality in pre-trained language models[C]//In Proceedings of the 37th International Conference on Neural Information Processing Systems. Red Hook: NIP, 2023: 10622-10643.
[61] LOTFI E, DE BRUYN M, BUHMANN J, et al. PersonalityChat: conversation distillation for personalized dialog modeling with facts and traits[C]//Proceedings of the Third Workshop on Natural Language Generation, Evaluation, and Metrics. Singapore: ACL, 2023: 353-371.
[62] ORABY S, REED L, SHARATH T S, et al. Neural MultiVoice models for expressing novel personalities in dialog[C]//Interspeech 2018. Hyderabad: ISCA, 2018: 3057-3061.
[63] ZHANG Jian, WANG Zhiyuan, WANG Zhangqi, et al. MAPS: a multi-agent framework based on big seven personality and socratic guidance for multimodal scientific problem solving[EB/OL]. (2025-03-21)[2025-04-16]. http://arxiv.org/abs/2503.16905.
[64] SANTURKAR S, DURMUS E, LADHAK F, et al Whose opinions do language models reflect?[C]//Proceedings of the 40th International Conference on Machine Learning. Honolulu: JMLR, 2023: 29971-30004.
[65] LI Xiyun, NI Ziyi, RUAN Jingqing, et al. Mixture of personality improved spiking actor network for efficient multi-agent cooperation[J]. Frontiers in neuroscience, 2023, 17: 1219405.
[66] ZHONG Wanjun, GUO Lianghong, GAO Qiqi, et al. MemoryBank: enhancing large language models with long-term memory[EB/OL]. (2023-05-17)[2025-04-16]. http://arxiv.org/abs/2305.10250.
[67] SPATHELF M, BENDEL O. The SPACE THEA project[EB/OL]. (2022-06-17)[2025-04-16]. http://arxiv.org/abs/2206.10390.
[68] HE Zihong, ZHANG Changwang. AFSPP: agent framework for shaping preference and personality with large language models[EB/OL]. (2024-01-05)[2025-04-16]. http://arxiv.org/abs/2401.02870.
[69] WANG Weixuan, CAI Xiaoling, HUANG C H, et al. Emily: developing an emotion-affective open-domain chatbot with knowledge graph-based persona[EB/OL]. (2021-09-18)[2025-04-16]. http://arxiv.org/abs/2109.08875.
[70] SANG Yisi, MOU Xiangyang, YU Mo, et al. MBTI personality prediction for fictional characters using movie scripts[EB/OL]. (2022-10-20)[2025-04-16]. http://arxiv.org/abs/2210.10994.
[71] BARUA A, BRASE G, DONG K, et al. On the psychology of GPT-4: moderately anxious, slightly masculine, honest, and humble[EB/OL]. (2024-02-01)[2025-08-20]. http://arxiv.org/abs/2402.01777.
[72] GHAFURIAN M, ELLARD C, DAUTENHAHN K. Social companion robots to reduce isolation: a perception change due to COVID-19[EB/OL]. (2020-08-12)[2025-04-16]. http://arxiv.org/abs/2008.05382.
[73] LIU Yifan, WEI Wei, LIU Jiayi, et al. Improving personality consistency in conversation by persona extending[C]//Proceedings of the 31st ACM International Conference on Information & Knowledge Management. Atlanta: ACM, 2022: 1350-1359.
[74] PARK J, PARK C, LIM H. CharacterGPT: a persona reconstruction framework for role-playing agents[EB/OL]. (2024-05-30)[2025-04-17]. http://arxiv.org/abs/2405.19778.
[75] YE Haoran, JIN Jing, XIE Yuhang, et al. Large language model psychometrics: a systematic review of evaluation, validation, and enhancement[EB/OL]. (2025-05-13)[2025-08-20]. http://arxiv.org/abs/2505.08245.
[76] HUANG J, WANG Wenxuan, LI E J, et al. Who is ChatGPT? Benchmarking LLMs’ psychological portrayal using PsychoBench[EB/OL]. (2023-10-02)[2025-04-16]. http://arxiv.org/abs/2310.01386.
[77] KARINSHAK E, HU A, KONG K, et al. LLM-GLOBE: a benchmark evaluating the cultural values embedded in LLM output[EB/OL]. (2024-11-09)[2025-08-20]. http://arxiv.org/abs/2411.06032.
[78] YE Haoran, XIE Yuhang, REN Yyuanyi, et al. Measuring human and AI values based on generative psychometrics with large language models[C]//Proceedings of the Thirty-Ninth AAAI Conference on Artificial Intelligence. Washington: AAAI Press, 2025: 26400-26408.
[79] CERON T, FALK N, BARI? Ana, et al. Beyond prompt brittleness: evaluating the reliability and consistency of political worldviews in LLMs[J]. Transactions of the association for computational linguistics, 2024, 12: 1378-1400.
[80] LIM M Y, LOPES J D A, ROBB D A, et al. We are all individuals: the role of robot personality and human traits in trustworthy interaction[C]//2022 31st IEEE International Conference on Robot and Human Interactive Communication (RO-MAN). Napoli: IEEE, 2022: 538-545.
[81] SABOUR S, LIU Siyang, ZHANG Zheyuan, et al. EmoBench: evaluating the emotional intelligence of large language models[EB/OL]. (2024-02-19)[2024-12-09]. http://arxiv.org/abs/2402.12071.
[82] HODSON N, WILLIAMSON S. Can large language models replace therapists? evaluating performance at simple cognitive behavioral therapy tasks[J]. JMIR AI, 2024, 3: e52500.
[83] TSENG Y M, HUANG Yuchao, HSIAO T Y, et al. Two tales of persona in LLMs: a survey of role-playing and personalization[EB/OL]. (2024-06-03)[2024-11-28]. http://arxiv.org/abs/2406.01171.
[84] LI Yuan, HUANG Yue, WANG Hongyi, et al. Quantifying AI psychology: a psychometrics benchmark for large language models[EB/OL]. (2024-06-25)[2025-08-20]. http://arxiv.org/abs/2406.17675.
[85] JI Ke, LIAN Yixin, LI Linxu, et al. Enhancing persona consistency for LLMs’ role-playing using persona-aware contrastive learning[C]//Findings of the Association for Computational Linguistics: ACL 2025. Vienna: ACL, 2025: 26221–26238.
[86] DEWEY D. Learning what to value[C]//Artificial General Intelligence. Berlin: Springer Berlin Heidelberg, 2011: 309-314.
[87] ZHU Minjun, WENG Yixuan, YANG Linyi. Personality alignment of large language models[EB/OL]. (2024-08-21)[2025-02-18]. http://arxiv.org/abs/2408.11779.
[88] SZYMANSKI A, ZIEMS N, EICHER-MILLER H A, et al. Limitations of the LLM-as-a-judge approach for evaluating LLM outputs in expert knowledge tasks[C]//Proceedings of the 30th International Conference on Intelligent User Interfaces. Cagliari: ACM, 2025: 952-966.
[89] LI Zhiyuan, LIU Hong, ZHOU Denny, et al. Chain of thought empowers Transformers to solve inherently serial problems[EB/OL]. (2024-02-20)[2025-05-27]. http://arxiv.org/abs/2402.12875.
[90] LU Keming, YU Bowen, ZHOU Chang, et al. Large language models are superpositions of all characters: attaining arbitrary role-play via self-alignment[EB/OL]. (2024-01-23)[2025-05-27]. https://arxiv.org/abs/2401.12474.
[91] ZHENG Junhao, QIU Shengjie, SHI Chengming, et al. Towards lifelong learning of large language models: a survey[EB/OL]. (2024-06-10)[2025-05-27]. http://arxiv.org/abs/2406.06391.
[92] WANG Qian, WU Jiaying, TANG Zhenheng, et al. What limits LLM-based human simulation: LLMs or our design?[EB/OL]. (2025-01-15)[2025-05-27]. http://arxiv.org/abs/2501.08579.
[93] RIEMER M, ASHKTORAB Z, BOUNEFFOUF D, et al. Position: theory of mind benchmarks are broken for large language models[EB/OL]. (2024-12-27)[2025-05-27]. http://arxiv.org/abs/2412.19726.
Similar References:

Memo

-

Last Update: 1900-01-01

Copyright © CAAI Transactions on Intelligent Systems