pith. sign in

arxiv: 2606.19640 · v1 · pith:I7GCBOY6new · submitted 2026-06-17 · 💻 cs.CL · cs.AI· cs.HC

Creating Multilingual Mental Health Dialogue Datasets: Limits of Persona-Based Localization via Nationality and Language

Yunkai Xu , Saeed Abdullah This is my paper

Pith reviewed 2026-06-26 20:26 UTC · model grok-4.3

classification 💻 cs.CL cs.AIcs.HC
keywords multilingual mental health datasetspersona localizationclinical dialogue generationLLM evaluationdepression severity assessmentcultural adaptationsynthetic data limitations
0
0 comments X

The pith

Modifying only nationality and language in English mental health personas produces inconsistent depression severity across languages.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper tests whether English-centric clinical personas can be localized for other languages simply by swapping nationality and language parameters. Researchers generated dialogues in Mandarin, Bengali, and Hindi, then used multiple LLMs to score depression severity against the original English versions. The results show that these minimal changes often create clinical inconsistencies that the same models fail to detect reliably in non-English text. This reveals a systemic limit in current persona-based synthetic data methods for global mental health applications. The work therefore calls for more culturally grounded generation approaches instead of parameter tweaks alone.

Core claim

We modified nationality and language parameters in personas to generate clinical dialogues in Mandarin, Bengali, and Hindi. We then examined how different LLMs perform when evaluating the depression severity of these generated multilingual datasets against the baseline in English. Our findings indicate that just adding nationality and language parameters in personas might not be adequate, as it can introduce clinical inconsistency across languages. LLM judge models often exhibit inaccuracies in assessing depression severity in non-English texts, with performance varying across different models. This exposes the systemic limitations of applying English-centric personas to multilingual context

What carries the argument

The localization method that changes only nationality and language fields inside validated English clinical personas, followed by LLM-based consistency checks on generated dialogue depression severity.

If this is right

  • Simple nationality and language swaps in personas are insufficient to maintain clinical consistency across languages.
  • LLM judges show language-dependent inaccuracies when scoring depression severity, limiting their use for multilingual validation.
  • English-centric persona libraries cannot be directly extended to other languages without additional adaptation steps.
  • Equitable global mental health AI systems require generation methods that incorporate cultural context beyond basic parameters.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Native clinician review or culturally specific instruments may be needed to validate multilingual synthetic data before use.
  • The observed model-to-model variation in non-English scoring suggests that evaluation pipelines themselves need language-specific calibration.
  • Future dataset creation could test whether adding explicit cultural values or symptom expression norms reduces the inconsistencies found here.

Load-bearing premise

That comparing LLM depression-severity scores on non-English dialogues against an English baseline reliably reveals true clinical inconsistency.

What would settle it

A controlled study in which native-speaking clinicians rate the generated Mandarin, Bengali, and Hindi dialogues as having the same depression severity distribution as the English originals, with no detectable inconsistencies.

Figures

Figures reproduced from arXiv: 2606.19640 by Saeed Abdullah, Yunkai Xu.

Figure 1
Figure 1. Figure 1: Overview of the multilingual synthetic dialogue generation and evaluation workflow. Example dialogue [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
read the original abstract

AI and large language models (LLMs) have emerged as promising tools to address global mental health challenges. Despite the global nature of these challenges, there remains a critical shortage of high-quality datasets for training and evaluating such systems. To mitigate this gap, researchers increasingly generate synthetic clinical personas to simulate user data and test digital mental health support systems. However, most validated personas rely on English-centric contexts. This paper investigates whether similar persona-based methods can be used to generate multilingual mental health datasets. We modified nationality and language parameters in personas to generate clinical dialogues in Mandarin, Bengali, and Hindi. We then examined how different LLMs perform when evaluating the depression severity of these generated multilingual datasets against the baseline in English. Our findings indicate that just adding nationality and language parameters in personas might not be adequate, as it can introduce clinical inconsistency across languages. LLM judge models often exhibit inaccuracies in assessing depression severity in non-English texts, with performance varying across different models. This exposes the systemic limitations of applying English-centric personas to multilingual contexts. Ultimately, our work highlights the urgent need for culturally responsive data generation to ensure equitable mental health systems globally.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The manuscript investigates whether modifying English-centric clinical personas with nationality and language parameters can produce consistent multilingual mental health dialogue datasets. Dialogues are generated in Mandarin, Bengali, and Hindi; LLM judges then score depression severity in these outputs against English baselines. The central finding is that persona localization introduces clinical inconsistency across languages and that LLM judges exhibit inaccuracies and model-dependent variation when evaluating non-English text.

Significance. If the methodological concerns are addressed, the work would usefully document practical limits of simple persona localization for clinical data and reinforce the need for culturally grounded dataset creation in mental-health NLP. The empirical focus on three languages and multiple judge models provides a concrete starting point, though the absence of ground-truth validation currently limits the strength of the conclusions.

major comments (2)
  1. [Experimental design / evaluation protocol] The central claim—that nationality/language modifications introduce clinical inconsistency—rests on discrepancies in LLM-judge severity scores between English and target-language dialogues. Because the paper simultaneously reports that the same judges are inaccurate on non-English text, any observed discrepancy could arise from judge failure modes rather than from the persona changes themselves. This circularity is load-bearing for the main conclusion.
  2. No information is supplied on the number of generated dialogues, the precise depression-severity metric or scale employed by the judges, statistical tests for cross-language differences, or any human-expert validation of the LLM judgments. These omissions prevent assessment of whether the reported inconsistencies are reliable or merely artifacts of small or uncontrolled samples.
minor comments (1)
  1. [Abstract] The abstract conflates two distinct observations (inconsistency in generated data vs. judge inaccuracy); separating them would improve clarity.

Simulated Author's Rebuttal

2 responses · 1 unresolved

We thank the referee for the constructive feedback. We address the two major comments point by point below, clarifying our experimental approach and committing to revisions where the manuscript can be strengthened without altering its core findings.

read point-by-point responses

  1. Referee: The central claim—that nationality/language modifications introduce clinical inconsistency—rests on discrepancies in LLM-judge severity scores between English and target-language dialogues. Because the paper simultaneously reports that the same judges are inaccurate on non-English text, any observed discrepancy could arise from judge failure modes rather than from the persona changes themselves. This circularity is load-bearing for the main conclusion.

    Authors: We agree that the evaluation protocol creates interdependence between the measured inconsistencies and judge reliability. Our intent was to demonstrate that persona localization produces outputs whose clinical properties (as scored by LLMs) diverge across languages, while separately documenting judge inaccuracy via model-to-model variation. The central finding is therefore the joint limitation rather than an isolated claim about persona effects. We will revise the manuscript to explicitly discuss this interdependence, reframe the conclusion around the need for improved multilingual evaluation methods, and avoid language that attributes discrepancies solely to persona changes. revision: partial

  2. Referee: No information is supplied on the number of generated dialogues, the precise depression-severity metric or scale employed by the judges, statistical tests for cross-language differences, or any human-expert validation of the LLM judgments. These omissions prevent assessment of whether the reported inconsistencies are reliable or merely artifacts of small or uncontrolled samples.

    Authors: We will add the requested details on the number of generated dialogues, the exact depression-severity metric and scale used by each judge model, and the statistical tests performed for cross-language comparisons. These elements exist in our experimental logs and can be reported in the revision. We did not conduct human-expert validation of the LLM judgments; the study was designed to examine LLM-as-judge behavior rather than to benchmark against clinicians. revision: partial

standing simulated objections not resolved
  • Human-expert validation of the LLM judgments (not performed in the original study)

Circularity Check

0 steps flagged

No circularity; direct empirical reporting of generation and LLM evaluation results

full rationale

The paper is an observational empirical study: personas are modified with nationality/language, dialogues are generated, and LLM judges score depression severity. No equations, fitted parameters, predictions derived from inputs, or self-citation chains are present. Claims rest on experimental observations rather than any derivation that reduces to prior inputs by construction. The noted limitation (LLM judge accuracy in non-English) is a methodological concern but does not constitute circularity under the defined patterns.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The paper relies on standard assumptions from the synthetic data and LLM evaluation literature without introducing new fitted numbers or postulated entities; the central claim rests on the validity of the experimental setup rather than new axioms.

axioms (2)
  • domain assumption Synthetic clinical personas can be localized via nationality and language parameters to produce dialogues in target languages
    This is the core method tested in the paper for generating the multilingual datasets.
  • domain assumption LLM judge models provide a usable proxy for assessing depression severity in generated dialogues
    This underpins the comparison of multilingual outputs against the English baseline.

pith-pipeline@v0.9.1-grok · 5731 in / 1492 out tokens · 28842 ms · 2026-06-26T20:26:01.663334+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

81 extracted references · 68 canonical work pages · 9 internal anchors

  1. [1]

    and Mensa-Kwao, Augustina and Gonese, Gloria and Kamamia, Christine K

    Moitra, Modhurima and Owens, Shanise and Hailemariam, Maji and Wilson, Katherine S. and Mensa-Kwao, Augustina and Gonese, Gloria and Kamamia, Christine K. and White, Belinda and Young, Dorraine M. and Collins, Pamela Y. , urldate =. Global Mental Health: Where We Are and Where We Are Going , volume =. Current Psychiatry Reports , shortjournal =. 2023 , la...

    work page doi:10.1007/s11920-023-01426-8 2023

  2. [2]

    Temporal and spatial trend analysis of all-cause depression burden based on Global Burden of Disease (

    Liu, Junjiao and Liu, Yueyang and Ma, Wenjun and Tong, Yan and Zheng, Jianzhong , urldate =. Temporal and spatial trend analysis of all-cause depression burden based on Global Burden of Disease (. Scientific Reports , shortjournal =. 2024 , langid =. doi:10.1038/s41598-024-62381-9 , abstract =

    work page doi:10.1038/s41598-024-62381-9 2024

  3. [3]

    Tasdik and Anwar, Tasnim and Christopher, Enryka and Hossain, Sahadat and Hossain, Md Mahbub and Koly, Kamrun Nahar and Saif-Ur-Rahman, K

    Hasan, M. Tasdik and Anwar, Tasnim and Christopher, Enryka and Hossain, Sahadat and Hossain, Md Mahbub and Koly, Kamrun Nahar and Saif-Ur-Rahman, K. M. and Ahmed, Helal Uddin and Arman, Nazish and Hossain, Saima Wazed , urldate =. The current state of mental healthcare in Bangladesh: part 1 – an updated country profile , volume =. 2021 , langid =. doi:10....

    work page doi:10.1192/bji.2021.41 2021

  4. [4]

    Proceedings of the 34th

    Wang, Xi and Perez, Anxo and Parapar, Javier and Crestani, Fabio , urldate =. Proceedings of the 34th. 2025 , file =. doi:10.1145/3746252.3761617 , series =

    work page doi:10.1145/3746252.3761617 2025

  5. [5]

    Gender differences in depression and anxiety: The role of age , volume =

    Faravelli, Carlo and Alessandra Scarpato, Maria and Castellini, Giovanni and Lo Sauro, Carolina , urldate =. Gender differences in depression and anxiety: The role of age , volume =. Psychiatry Research , shortjournal =. 2013 , keywords =. doi:10.1016/j.psychres.2013.09.027 , shorttitle =

    work page doi:10.1016/j.psychres.2013.09.027 2013

  6. [6]

    The Journal of Nervous and Mental Disease , author =

    The Role of Age in the Relationship of Gender and Marital Status to Depression , volume =. The Journal of Nervous and Mental Disease , author =. 1982 , langid =

    1982

  7. [7]

    Diagnostic and Statistical Manual of Mental Disorders:

  8. [8]

    Language Patterns Discriminate Mild Depression From Normal Sadness and Euthymic State , volume =

    Smirnova, Daria and Cumming, Paul and Sloeva, Elena and Kuvshinova, Natalia and Romanov, Dmitry and Nosachev, Gennadii , urldate =. Language Patterns Discriminate Mild Depression From Normal Sadness and Euthymic State , volume =. 2018 , keywords =. doi:10.3389/fpsyt.2018.00105 , journal =

    work page doi:10.3389/fpsyt.2018.00105 2018

  9. [9]

    In an Absolute State: Elevated Use of Absolutist Words Is a Marker Specific to Anxiety, Depression, and Suicidal Ideation , volume =

    Al-Mosaiwi, Mohammed and Johnstone, Tom , urldate =. In an Absolute State: Elevated Use of Absolutist Words Is a Marker Specific to Anxiety, Depression, and Suicidal Ideation , volume =. Clinical Psychological Science , publisher =. 2018 , keywords =. doi:10.1177/2167702617747074 , shorttitle =

    work page doi:10.1177/2167702617747074 2018

  10. [10]

    Sociological Methodology , author =

    Linking Life Histories and Mental Health: A Person-Centered Strategy , volume =. Sociological Methodology , author =. 1998 , langid =. doi:10.1111/0081-1750.00041 , shorttitle =

    work page doi:10.1111/0081-1750.00041 1998

  11. [11]

    and Gore, Susan and Colten, Mary Ellen , year =

    Aseltine, Robert H. and Gore, Susan and Colten, Mary Ellen , year =. Depression and the social developmental context of adolescence , volume =. Journal of Personality and Social Psychology , publisher =. doi:10.1037/0022-3514.67.2.252 , pages =

    work page doi:10.1037/0022-3514.67.2.252

  12. [12]

    , urldate =

    Brown, George W. , urldate =. Social Roles, Context and Evolution in the Origins of Depression , volume =. Journal of Health and Social Behavior , publisher =. doi:10.2307/3090203 , pages =

    work page doi:10.2307/3090203

  13. [13]

    , year =

    Bickley, Lynn and Szilagyi, Peter G. , year =. Bates' Guide to Physical Examination and History-Taking , isbn =

  14. [14]

    A meta-analysis of the problematic social media use and mental health , volume =

    Huang, Chiungjung , urldate =. A meta-analysis of the problematic social media use and mental health , volume =. International Journal of Social Psychiatry , shortjournal =. doi:10.1177/0020764020978434 , abstract =

    work page doi:10.1177/0020764020978434

  15. [15]

    and Negy, Charles , urldate =

    Berryman, Chloe and Ferguson, Christopher J. and Negy, Charles , urldate =. Social Media Use and Mental Health among Young Adults , volume =. Psychiatric Quarterly , shortjournal =. 2018 , langid =. doi:10.1007/s11126-017-9535-6 , abstract =

    work page doi:10.1007/s11126-017-9535-6 2018

  16. [16]

    One Persona, Many Cues, Different Results: How Sociodemographic Cues Impact LLM Personalization

    Weeber, Franziska and Neplenbroek, Vera and Batzner, Jan and Padó, Sebastian , urldate =. One Persona, Many Cues, Different Results: How Sociodemographic Cues Impact. 2026 , eprinttype =. doi:10.48550/arXiv.2601.18572 , shorttitle =. 2601.18572 [cs] , keywords =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2601.18572 2026

  17. [17]

    and Craighead, W

    Weiner, Irving B. and Craighead, W. Edward , year =. The Corsini Encyclopedia of Psychology, Volume 1 , isbn =

  18. [18]

    and Zhang, Xiao Chi and Cameron, Kenzie A

    Papanagnou, Dimitrios and Klein, Matthew R. and Zhang, Xiao Chi and Cameron, Kenzie A. and Doty, Amanda and. Developing standardized patient-based cases for communication training: lessons learned from training residents to communicate diagnostic uncertainty , volume =. Advances in Simulation , shortjournal =. 2021 , file =. doi:10.1186/s41077-021-00176-y...

    work page doi:10.1186/s41077-021-00176-y 2021

  19. [19]

    GPT-4 Technical Report

    2024 , eprinttype =. doi:10.48550/arXiv.2303.08774 , abstract =. 2303.08774 [cs] , keywords =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2303.08774 2024

  20. [20]

    The Llama 3 Herd of Models

    Grattafiori, Aaron and Dubey, Abhimanyu and Jauhri, Abhinav and Pandey, Abhinav and others , urldate =. The Llama 3 Herd of Models , url =. 2024 , eprinttype =. doi:10.48550/arXiv.2407.21783 , abstract =. 2407.21783 [cs] , keywords =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2407.21783 2024

  21. [21]

    Qwen3 Technical Report

    Yang, An and Li, Anfeng and Yang, Baosong and others , urldate =. Qwen3 Technical Report , url =. 2025 , eprinttype =. doi:10.48550/arXiv.2505.09388 , abstract =. 2505.09388 [cs] , keywords =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2505.09388 2025

  22. [22]

    DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning

    DeepSeek-AI and Guo, Daya and Yang, Dejian and others , urldate =. Nature , shortjournal =. 2025 , eprinttype =. doi:10.1038/s41586-025-09422-z , shorttitle =. 2501.12948 [cs] , keywords =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1038/s41586-025-09422-z 2025

  23. [23]

    DeepSeek-V3.2: Pushing the Frontier of Open Large Language Models

    DeepSeek-AI and Liu, Aixin and Mei, Aoxue and Lin, Bangcai and Xue, Bing and others , urldate =. 2025 , eprinttype =. doi:10.48550/arXiv.2512.02556 , shorttitle =. 2512.02556 [cs] , keywords =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2512.02556 2025

  24. [24]

    Depression

    Chen, Zhuang and Deng, Jiawen and Zhou, Jinfeng and Wu, Jincenzi and Qian, Tieyun and Huang, Minlie , editor =. Depression. Proceedings of the 2024. doi:10.18653/v1/2024.naacl-long.452 , urldate =

    work page doi:10.18653/v1/2024.naacl-long.452 2024

  25. [25]

    doi:10.48550/arXiv.2501.17510 , urldate =

    Ignashina, Mariia and Bondaronek, Paulina and Santel, Dan and Pestian, John and Ive, Julia , year = 2025, number =. doi:10.48550/arXiv.2501.17510 , urldate =. 2501.17510 , primaryclass =

    work page doi:10.48550/arxiv.2501.17510 2025

  26. [26]

    Customizable

    Li, Yi and Ding, Xuanxuan and Chen, Yifan and Li, Yeye and Ma, Nan , year = 2025, series =. Customizable. Proceedings of the 2025. doi:10.1145/3715336.3735795 , urldate =

    work page doi:10.1145/3715336.3735795 2025

  27. [27]

    Generative

    Zhang, Qiyang and Zhang, Renwen and Xiong, Yiying and Sui, Yuan and Tong, Chang and Lin, Fu-Hung , year = 2025, journal =. Generative. doi:10.2196/78238 , urldate =

    work page doi:10.2196/78238 2025

  28. [28]

    Better to

    Jin, Yiqiao and Chandra, Mohit and Verma, Gaurav and Hu, Yibo and Choudhury, Munmun De and Kumar, Srijan , year = 2023, number =. Better to. doi:10.48550/arXiv.2310.13132 , urldate =. 2310.13132 , primaryclass =

    work page doi:10.48550/arxiv.2310.13132 2023

  29. [29]

    Raihan, Nishat and Puspo, Sadiya Sayara Chowdhury and Bucur, Ana-Maria and Chancellor, Stevie and Zampieri, Marcos , year = 2026, number =. Large. doi:10.48550/arXiv.2602.02440 , urldate =. 2602.02440 , primaryclass =

    work page doi:10.48550/arxiv.2602.02440 2026

  30. [30]

    Synthetic

    Kang, Andrea and Chen, Jun Yu and. Synthetic. doi:10.48550/arXiv.2411.17672 , urldate =. 2411.17672 , primaryclass =

    work page doi:10.48550/arxiv.2411.17672

  31. [31]

    Computers in Biology and Medicine , volume =

    Detecting the Clinical Features of Difficult-to-Treat Depression Using Synthetic Data from Large Language Models , author =. Computers in Biology and Medicine , volume =. doi:10.1016/j.compbiomed.2025.110246 , urldate =

    work page doi:10.1016/j.compbiomed.2025.110246 2025

  32. [32]

    Ge, Tao and Chan, Xin and Wang, Xiaoyang and Yu, Dian and Mi, Haitao and Yu, Dong , year = 2025, number =. Scaling. doi:10.48550/arXiv.2406.20094 , urldate =. 2406.20094 , primaryclass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2406.20094 2025

  33. [33]

    Proceedings of the 2024

    Na, Hongbin , editor =. Proceedings of the 2024

    2024

  34. [34]

    and Chui, Celine Sze Ling and Ip, Patrick , year = 2025, journal =

    Chen, Chen and Lam, Kok Tai and Yip, Ka Man and So, Hung Kwan and Lum, Terry Yat Sang and Wong, Ian Chi Kei and Yam, Jason C. and Chui, Celine Sze Ling and Ip, Patrick , year = 2025, journal =. Comparison of an. doi:10.2196/65785 , urldate =

    work page doi:10.2196/65785 2025

  35. [35]

    The Promise of Generative

    Chakraborty, Tanmoy and Sinha Deb, Koushik and Kulkarni, Himanshu and Masud, Sarah and Math, Suresh Bada and Oke, Gayatri and Sagar, Rajesh and Sharma, Mona , year = 2025, journal =. The Promise of Generative. doi:10.1038/s42256-025-00992-1 , urldate =

    work page doi:10.1038/s42256-025-00992-1 2025

  36. [36]

    Ashraful , editor =

    Ahmed, Istiaq and Mohtasim, Syed Niaz and Arpita, Faiza Omar and Islam, Ashraful and Amin, M. Ashraful , editor =. A. doi:10.1007/978-3-031-78561-0_1 , isbn =

    work page doi:10.1007/978-3-031-78561-0_1

  37. [37]

    R. A. Proceedings of the 28th. doi:10.1145/3340631.3394879 , urldate =

    work page doi:10.1145/3340631.3394879

  38. [38]

    Cooperative

    Gui, Tao and Zhu, Liang and Zhang, Qi and Peng, Minlong and Zhou, Xu and Ding, Keyu and Chen, Zhigang , year = 2019, journal =. Cooperative. doi:10.1609/aaai.v33i01.3301110 , urldate =

    work page doi:10.1609/aaai.v33i01.3301110 2019

  39. [39]

    and Kim, Jinman and Khushi, Matloob , year = 2022, series =

    Naseem, Usman and Dunn, Adam G. and Kim, Jinman and Khushi, Matloob , year = 2022, series =. Early. Proceedings of the. doi:10.1145/3485447.3512128 , urldate =

    work page doi:10.1145/3485447.3512128 2022

  40. [40]

    Overview of

    Parapar, Javier and Perez, Anxo and Wang, Xi and Crestani, Fabio , editor =. Overview of. Experimental. doi:10.1007/978-3-032-04354-2_15 , isbn =

    work page doi:10.1007/978-3-032-04354-2_15

  41. [41]

    Language Resources and Evaluation , volume =

    dos Santos, Wesley Ramos and. Language Resources and Evaluation , volume =. doi:10.1007/s10579-022-09633-0 , urldate =

    work page doi:10.1007/s10579-022-09633-0

  42. [42]

    doi:10.48550/arXiv.2508.12733 , abstract =

    Ning, Zhiyuan and Gu, Tianle and Song, Jiaxin and Hong, Shixin and Li, Lingyu and Liu, Huacan and Li, Jie and Wang, Yixu and Lingyu, Meng and Teng, Yan and Wang, Yingchun , year = 2025, number =. doi:10.48550/arXiv.2508.12733 , urldate =. 2508.12733 , primaryclass =

    work page doi:10.48550/arxiv.2508.12733 2025

  43. [43]

    doi:10.48550/arXiv.2506.19468 , urldate =

    Han, Wenhan and Zhang, Yifan and Chen, Zhixun and Liu, Binbin and Lin, Haobin and Zhang, Bingni and Wang, Taifeng and Pechenizkiy, Mykola and Fang, Meng and Zheng, Yin , year = 2025, number =. doi:10.48550/arXiv.2506.19468 , urldate =. 2506.19468 , primaryclass =

    work page doi:10.48550/arxiv.2506.19468 2025

  44. [44]

    doi:10.48550/arXiv.2503.10497 , urldate =

    Xuan, Weihao and Yang, Rui and Qi, Heli and Zeng, Qingcheng and Xiao, Yunze and Feng, Aosong and Liu, Dairui and Xing, Yun and Wang, Junjue and Gao, Fan and Lu, Jinghui and Jiang, Yuang and Li, Huitao and Li, Xin and Yu, Kunyu and Dong, Ruihai and Gu, Shangding and Li, Yuekang and Xie, Xiaofei and. doi:10.48550/arXiv.2503.10497 , urldate =. 2503.10497 , p...

    work page doi:10.48550/arxiv.2503.10497

  45. [45]

    and Murphy, David and Tabirca, Sabin , year = 2025, journal =

    Ronan, Isabel and Crowley, Patrice and Rombouts, Eva and Cornally, Nicola and Saab, Mohamad M. and Murphy, David and Tabirca, Sabin , year = 2025, journal =. A. doi:10.1016/j.jbi.2025.104936 , urldate =

    work page doi:10.1016/j.jbi.2025.104936 2025

  46. [46]

    Wang, Ke and Zhu, Jiahui and Ren, Minjie and Liu, Zeming and Li, Shiwei and Zhang, Zongye and Zhang, Chenkai and Wu, Xiaoyu and Zhan, Qiqi and Liu, Qingjie and Wang, Yunhong , year = 2024, number =. A. doi:10.48550/arXiv.2410.12896 , urldate =. 2410.12896 , primaryclass =

    work page doi:10.48550/arxiv.2410.12896 2024

  47. [47]

    Zhezherau, Alexey and Yanockin, Alexei , year = 2024, number =. Hybrid. doi:10.48550/arXiv.2410.09168 , urldate =. 2410.09168 , primaryclass =

    work page doi:10.48550/arxiv.2410.09168 2024

  48. [48]

    Frontiers in Psychiatry , volume =

    Leveraging Reddit Data for Context-Enhanced Synthetic Health Data Generation to Identify Low Self Esteem , author =. Frontiers in Psychiatry , volume =. doi:10.3389/fpsyt.2025.1726100 , urldate =

    work page doi:10.3389/fpsyt.2025.1726100 2025

  49. [49]

    doi:10.48550/arXiv.2602.11684 , urldate =

    Sabour, Sahand and NG, TszYam and Huang, Minlie , year = 2026, number =. doi:10.48550/arXiv.2602.11684 , urldate =. 2602.11684 , primaryclass =

    work page doi:10.48550/arxiv.2602.11684 2026

  50. [50]

    Evaluating

    Bhowmik, Shimanto and Dipto, Tawsif Tashwar and Islam, Md Sazzad and Hsu, Sheryl and Reasat, Tahsin , year = 2025, number =. Evaluating. doi:10.48550/arXiv.2507.23248 , urldate =. 2507.23248 , primaryclass =

    work page doi:10.48550/arxiv.2507.23248 2025

  51. [51]

    Personas -

    Nielsen, Lene , year = 2019, publisher =. Personas -

    2019

  52. [52]

    Personas: Practice and Theory , booktitle =

    Pruitt, John and Grudin, Jonathan , year = 2003, series =. Personas: Practice and Theory , booktitle =. doi:10.1145/997078.997089 , urldate =

    work page doi:10.1145/997078.997089 2003

  53. [53]

    Salminen, Joni and Amin, Danial and Jung, Soon-Gyo and Jansen, Bernard , year = 2025, series =. The. Proceedings of the. doi:10.1145/3745900.3746108 , urldate =

    work page doi:10.1145/3745900.3746108 2025

  54. [54]

    Available: https://doi.org/10.1016/j.ijhcs.2025.103445

    PersonaCraft: Leveraging Language Models for Data-Driven Persona developmentPersonaCraft , author =. International Journal of Human-Computer Studies , volume =. doi:10.1016/j.ijhcs.2025.103445 , urldate =

    work page doi:10.1016/j.ijhcs.2025.103445 2025

  55. [55]

    Wu, Shenghan and Zhu, Yimo and Hsu, Wynne and Lee, Mong-Li and Deng, Yang , editor =. From. Proceedings of the 2025. doi:10.18653/v1/2025.emnlp-main.277 , urldate =

    work page doi:10.18653/v1/2025.emnlp-main.277 2025

  56. [56]

    Faithful

    Jandaghi, Pegah and Sheng, Xianghai and Bai, Xinyi and Pujara, Jay and Sidahmed, Hakim , editor =. Faithful. Proceedings of the 6th

  57. [57]

    Synthetic

    Kaur, Arshnoor and Aird, Amanda and Borman, Harris and Nicastro, Andrea and Leontjeva, Anna and Pizzato, Luiz and Jermyn, Dan , year = 2025, pages =. Synthetic. Proceedings of the 33rd. doi:10.1145/3699682.3728339 , urldate =

    work page doi:10.1145/3699682.3728339 2025

  58. [58]

    doi:10.48550/arXiv.2503.16527 , urldate =

    Li, Ang and Chen, Haozhe and Namkoong, Hongseok and Peng, Tianyi , year = 2025, number =. doi:10.48550/arXiv.2503.16527 , urldate =. 2503.16527 , primaryclass =

    work page doi:10.48550/arxiv.2503.16527 2025

  59. [59]

    Batzner, Jan and Stocker, Volker and Tang, Bingjun and Natarajan, Anusha and Chen, Qinhao and Schmid, Stefan and Kasneci, Gjergji , year = 2025, journal =. Whose. doi:10.1609/aies.v8i1.36553 , urldate =

    work page doi:10.1609/aies.v8i1.36553 2025

  60. [60]

    (eds.) Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)

    Zhang, Saizheng and Dinan, Emily and Urbanek, Jack and Szlam, Arthur and Kiela, Douwe and Weston, Jason , editor =. Personalizing. Proceedings of the 56th. doi:10.18653/v1/P18-1205 , urldate =

    work page doi:10.18653/v1/p18-1205

  61. [61]

    and Ma, Yiming and Song, Xingyu and Xu, Xiaohang and Diab, Mona and Li, Irene and Ng, Ka Chung , year = 2026, number =

    Xiao, Yunze and He, Tingyu and Wang, Lionel Z. and Ma, Yiming and Song, Xingyu and Xu, Xiaohang and Diab, Mona and Li, Irene and Ng, Ka Chung , year = 2026, number =. doi:10.48550/arXiv.2503.21679 , urldate =. 2503.21679 , primaryclass =

    work page doi:10.48550/arxiv.2503.21679 2026

  62. [62]

    Agarwal, Kaustubh and Dhingra, Bhavya , editor =. Deep. Proceedings of the 18th

  63. [63]

    PAWS - X : A Cross-lingual Adversarial Dataset for Paraphrase Identification

    Yang, Yinfei and Zhang, Yuan and Tar, Chris and Baldridge, Jason , editor =. Proceedings of the 2019. doi:10.18653/v1/D19-1382 , urldate =

    work page doi:10.18653/v1/d19-1382 2019

  64. [64]

    and Antypas, Dimosthenis and Borkakoty, Hsuvas and Kim, Eunsu and

    Myung, Junho and Lee, Nayeon and Zhou, Yi and Jin, Jiho and Putri, Rifki A. and Antypas, Dimosthenis and Borkakoty, Hsuvas and Kim, Eunsu and. Advances in Neural Information Processing Systems , volume =

  65. [65]

    Restrepo, David and Wu, Chenwei and Tang, Zhengxu and Shuai, Zitao and Phan, Thao Nguyen Minh and Ding, Jun-En and Dao, Cong-Tinh and Gallifant, Jack and Dychiao, Robyn Gayle and Artiaga, Jose Carlo and Bando, Andr. Multi-. Proceedings of the AAAI Conference on Artificial Intelligence , volume =. doi:10.1609/aaai.v39i27.35053 , urldate =

    work page doi:10.1609/aaai.v39i27.35053

  66. [66]

    Generating personas using LLMs and assessing their viability,

    Schuller, Andreas and Janssen, Doris and Blumenr. Generating Personas Using. Extended. doi:10.1145/3613905.3650860 , urldate =

    work page doi:10.1145/3613905.3650860

  67. [67]

    Kamruzzaman, Mahammed and Al Monsur, Abdullah and Kim, Gene Louis and Chhabra, Anshuman , editor =. From. Proceedings of the 14th

  68. [68]

    Exploring

    Kamruzzaman, Mahammed and Kim, Gene Louis , editor =. Exploring. Proceedings of the 2025. doi:10.18653/v1/2025.emnlp-main.181 , urldate =

    work page doi:10.18653/v1/2025.emnlp-main.181 2025

  69. [69]

    Liu, Yang and Iter, Dan and Xu, Yichong and Wang, Shuohang and Xu, Ruochen and Zhu, Chenguang , editor =. G-. Proceedings of the 2023. doi:10.18653/v1/2023.emnlp-main.153 , urldate =

    work page doi:10.18653/v1/2023.emnlp-main.153 2023

  70. [70]

    Fu, Xiyan and Liu, Wei , editor =. How. Findings of the. doi:10.18653/v1/2025.findings-emnlp.587 , urldate =

    work page doi:10.18653/v1/2025.findings-emnlp.587 2025

  71. [71]

    Humans or

    Chen, Guiming Hardy and Chen, Shunian and Liu, Ziche and Jiang, Feng and Wang, Benyou , editor =. Humans or. Proceedings of the 2024. doi:10.18653/v1/2024.emnlp-main.474 , urldate =

    work page doi:10.18653/v1/2024.emnlp-main.474 2024

  72. [72]

    Justice or Prejudice? Quantifying Biases in LLM-as-a-Judge

    Ye, Jiayi and Wang, Yanbo and Huang, Yue and Chen, Dongping and Zhang, Qihui and Moniz, Nuno and Gao, Tian and Geyer, Werner and Huang, Chao and Chen, Pin-Yu and Chawla, Nitesh V. and Zhang, Xiangliang , year = 2024, number =. Justice or. doi:10.48550/arXiv.2410.02736 , urldate =. 2410.02736 , primaryclass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2410.02736 2024

  73. [73]

    Teja, J. S. and Narang, R. L. and Aggarwal, A. K. , year = 1971, journal =. Depression. doi:10.1192/bjp.119.550.253 , urldate =

    work page doi:10.1192/bjp.119.550.253 1971

  74. [74]

    Somatic in the East, Psychological in the West?: Investigating Clinically-Grounded Cross-Cultural Depression Symptom Expression in LLMs

    Sakai, Shintaro and An, Jisun and Kang, Migyeong and Kwak, Haewoon , year = 2025, number =. Somatic in the. doi:10.48550/arXiv.2508.03247 , urldate =. 2508.03247 , primaryclass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2508.03247 2025

  75. [75]

    Framework for

    Peng, Shixin and Jiang, Kun and Yang, Yu and Chen, Jingying and Xu, Guandong , editor =. Framework for. Behavioural and. doi:10.1007/978-981-95-7138-3_26 , isbn =

    work page doi:10.1007/978-981-95-7138-3_26

  76. [76]

    2023 , url =

    Ethnologue: Languages of the World , author =. 2023 , url =

    2023

  77. [77]

    Goodmann and Sariah Daouk and Megan Sullivan and Juan Cabrera and Nancy H

    Danielle R. Goodmann and Sariah Daouk and Megan Sullivan and Juan Cabrera and Nancy H. Liu and Suzanne Barakat and Ricardo F. Muñoz and Yan Leykin , keywords =. Factor analysis of depression symptoms across five broad cultural groups , journal =. 2021 , issn =. doi:https://doi.org/10.1016/j.jad.2020.12.159 , url =

    work page doi:10.1016/j.jad.2020.12.159 2021

  78. [78]

    Depression and anxiety symptoms in adolescents across 30 countries: Cross-national measurement invariance and relationships with subjective well-being , journal =

    Jovanovi. Depression and anxiety symptoms in adolescents across 30 countries: Cross-national measurement invariance and relationships with subjective well-being , journal =. 2026 , issn =. doi:10.1016/j.jad.2026.121693 , url =

    work page doi:10.1016/j.jad.2026.121693 2026

  79. [79]

    Communications Medicine , year=

    Demographic variation in symptoms of depression and anxiety across 22 Global Flourishing Study countries , author=. Communications Medicine , year=

  80. [80]

    Linguistic

    Elwahsh, Sarah and Stern, Nora and Singh, Aneesha and Ayobi, Amid , year = 2025, series =. Linguistic. Proceedings of the 7th. doi:10.1145/3719160.3736615 , url =

    work page doi:10.1145/3719160.3736615 2025

Showing first 80 references.