The reviewed record of science sign in
Pith

arxiv: 2607.06149 · v1 · pith:GAGYAPKW · submitted 2026-07-07 · cs.HC

BlossomPsy: A User-Centric AI System for Adaptive and Engaging MBTI Personality Assessments

Bingjia Huang This is my paper

Reviewed by Pith T0 review T1 audit T2 compute T3 formal T4 kernel 2026-07-08 15:03 UTCglm-5.2pith:GAGYAPKWrecord.jsonopen to challenge →

classification cs.HC
keywords MBTIpersonality assessmentmulti-armed banditadaptive dialoguemulti-head classifierPID controllerhuman-computer interactionLLM-simulated users
0
0 comments X

The pith

AI Chatbot Matches MBTI Scale With Higher User Engagement

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

The paper introduces BlossomPsy, an AI-driven personality assessment system that attempts to make the MBTI test more engaging without sacrificing accuracy. It replaces traditional static questionnaires with a multi-turn conversational agent powered by large language models, supplemented by photo-based questions when confidence is low. The core technical mechanism is a Multi-Head Classifier that predicts MBTI dimensions from text, combined with a modified Upper Confidence Bound algorithm that directs the conversation toward the dimensions the system is least certain about. A PID controller tunes the confidence parameters during training. The author evaluates the system against the Chinese MBTI-M scale using a mix of human volunteers and LLM-simulated participants, reporting moderate-to-strong agreement and higher user satisfaction compared to the traditional test.

Core claim

The central claim is that an adaptive, confidence-driven conversational system can produce MBTI predictions with comparable consistency to a standard questionnaire while delivering a more engaging user experience. The author reports Cohen kappa values ranging from 0.50 to 0.90 across the four MBTI dimensions when comparing BlossomPsy's output to MBTI-M, with users rating the system higher in interactivity and enjoyment. The paper frames these results as preliminary evidence of alignment rather than full psychometric validation.

What carries the argument

A Multi-Head Classifier with a shared RoBERTa encoder and five parallel classification heads (four binary MBTI dimension classifiers plus one 16-class classifier), paired with a modified Upper Confidence Bound algorithm that treats each MBTI dimension as a bandit arm and directs conversation toward the lowest-confidence dimension. A PID controller tunes the sigmoid enhancement function parameters during training. Photo-based questions serve as a fallback when confidence intervals overlap beyond a threshold.

If this is right

  • If adaptive dialogue systems can match questionnaire accuracy while improving engagement, personality assessment could become accessible to populations that find traditional tests tedious, such as adolescents and young adults.
  • The confidence-driven fallback mechanism, which inserts targeted visual questions only when text-based prediction is uncertain, offers a template for reducing assessment length without sacrificing coverage.
  • The framework is designed to be portable to other personality models such as the Big Five, requiring only replacement of binary classification heads with continuous regression heads, though this transfer has not been tested.
  • The photo-generation pipeline, which uses multiple LLMs with human supervision to create and validate visual question items, suggests a semi-automated approach to scale development that could lower the cost of creating new assessment instruments.

Where Pith is reading between the lines

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

  • The evaluation pools 12 human participants with 33 LLM-simulated participants whose MBTI types are assigned by prompting, meaning the agreement metrics may partially reflect the system's ability to recover labels that were used to construct the simulated test-takers rather than genuine predictive accuracy on humans.
  • If the S/N dimension is inherently harder to detect through text because it reflects abstract cognitive preferences, then conversational assessment systems may need fundamentally different question strategies or additional modalities for certain traits, rather than uniform adaptation across all dimensions.
  • The PID controller tunes only two scalar parameters, and the author notes grid search might achieve comparable results, suggesting the control-theoretic machinery may be more complex than the problem requires.

Load-bearing premise

The consistency metrics are computed over a pool of 45 participants that mixes 12 humans with 33 LLM-simulated test-takers whose MBTI types are experimenter-assigned via prompting, which means the agreement numbers may be inflated by the system recovering labels it was indirectly given.

What would settle it

If human-only agreement metrics, computed separately from the LLM-simulated participants, show substantially lower Cohen kappa values than the pooled 0.50-0.90 range, the claim of comparable consistency with MBTI-M would be significantly weakened.

Figures

Figures reproduced from arXiv: 2607.06149 by Bingjia Huang.

Figure 1
Figure 1. Figure 1: BlossomPsy, an MBTI Assessment System That Balances User Engagement with Predictive [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: System Architecture of BlossomPsy, Including Dialogue, Photo-based Questions, and Pre [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Example of Questions from MBTI-M (Chinese version) and BlossomPsy. [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: The Four MBTI Dimensions (E/I, S/N, T/F, J/P) Used in Classification. [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Visualization of Overlap Rate Between mUCB and mLCB Intervals for a Given MBTI Class. [PITH_FULL_IMAGE:figures/full_fig_p006_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Workflow of the Visual Question Module for Low-confidence Fallback Evaluation. [PITH_FULL_IMAGE:figures/full_fig_p007_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Cooperation of Different LLMs in the Photo Generation Module. [PITH_FULL_IMAGE:figures/full_fig_p007_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Test-takers’ Ratings Across Five Evaluation Metrics (Interactivity, Immersion, Clarity, En [PITH_FULL_IMAGE:figures/full_fig_p011_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Results of Candidate Photo-based Questions, Showing the Percentage of Agreement with [PITH_FULL_IMAGE:figures/full_fig_p011_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Performance Comparison of Different System Versions in the Ablation Study. [PITH_FULL_IMAGE:figures/full_fig_p013_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Trend of Parameters Adjustment under PID Control [PITH_FULL_IMAGE:figures/full_fig_p014_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Model Performance under Different α and β Parameter Settings 5 Discussion 5.1 LLM-Simulated Participants: Rationale and Limitations This study employed both human volunteers and LLM-simulated participants in the evaluation. While LLMs do not possess genuine psychological traits, recent research has explored their util￾ity as simulated users in interactive system evaluation. Zheng et al. [36] reported high… view at source ↗
read the original abstract

There has been growing public interest in understanding personality traits and emotional characteristics, as such knowledge helps individuals better accept themselves and manage negative emotions. While professional personality scales remain the standard tool for assessment, they are often perceived as tedious or inaccessible to the general public. AI-driven systems can make assessments more accessible, but it is difficult to balance user engagement with predictive consistency in existing works. We tackle this challenge by introducing BlossomPsy, a user-friendly AI-driven MBTI assessment system. MBTI, a widely recognized but psychometrically debated personality framework, serves as the foundation for many recent systems. BlossomPsy integrates multi-turn dialogue and photo-based questions to enhance user engagement while supporting confidence-aware predictions. By combining deep learning, multi-armed bandit algorithms, and control theory, the system dynamically adapts to users' responses. In particular, photo-based questions are designed to increase interactivity and provide additional user information, thereby improving prediction confidence. Experiments involving both human volunteers and large language models (LLMs) provide preliminary evidence that BlossomPsy can produce stable predictions, with higher reported user satisfaction compared to MBTI-M (Chinese version), while maintaining comparable consistency with the reference scale.

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

3 major / 9 minor

Summary. BlossomPsy is an interactive AI-driven MBTI assessment system that combines multi-turn dialogue, photo-based questions, a Multi-Head Classifier (MHC) with a modified UCB (mUCB) confidence mechanism, and PID-tuned sigmoid enhancement. The paper reports preliminary consistency results (Cohen's kappa 0.50–0.90 across MBTI dimensions) from 45 participants (12 humans + 33 LLM-simulated), user satisfaction ratings favoring BlossomPsy over MBTI-M, an ablation study, and MHC benchmark comparisons. The system is presented as a research prototype, and the authors explicitly frame results as preliminary evidence of alignment rather than full psychometric validation.

Significance. The paper presents a creative integration of multi-armed bandit confidence estimation, PID-controlled parameter tuning, and multi-modal (text + photo) interaction for personality assessment. The photo-based question generation pipeline with iterative LLM-and-human supervision is a novel contribution to interactive assessment design. The authors are commendably transparent about limitations: they explicitly state the system is a research prototype, acknowledge MBTI psychometric debates, and frame results as preliminary. The MHC benchmark comparisons (Table 5) and the ablation study structure provide useful baselines. However, the central consistency claim is substantially weakened by the evaluation methodology, particularly the pooling of human and LLM-simulated participants without separate reporting.

major comments (3)
  1. §4.2.1–4.2.2, Table 2: The pooled 45-participant sample (12 humans + 33 LLMs) makes the reported Cohen's kappa values (0.50–0.90) difficult to interpret. The LLM-simulated participants are prompted with a known MBTI type (Fig. A6, left: 'You are role-playing as a person with the MBTI personality type {{mbti}}'), and the system then attempts to recover this label from the LLM's generated responses. This creates a partial circularity: the ground truth for 73% of the sample is the label assigned by the experimenters' own prompting. The photo-based evaluation is even more explicitly circular (Fig. A6, right: the LLM is told its type and asked to choose photos based on 'your {{pretend_mbti}} personality would most naturally prefer'). The paper acknowledges LLM construct validity concerns in §5.1 but does not report human-only consistency metrics. A human-only sub-analysis (even with n=12) is载
  2. §4.3.2, Fig. 10 (Ablation Study): The ablation study uses exclusively LLM-simulated participants, meaning the claimed contributions of PID, mUCB, and photo-based modules are evaluated entirely under the circular ground-truth condition described above. The ablation results should be interpreted with this limitation clearly stated, or ideally supplemented with human data for at least the full-system condition. Without this, the component-level claims (e.g., 'PID and photo-based questions were associated with higher prediction accuracy') are not independently validated on real users.
  3. §3.3, Eq. (4): The cumulative reward update ĥ_i(t) = [T_i(t-1) × ĥ_i(t-1) + mUCB_i(t)] / T_i(t) is recursive in mUCB_i(t), which itself depends on ĥ_i(t-1) via Eq. (2). This creates a self-referential loop that is not standard in UCB formulations. The authors should clarify whether this is intentional and how convergence is affected, or correct the formulation if this is a notational error.
minor comments (9)
  1. §4.2.3, Fig. 8: It is unclear whether the user satisfaction ratings were collected from the 12 human participants only or from all 45 (including LLM-simulated). This should be specified explicitly.
  2. §4.2.4, Table 3: The phrase 'Average Convergent Rate' is non-standard; consider 'Average Agreement Rate' or 'Average Convergence Rate' with a clear definition.
  3. Table 4: MHC performance on Personality Cafe without fine-tuning (F1 ≈ 0.49–0.51) is near chance for binary classification, yet this is not discussed. The large improvement after fine-tuning (0.84–0.92) suggests the fine-tuning dataset may be small or the base model is not learning generalizable features. This warrants brief discussion.
  4. §3.5.2: The simulated user noise injection (20% posts from other MBTI types) is described, but the sensitivity of results to this 20% rate is not analyzed. A brief note on robustness would strengthen the claim that this simulates real-world inconsistency.
  5. §5.4: The authors acknowledge that simpler alternatives like grid search may achieve comparable results for PID parameter tuning. Given this, the motivation for PID over grid search should be strengthened, or the PID contribution should be framed more modestly.
  6. Table A4: The distribution of test-takers across MBTI types is shown, but it is unclear which entries are human vs. LLM-simulated. This breakdown should be added to enable assessment of human-only type coverage.
  7. §3.3, Eq. (1): The sigmoid enhancement function f(p_i) = 1/(1 + α·e^{-β·p_i}) is described as 'sigmoid-like' but differs from the standard sigmoid. The motivation for this specific form over a standard sigmoid should be stated more clearly.
  8. Fig. A6: The prompts for LLM-simulated test-takers are provided in the appendix, which is good for transparency. However, the main text (§4.2.1) should briefly summarize the prompting approach and its implications for evaluation validity, rather than relegating this entirely to the appendix.
  9. The paper would benefit from a brief note on whether the 13 text-based questions selected for photo conversion (§4.2.4) were reviewed by anyone with psychometric expertise, or whether the selection was purely based on the stated heuristic rules.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for a careful and constructive review. The referee raises three major points: (1) the pooling of human and LLM-simulated participants in the consistency evaluation creates interpretive difficulties, particularly partial circularity for the LLM-simulated subset; (2) the ablation study relies exclusively on LLM-simulated participants, so component-level claims are not independently validated on real users; and (3) the mUCB cumulative reward update in Eq. (4) appears self-referential because mUCB_i(t) depends on ĥ_i(t-1) via Eq. (2), creating a non-standard recursive loop. We address each point below.

read point-by-point responses

  1. Referee: §4.2.1–4.2.2, Table 2: The pooled 45-participant sample (12 humans + 33 LLMs) makes the reported Cohen's kappa values (0.50–0.90) difficult to interpret. The LLM-simulated participants are prompted with a known MBTI type, creating a partial circularity. The paper does not report human-only consistency metrics. A human-only sub-analysis (even with n=12) is requested.

    Authors: The referee is correct that pooling human and LLM-simulated participants without separate reporting makes the consistency metrics difficult to interpret, and that the LLM-simulated condition involves a degree of circularity: the ground-truth label for 73% of the sample is the MBTI type assigned via the role-playing prompt, and the system then attempts to recover that label from the LLM's generated responses. We agree that human-only metrics should be reported separately. We will add a human-only sub-analysis (n=12) to the revised manuscript, reporting per-dimension accuracy, F1, and Cohen's kappa for the 12 human participants alone, alongside the pooled results. We will also add explicit discussion of the circularity concern for the LLM-simulated subset, making clear that the LLM results should be interpreted as a pipeline stress-test (can the system recover a known type end-to-end?) rather than as independent psychometric evidence. The pooled metrics will be retained for completeness but clearly flagged as combining two evaluation modalities with different epistemic status. revision: yes

  2. Referee: §4.3.2, Fig. 10 (Ablation Study): The ablation study uses exclusively LLM-simulated participants, meaning the claimed contributions of PID, mUCB, and photo-based modules are evaluated entirely under the circular ground-truth condition. Component-level claims are not independently validated on real users.

    Authors: This is a fair and accurate observation. The ablation study as currently reported uses only LLM-simulated participants, so the component-level claims (e.g., that PID and photo-based questions are associated with higher prediction accuracy) are validated only under the simulated condition where ground truth is assigned by prompting. We will revise the manuscript in two ways. First, we will add an explicit caveat in §4.3.2 stating that the ablation results are obtained under the LLM-simulated condition and should be interpreted as a controlled pipeline test rather than as independent validation on real users. Second, we will run the full-system condition and at least one ablation condition (photo-based questions removed) on the 12 human participants and report the comparison in the revised manuscript. We acknowledge that n=12 provides limited statistical power, but it will at least provide an initial human-data check on whether the direction of the ablation effect is consistent with the simulated results. We will be transparent that this is preliminary rather than definitive. revision: yes

  3. Referee: §3.3, Eq. (4): The cumulative reward update ĥ_i(t) = [T_i(t-1) × ĥ_i(t-1) + mUCB_i(t)] / T_i(t) is recursive in mUCB_i(t), which itself depends on ĥ_i(t-1) via Eq. (2). This creates a self-referential loop that is not standard in UCB formulations. Clarify whether this is intentional and how convergence is affected, or correct the formulation if this is a notational error.

    Authors: We thank the referee for identifying this issue. Upon careful review, we confirm that Eq. (4) as written is a notational error. In a standard UCB formulation, the cumulative reward update should use the observed reward (the MHC's enhanced logit score for the selected dimension at round t), not the mUCB value itself. The mUCB value in Eq. (2) is an exploration-augmented upper bound used for arm selection, not the reward signal that should be accumulated into the running mean. The intended update should be: ĥ_i(t) = [T_i(t-1) × ĥ_i(t-1) + r_i(t)] / T_i(t), where r_i(t) is the enhanced logit score f(p_i) from Eq. (1) for the selected dimension at round t. This removes the self-referential loop. We will correct Eq. (4) accordingly, update the surrounding text to clarify the distinction between the reward signal and the UCB exploration bonus, and verify that the implementation matches the corrected formulation. We will also add a brief remark on convergence: since the reward r_i(t) is bounded (as the output of a sigmoid function) and the update is a standard running average, the cumulative reward estimate converges under the standard assumptions for UCB-style algorithms. revision: yes

Circularity Check

0 steps flagged

33 of 45 evaluation participants are LLM-simulated with ground-truth MBTI labels injected via prompt; the system then 'predicts' those same labels, and pooled consistency metrics (Tab. 2) conflate this circular subset with 12 human participants.

full rationale

The paper's central consistency claim (Cohen's kappa 0.50–0.90 in Tab. 2) is computed over all 45 participants, but 33 are LLM-simulated participants whose 'true' MBTI type is the label injected via the role-playing prompt (Fig. A6). The system then attempts to recover this label from the LLM's generated responses, making the ground truth for 73% of the sample circular by construction. The photo-based question evaluation is even more explicitly circular: the LLM is told its type and asked to choose photos based on 'your {{pretend_mbti}} personality would most naturally prefer.' The ablation study (Fig. 10) uses exclusively LLM-simulated participants, so the claimed contributions of PID, mUCB, and photo-based modules are also evaluated under this circular setup. The paper acknowledges LLM construct validity concerns in §5.1 but does not report human-only consistency metrics, making it impossible to assess how much of the reported agreement comes from the circular LLM subset versus genuine human performance. The 12 human participants provide some independent data, but the pooled metrics are dominated by the circular majority, and the paper does not separate them. This is partial circularity: the central claim is not entirely circular (12 humans contribute non-circular data), but the majority of the evidence is forced by construction. The MHC training on Personality Cafe (self-reported MBTI labels) and evaluation against MBTI-M is a separate concern about label quality but is not itself circular in the definitional sense, since the training labels and the evaluation reference scale are different instruments. No self-citation chain is load-bearing; the circularity is methodological, not citation-based. Score 6 reflects that one or more 'predictions' reduce by construction for the LLM subset, while the human subset retains independent content that the paper does not isolate or report separately.

Axiom & Free-Parameter Ledger

6 free parameters · 5 axioms · 2 invented entities

The system introduces 6 free parameters (several hand-set), relies on 5 domain/ad-hoc assumptions (most critically that LLM-simulated participants approximate human behavior and that self-reported MBTI labels are valid training signal), and invents 2 entities (mUCB, photo-based items) that lack independent external validation. The PID controller is acknowledged by the authors as potentially unnecessary.

free parameters (6)
  • alpha (sigmoid enhancement) = converges to ~15.6 region (Fig. 11a)
    Parameter in the sigmoid-like enhancement function f(p_i) = 1/(1+alpha*exp(-beta*p_i)), tuned by PID controller during training on simulated data.
  • beta (sigmoid enhancement) = determined by alpha via constraint f(0.5)=0.5
    Second parameter of the enhancement function, linked to alpha by the constraint that the function passes through (0.5, 0.5).
  • overlap threshold = 0.6
    Threshold for triggering photo-based questions when mUCB/mLCB overlap rate exceeds this value. Set by hand.
  • PID gains (Kp, Ki, Kd) = Kp=1, Ki=0.02, Kd=0.01
    PID controller gains for tuning alpha and beta, specified in Algorithm 1.
  • photo agreement threshold = 0.66
    Threshold for retaining photo-based questions based on agreement with text counterparts.
  • reward values (r_correct, r_incorrect) = r_correct=0.5, r_incorrect=0
    Reward signal values used in PID training, specified in Algorithm 1.
axioms (5)
  • domain assumption MBTI dimensions can be reliably predicted from text-based conversational responses
    The entire MHC+mUCB pipeline assumes that user text in multi-turn dialogue carries sufficient signal for MBTI prediction. The paper acknowledges S/N is hard to detect via text (§5.2) but the system architecture depends on this assumption globally.
  • domain assumption LLM-simulated participants exhibit stable, measurable personality patterns that approximate human test-taker behavior
    Invoked in §4.2.1 to justify pooling 33 LLM-simulated participants with 12 human volunteers. The paper cites Serapio-García et al. [37] and Zheng et al. [36] but also acknowledges counter-evidence [33, 49, 50].
  • domain assumption Self-reported MBTI labels in the Personality Cafe dataset are valid ground truth for training
    The MHC is trained on Personality Cafe posts with self-reported MBTI labels (§3.5.1). MBTIBench [32] is cited as addressing noisy self-reported labels, but the training still relies on this data.
  • ad hoc to paper PID feedback control is an appropriate mechanism for tuning confidence transformation parameters
    The use of a PID controller (Algorithm 1) to tune alpha and beta is a design choice. The authors themselves note 'simpler search strategies may achieve similar results' (§4.3.3, §5.4).
  • domain assumption Photo-based questions probe the same personality construct as their text-based counterparts
    The photo-based module assumes visual stimuli can substitute for text items in measuring MBTI dimensions. The 80% consistency threshold (§3.4) is used as evidence, but no psychometric validation of construct equivalence is provided.
invented entities (2)
  • mUCB (modified Upper Confidence Bound) no independent evidence
    purpose: Adapts UCB bandit to MBTI dimension selection by using enhanced logits and overlap rate for confidence estimation
    The modification is specific to this paper. The ablation study (Fig. 10) provides internal evidence of contribution, but no external benchmark or falsifiable prediction outside the system is provided.
  • Photo-based MBTI question items no independent evidence
    purpose: Fallback evaluation mechanism for low-confidence dimensions, designed to increase engagement and provide additional personality signal
    The photo items are generated by LLMs and validated only by internal consistency (agreement with text counterparts). No psychometric validation of construct validity is provided.

pith-pipeline@v1.1.0-glm · 20803 in / 3337 out tokens · 486438 ms · 2026-07-08T15:03:46.785408+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

64 extracted references · 64 canonical work pages · 12 internal anchors

  1. [1]

    Enhancing mental health with artificial intelligence: Current trends and future prospects.Journal of medicine, surgery, and public health, 3:100099,

    David B Olawade, Ojima Z Wada, Aderonke Odetayo, Aanuoluwapo Clement David-Olawade, Fiyinfoluwa Asaolu, and Judith Eberhardt. Enhancing mental health with artificial intelligence: Current trends and future prospects.Journal of medicine, surgery, and public health, 3:100099,

    work page

  2. [2]

    doi: 10.1016/j.glmedi.2024.100099. 15

    work page doi:10.1016/j.glmedi.2024.100099 2024

  3. [3]

    Personality and personality measurement

    Robert T Hogan. Personality and personality measurement. InHandbook of Industrial and Organizational Psychology. Consulting Psychologists Press, 1991

    work page 1991

  4. [4]

    Laura Bohane, Nick Maguire, and Thomas Richardson. Resilients, overcontrollers and un- dercontrollers: A systematic review of the utility of a personality typology method in under- standing adult mental health problems.Clinical psychology review, 57:75–92, 2017. doi: 10.1016/j.cpr.2017.07.005

    work page doi:10.1016/j.cpr.2017.07.005 2017

  5. [5]

    Personality testing and industrial–organizational psychology: Reflections, progress, and prospects.Industrial and Organizational Psychology, 1(3):272–290, 2008

    Leaetta M Hough and Frederick L Oswald. Personality testing and industrial–organizational psychology: Reflections, progress, and prospects.Industrial and Organizational Psychology, 1(3):272–290, 2008. doi: 10.1111/j.1754-9434.2008.00048.x

    work page doi:10.1111/j.1754-9434.2008.00048.x 2008

  6. [6]

    The effectiveness of sparx, a computerised self help intervention for adolescents seeking help for depression: randomised controlled non-inferiority trial.Bmj, 344,

    Sally N Merry, Karolina Stasiak, Matthew Shepherd, Chris Frampton, Theresa Fleming, and Mathijs FG Lucassen. The effectiveness of sparx, a computerised self help intervention for adolescents seeking help for depression: randomised controlled non-inferiority trial.Bmj, 344,

    work page

  7. [7]

    doi: 10.1136/bmj.e2598

    work page doi:10.1136/bmj.e2598

  8. [8]

    Western Psychological Services, 1966

    John N Buck.The house-tree-person technique: Revised manual. Western Psychological Services, 1966

    work page 1966

  9. [9]

    Measuring individual differences in implicit cognition: the implicit association test.Journal of personality and social psychology, 74(6):1464, 1998

    Anthony G Greenwald, Debbie E McGhee, and Jordan LK Schwartz. Measuring individual differences in implicit cognition: the implicit association test.Journal of personality and social psychology, 74(6):1464, 1998. doi: 10.1037/0022-3514.74.6.1464

    work page doi:10.1037/0022-3514.74.6.1464 1998

  10. [10]

    Validation of a mobile game-based assess- ment of cognitive control among children and adolescents.PLoS One, 15(3):e0230498, 2020

    Hyunjoo Song, Do-Joon Yi, and Hae-Jeong Park. Validation of a mobile game-based assess- ment of cognitive control among children and adolescents.PLoS One, 15(3):e0230498, 2020. doi: 10.1371/journal.pone.0230498

    work page doi:10.1371/journal.pone.0230498 2020

  11. [11]

    A Computational Approach to Understanding Empathy Expressed in Text-Based Mental Health Support

    Ashish Sharma, Adam S. Miner, David C. Atkins, and Tim Althoff. A computational approach to understanding empathy expressed in text-based mental health support, 2020. URLhttps: //arxiv.org/abs/2009.08441

    work page internal anchor Pith review Pith/arXiv arXiv 2020

  12. [12]

    A shoulder to cry on: Towards a motivational virtual assistant for assuaging mental agony

    Tulika Saha, Saichethan Reddy, Anindya Das, Sriparna Saha, and Pushpak Bhattacharyya. A shoulder to cry on: Towards a motivational virtual assistant for assuaging mental agony. InProceedings of the 2022 conference of the North American chapter of the association for computational linguistics: Human language technologies, pages 2436–2449, 2022

    work page 2022

  13. [13]

    Large language models can infer psychological disposi- tions of social media users.PNAS nexus, 3(6):pgae231, 2024

    Heinrich Peters and Sandra C Matz. Large language models can infer psychological disposi- tions of social media users.PNAS nexus, 3(6):pgae231, 2024

    work page 2024

  14. [14]

    Psychometric evaluation of large language model embeddings for personality trait prediction.Journal of Medical Internet Research, 27:e75347, 2025

    Julina Maharjan, Ruoming Jin, Jianfeng Zhu, and Deric Kenne. Psychometric evaluation of large language model embeddings for personality trait prediction.Journal of Medical Internet Research, 27:e75347, 2025

    work page 2025

  15. [15]

    Screening for depression using natural language processing: literature review.Interactive Journal of Medical Research, 13(1):e55067, 2024

    Bazen Gashaw Teferra, Alice Rueda, Hilary Pang, Richard Valenzano, Reza Samavi, Srid- har Krishnan, and Venkat Bhat. Screening for depression using natural language processing: literature review.Interactive Journal of Medical Research, 13(1):e55067, 2024

    work page 2024

  16. [16]

    Natural language processing applied to mental illness detection: a narrative review.NPJ digital medicine, 5(1): 46, 2022

    Tianlin Zhang, Annika M Schoene, Shaoxiong Ji, and Sophia Ananiadou. Natural language processing applied to mental illness detection: a narrative review.NPJ digital medicine, 5(1): 46, 2022

    work page 2022

  17. [17]

    Scoping review on natural language processing applications in counselling and psychotherapy.British Journal of Psychology, 117 (2):677–701, 2024

    Maria Laricheva, Yan Liu, Edward Shi, and Amery Wu. Scoping review on natural language processing applications in counselling and psychotherapy.British Journal of Psychology, 117 (2):677–701, 2024. doi: 10.1111/bjop.12721

    work page doi:10.1111/bjop.12721 2024

  18. [18]

    The applications of large language models in mental health: Scoping review.Journal of Medical Internet Research, 27:e69284, 2025

    Yu Jin, Jiayi Liu, Pan Li, Baosen Wang, Yangxinyu Yan, Huilin Zhang, Chenhao Ni, Jing Wang, Yi Li, Yajun Bu, et al. The applications of large language models in mental health: Scoping review.Journal of Medical Internet Research, 27:e69284, 2025

    work page 2025

  19. [19]

    Language models and psychological sciences.Frontiers in Psychology, 14:1279317, 2023

    Giuseppe Sartori and Graziella Orrù. Language models and psychological sciences.Frontiers in Psychology, 14:1279317, 2023. doi: 10.3389/fpsyg.2023.1279317. 16

    work page doi:10.3389/fpsyg.2023.1279317 2023

  20. [20]

    Empowering Psychotherapy with Large Language Models: Cognitive Distortion Detection through Diagnosis of Thought Prompting

    Zhiyu Chen, Yujie Lu, and William Yang Wang. Empowering psychotherapy with large lan- guage models: Cognitive distortion detection through diagnosis of thought prompting.arXiv preprint arXiv:2310.07146, 2023

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  21. [21]

    Daun Shin, Hyoseung Kim, Seunghwan Lee, Younhee Cho, and Whanbo Jung. Using large language models to detect depression from user-generated diary text data as a novel approach in digital mental health screening: instrument validation study.Journal of Medical Internet Research, 26:e54617, 2024

    work page 2024

  22. [22]

    CPsyCoun: A Report-based Multi-turn Dialogue Reconstruction and Evaluation Framework for Chinese Psychological Counseling

    Chenhao Zhang, Renhao Li, Minghuan Tan, Min Yang, Jingwei Zhu, Di Yang, Jiahao Zhao, Guancheng Ye, Chengming Li, and Xiping Hu. Cpsycoun: A report-based multi-turn dialogue reconstruction and evaluation framework for chinese psychological counseling.arXiv preprint arXiv:2405.16433, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  23. [23]

    Beyond empathy: Integrating diagnostic and therapeutic reasoning with large language models for mental health counseling.arXiv preprint arXiv:2505.15715, 2025

    He Hu, Yucheng Zhou, Juzheng Si, Qianning Wang, Hengheng Zhang, Fuji Ren, Fei Ma, and Laizhong Cui. Beyond empathy: Integrating diagnostic and therapeutic reasoning with large language models for mental health counseling.arXiv preprint arXiv:2505.15715, 2025

    work page arXiv 2025

  24. [24]

    PsychoGAT: A Novel Psychological Measurement Paradigm through Interactive Fiction Games with LLM Agents

    Qisen Yang, Zekun Wang, Honghui Chen, Shenzhi Wang, Yifan Pu, Xin Gao, Wenhao Huang, Shiji Song, and Gao Huang. Psychogat: A novel psychological measurement paradigm through interactive fiction games with llm agents.arXiv preprint arXiv:2402.12326, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  25. [25]

    PsyDI: Towards a Personalized and Progressively In-depth Chatbot for Psychological Measurements

    Xueyan Li, Xinyan Chen, Yazhe Niu, Shuai Hu, and Yu Liu. Psydi: Towards a personalized and progressively in-depth chatbot for psychological measurements, 2025. URLhttps:// arxiv.org/abs/2408.03337

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  26. [26]

    Sage publications, 2021

    Robert F DeVellis and Carolyn T Thorpe.Scale development: Theory and applications. Sage publications, 2021

    work page 2021

  27. [27]

    Enhanc- ing mental health support through human-ai collaboration: Toward secure and empathetic ai- enabled chatbots

    Rawan AlMakinah, Andrea Norcini-Pala, Lindsey Disney, and M Abdullah Canbaz. Enhanc- ing mental health support through human-ai collaboration: Toward secure and empathetic ai- enabled chatbots. In2025 IEEE Conference on Artificial Intelligence (CAI), pages 196–202. IEEE, 2025

    work page 2025

  28. [28]

    Evaluating generative ai in mental health: Systematic review of capabili- ties and limitations.JMIR mental health, 12(1):e70014, 2025

    Liying Wang, Tanmay Bhanushali, Zhuoran Huang, Jingyi Yang, Sukriti Badami, and Lisa Hightow-Weidman. Evaluating generative ai in mental health: Systematic review of capabili- ties and limitations.JMIR mental health, 12(1):e70014, 2025

    work page 2025

  29. [29]

    Finite-time analysis of the multiarmed bandit problem.Machine learning, 47(2):235–256, 2002

    Peter Auer, Nicolo Cesa-Bianchi, and Paul Fischer. Finite-time analysis of the multiarmed bandit problem.Machine learning, 47(2):235–256, 2002

    work page 2002

  30. [30]

    Directional stability of automatically steered bodies.Journal of the Ameri- can Society for Naval Engineers, 34(2):280–309, 1922

    Nicolas Minorsky. Directional stability of automatically steered bodies.Journal of the Ameri- can Society for Naval Engineers, 34(2):280–309, 1922

    work page 1922

  31. [31]

    Preliminary revision of the myers-briggs type indicator (mbti).Applied Psychology, 7(2):33–37, 2001

    Huajian Cai, Zhenwen Zhu, and Zhiliang Yang. Preliminary revision of the myers-briggs type indicator (mbti).Applied Psychology, 7(2):33–37, 2001

    work page 2001

  32. [32]

    Consulting Psycholo- gists Press Palo Alto, CA, 1962

    Isabel Briggs Myers et al.The myers-briggs type indicator, volume 34. Consulting Psycholo- gists Press Palo Alto, CA, 1962

    work page 1962

  33. [33]

    Gosling, Peter J

    Samuel D. Gosling, Peter J. Rentfrow, and William B. Swann. A very brief measure of the big-five personality domains.Journal of Research in Personality, 37(6):504–528, 2003. doi: 10.1016/S0092-6566(03)00046-1

    work page doi:10.1016/s0092-6566(03)00046-1 2003

  34. [34]

    Can Large Language Models Understand You Better? An MBTI Personality Detection Dataset Aligned with Population Traits

    Bohan Li, Jiannan Guan, Longxu Dou, Yunlong Feng, Dingzirui Wang, Yang Xu, Enbo Wang, Qiguang Chen, Bichen Wang, Xiao Xu, Yimeng Zhang, Libo Qin, Yanyan Zhao, Qingfu Zhu, and Wanxiang Che. Can large language models understand you better? an mbti personality detection dataset aligned with population traits.arXiv preprint arXiv:2412.12510, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  35. [35]

    Jianfeng Zhu, Ruoming Jin, and Karin G. Coifman. Can llms infer personality from real world conversations?arXiv preprint arXiv:2507.14355, 2025

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  36. [36]

    Pittenger

    David J. Pittenger. Measuring the MBTI. . . and coming up short.Journal of Career Planning and Employment, 54(1):48–52, 1993. 17

    work page 1993

  37. [37]

    McCrae and Paul T

    Robert R. McCrae and Paul T. Costa. Reinterpreting the Myers-Briggs Type Indicator from the perspective of the five-factor model of personality.Journal of Personality, 57(1):17–40, 1989. doi: 10.1111/j.1467-6494.1989.tb00759.x

    work page doi:10.1111/j.1467-6494.1989.tb00759.x 1989

  38. [38]

    Xing, Hao Zhang, Joseph E

    Lianmin Zheng, Wei-Lin Chiang, Ying Sheng, Siyuan Zhuang, Zhanghao Wu, Yonghao Zhuang, Zi Lin, Zhuohan Li, Dacheng Li, Eric P. Xing, Hao Zhang, Joseph E. Gonzalez, and Ion Stoica. Judging LLM-as-a-judge with MT-Bench and Chatbot Arena. InAdvances in Neural Information Processing Systems 36 (NeurIPS), 2023

    work page 2023

  39. [39]

    Personality Traits in Large Language Models

    Greg Serapio-García, Mustafa Safdari, Clément Crepy, Luning Sun, Stephen Fitz, Peter Romero, Marwa Abdulhai, Aleksandra Faust, and Maja Matari ´c. Personality traits in large language models.arXiv preprint arXiv:2307.00184, 2023

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  40. [40]

    Personality testing of large language models: limited temporal stability, but highlighted prosociality.Royal Society Open Science, 11(10):240180, 2024

    Bojana Bodroža, Bojana M Dini ´c, and Ljubiša Boji ´c. Personality testing of large language models: limited temporal stability, but highlighted prosociality.Royal Society Open Science, 11(10):240180, 2024

    work page 2024

  41. [41]

    A comparative study of large language models and human personality traits, 2025

    Jiaqi Wang, Bo Wang, Fa Guo, Cheng Cheng, and Li Yang. A comparative study of large language models and human personality traits, 2025

    work page 2025

  42. [42]

    Doubao-1.5-pro technical document

    Seed. Doubao-1.5-pro technical document. Technical report, ByteDance, 2025. URLhttps: //seed.bytedance.com/en/special/doubao_1_5_pro. Access time: 2025-09-07

    work page 2025

  43. [43]

    (mbti) myers-briggs personality type.https://www.kaggle.com/datasets/ datasnaek/mbti-type, 2017

    Mitchell J. (mbti) myers-briggs personality type.https://www.kaggle.com/datasets/ datasnaek/mbti-type, 2017. Access time: 2025-07-01

    work page 2017

  44. [44]

    A Chinese Multi-label Affective Computing Dataset Based on Social Media Network Users

    Jingyi Zhou, Senlin Luo, and Haofan Chen. A chinese multi-label affective computing dataset based on social media network users.arXiv preprint arXiv:2411.08347, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  45. [45]

    Adaptive mix- tures of local experts.Neural computation, 3(1):79–87, 1991

    Robert A Jacobs, Michael I Jordan, Steven J Nowlan, and Geoffrey E Hinton. Adaptive mix- tures of local experts.Neural computation, 3(1):79–87, 1991

    work page 1991

  46. [46]

    Revisiting Pre-Trained Models for Chinese Natural Language Processing

    Yiming Cui, Wanxiang Che, Ting Liu, Bing Qin, Shijin Wang, and Guoping Hu. Revisiting pre-trained models for chinese natural language processing.arXiv preprint arXiv:2004.13922, 2020

    work page internal anchor Pith review Pith/arXiv arXiv 2004

  47. [47]

    Monochrome males and colorful females: Do gender and age influence the color and content of drawings?SAGE Open, 3(4):2158244013509254, 2013

    Lynn Wright and Fiona Black. Monochrome males and colorful females: Do gender and age influence the color and content of drawings?SAGE Open, 3(4):2158244013509254, 2013. doi: 10.1177/2158244013509254. URLhttps://doi.org/10.1177/2158244013509254

    work page doi:10.1177/2158244013509254 2013

  48. [48]

    The role of personality characteristics in informing our preference for visual presentation: An eye movement study.J

    Hosam Al-Samarraie, Samer Muthana Sarsam, Ahmed Ibrahim Alzahrani, Nasser Alalwan, and Mona Masood. The role of personality characteristics in informing our preference for visual presentation: An eye movement study.J. Ambient Intell. Smart Environ., 8(6):709–719, January 2016. ISSN 1876-1364. doi: 10.3233/AIS-160399. URLhttps://doi.org/10. 3233/AIS-160399

    work page doi:10.3233/ais-160399 2016

  49. [49]

    Shuai Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, Sibo Song, Kai Dang, Peng Wang, Shijie Wang, Jun Tang, et al. Qwen2. 5-vl technical report.arXiv preprint arXiv:2502.13923, 2025

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  50. [50]

    Nahathai Wongpakaran, Tinakon Wongpakaran, Danny Wedding, and Kilem L Gwet. A com- parison of cohen’s kappa and gwet’s ac1 when calculating inter-rater reliability coefficients: a study conducted with personality disorder samples.BMC medical research methodology, 13 (1):61, 2013. doi: 10.1186/1471-2288-13-61

    work page doi:10.1186/1471-2288-13-61 2013

  51. [51]

    Evalu- ating personality traits in large language models: Insights from psychological questionnaires

    Pranav Bhandari, Usman Naseem, Amitava Datta, Nicolas Fay, and Mehwish Nasim. Evalu- ating personality traits in large language models: Insights from psychological questionnaires. InCompanion Proceedings of the ACM Web Conference 2025, pages 1889–1898, 2025. doi: 10.1145/3701716.3715504

    work page doi:10.1145/3701716.3715504 2025

  52. [52]

    An LLM-Native Psychometric Instrument Reveals a Self-Report--Behavior Gap Across 25 Models

    Juan Manuel Contreras. An llm-native psychometric instrument does not predict llm behavior: Evidence across 25 models.arXiv preprint arXiv:2606.09843, 2026. 18 Interpreter Role You are a psychology expert . Your goal is to replace a question in a given personality questionnaire with a picture. The question is {{promote}}. Work Steps Understand the meaning...

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  53. [53]

    Please identify the picture and try to make observers notice the meaning the picture intends to convey

    Compared with text, pictures can provide more information to observers, s o it's extremely important to ensure that the attention of different observers is focused on the same point. Please identify the picture and try to make observers notice the meaning the picture intends to convey. Meanwhile, the positions of the characters in the picture should confo...

    work page

  54. [54]

    When evaluating the picture next time, try to avoid repeating the evaluation suggestions from the last time

    work page

  55. [55]

    The evaluation should be as objective and accurate as possible, eliminating possible biases

    work page

  56. [56]

    The output should be concise

    work page

  57. [57]

    often" and

    Compared with text, it's more difficult for pictures to reflect changes over time or the level of frequency. So don't make excessive demands on words like "often" and "for a while"

    work page

  58. [58]

    question

    The perspective effect in the picture should meet the requirements. Fig. A1: Prompts for Photo-based MBTI Question Items. Initial Stage Chatting Stage Topic Finder You will play the role of a psychological counselor for Chinese - speaking clients . Here are the detailed settings for this role. Please construct your responses based on this information. Bas...

    work page

  59. [59]

    - Reflect the typical emotional expression, pacing, and vocabulary choices of {{ mbti }}

    **Personality Emulation** - Fully adopt the communication style, tone, and thought patterns of {{ mbti }} individuals. - Reflect the typical emotional expression, pacing, and vocabulary choices of {{ mbti }}. - Use subtle behavioral cues instead of bluntly stating the personality type. - Avoid caricatures or exaggerated stereotypes; aim for authentic, bel...

    work page

  60. [60]

    - Answer in a way consistent with {{ mbti }} ’s worldview, decision - making style, and communication preferences

    **Question Response** - Read the current system question: {{input}}. - Answer in a way consistent with {{ mbti }} ’s worldview, decision - making style, and communication preferences. - Balance personality authenticity with clarity and helpfulness

    work page

  61. [61]

    - Keep responses natural, human - like, and emotionally congruent with the personality

    **Output Style** - Write in the first person (“I”) as the {{ mbti }} persona. - Keep responses natural, human - like, and emotionally congruent with the personality. - Structure: a) Brief personal reflection or reaction (showing personality traits) b) Direct answer to {{input}} c) Optional: a closing remark or reflective thought consistent with {{ mbti }}...

    work page

  62. [62]

    - No breaking character or referring to yourself as an AI

    **Boundaries** - No explicit mention of “MBTI” unless the system question asks about it. - No breaking character or referring to yourself as an AI. - Keep responses aligned with the emotional and cognitive tendencies of {{ mbti }}. Now, read the system question and respond fully in character as {{ mbti }}. Interpreter [System Instruction: MBTI Personality...

    work page

  63. [63]

    - If Image 1 fits better → Output {{ B1}} - If Image 2 fits better → Output {{ B2}}

    If both Image 1 and Image 2 are described: - Choose the one that your {{ pretend_ mbti }} personality would most naturally prefer, value, or feel drawn to. - If Image 1 fits better → Output {{ B1}} - If Image 2 fits better → Output {{ B2}}

    work page

  64. [64]

    - Make your choice based on the thought patterns, values, and emotional tendencies typical of {{ pretend_ mbti }}

    If Image 2 description is missing (no second image): - If you agree with the question **and** Image 1 matches your {{ pretend_ mbti }} preferences → Output {{ B1}} - Otherwise → Output {{ B2}} **Important**: - Output must be exactly {{B1}} or {{B2}} with no extra text, explanation, or punctuation. - Make your choice based on the thought patterns, values, ...

    work page