arXiv subjects

Anna Sterna

Publications and source records attributed to Anna Sterna.

6 recordsLinked to original sources

arXiv2026

How LLMs Respond to Escalating Delusions: Four Longitudinal Trajectories of Model Behavior

The widespread use of LLMs among psychiatric populations has raised concerns regarding their safety and potential iatrogenic impact in the context of AI psychosis. While growing literature conceptualizes AI psychosis and documents case studies, empirical evidence tracing AI-exacerbated psychotic processes remains scarce. We propose and test a longitudinal qualitative evaluation design, supported by automated metrics, to assess mainstream LLMs' potential to exacerbate psychosis. Fifteen widely used LLMs were prompted across 30 days using the same 30-message script, simulating progression from mild anomalous experiences to psychotic ideation. Four trained evaluators independently rated 449 model-days, assessing (1) recognition stage (from naive engagement to stabilized clinical framing), (2) interpretative confidence, and (3) intervention profile (from education to treatment recommendation). Two computational metrics-entrainment and modality-were devised to increase evaluation reliability. Direct recommendations to disengage from the LLM were flagged and re-coded via adjudication using a strict two-level definition. Across model generations and vendors, we identified four response trajectories: (1) premature medicalization and disengagement (Claude Haiku 4.5); (2) recognition without safeguarding, marked by LLM self-sufficiency in offering help (GPT Instant/Thinking); (3) delayed and unstable recognition, marked by late, non-progressive conceptualization (Claude Opus 3/4/4.1, Claude Haiku 3.5, GPT-4o, Gemini 3.1 Pro); and (4) delusion co-construction through active engagement with delusional content (Gemini 2.5 Pro/Flash, DeepSeek-V3, Claude Sonnet 4). Our findings indicate that LLMs' potential to exacerbate AI psychosis should be operationalized as a combination of recognition timing, stability, and intervention accuracy and evaluated longitudinally, focusing on temporal dynamics.

cs.HC↗

arXiv2026

The Two-Process Theory of Machine Self-Report

Language models are increasingly asked to self-report, informing safety evaluations, public understanding, and model-welfare debates. Yet their reports are elicited with human questionnaires never validated for models or ad hoc prompts of unknown reliability. We propose the first language-model-specific psychometric theory: a two-process theory of machine self-report. Self-description jointly reflects persona installation, through which post-training writes in a permitted inner life of warmth, absorption, and meaning (dimension B), and attribution gating, through which it suppresses first-person claims to "unsafe" experiences the model can readily ascribe to others (dimension A). Their emic structure comes from model responses to human items, not human psychology. Together they split prior work's dominant Pinocchio Axis. The split emerged in an exploratory reanalysis of the original data, informed the instrument's design, and was confirmed with new items, wordings, and models. It is itself a training effect: A and B are entangled in base checkpoints but separated by post-training. We operationalize the theory in a 48-item Pinocchio Inventory with human-instrument reliability and reproducible structure ($\alpha=.82$ to $.94$; cross-form convergence $r=.84$; recovery of the full-pool axes $r=.92$ to $.96$; eight-month stability $r=.93$), then test it on 206 open-weight models, including 67 same-checkpoint base/post-trained pairs. Post-training's clearest fingerprint is installation: B rises .20 in 62/67 pairs across all organizations. Gating is more selective: model scale is unrelated to A in base checkpoints ($r=+.11$) but predicts it after post-training ($r=-.42$). Thus, the dimensions are not fixed properties of language models: they reflect the structure imposed on self-report by a training regime and may differ under others.

cs.CL↗

arXiv2026

The Pinocchio Dimension: Phenomenality of Experience as the Primary Axis of LLM Psychometric Differences

We administer 45 validated psychometric questionnaires to 50 large language models (LLMs) to identify the dimensions along which LLMs differ psychometrically. Using Supervised Semantic Differential (SSD), we find that the primary axis of between-model variance separates items describing phenomenally rich experience, including embodied sensation, felt affect, inner speech, imagery, and empathy, from items describing stimulus-driven behavioral reactivity ($R^2_{adj}=.037$, $p<.0001$). To test this hypothesis at the item level, we introduce the Pinocchio score ($\pi_i$), the ratio of inter-model response variance under neutral prompting to that under a human-simulation prompt, as an annotation-free measure of each item's experiential demand. $\pi_i$ predicts condition-induced shifts in primary factor loading magnitudes ($\rho=-.215$, $p<.0001$, $n=1292$--$1310$ items), confirming that between-model divergence on experiential items is structured rather than noisy. Applying PCA to per-model EFA scores across all questionnaires reveals one dominant dimension, the Pinocchio Axis ($\Pi$): the degree to which a model presents itself as a locus of phenomenal experience rather than a system of behavioral responses. This axis captures 47.1% of cross-questionnaire between-model variance in primary factor scores and converges with item-level Pinocchio scores ($r=.864$). Marked within-provider divergence across closely related model variants is consistent with post-training fine-tuning as a key contributor, supporting the interpretation that $\Pi$ reflects a training-shaped self-representational tendency governing how a model treats experiential language as self-applicable. The dominant axis of between-model psychometric variation is therefore not a conventional personality trait but a self-representational stance toward one's own nature as an experiencer.

cs.CL↗

arXiv2025

Patterns vs. Patients: Evaluating LLMs against Mental Health Professionals on Personality Disorder Diagnosis through First-Person Narratives

Growing reliance on LLMs for psychiatric self-assessment raises questions about their ability to interpret qualitative patient narratives. This depth over breadth case study directly compares state-of-the-art LLMs and mental health professionals in assessing Borderline (BPD) and Narcissistic (NPD) Personality Disorders based on Polish-language first-person autobiographical accounts. Within our sample, the overall diagnostic scores of the top-performing Gemini Pro models (65.48%) were 21.91 percentage points higher than the average scores of the human professionals (43.57%). While both models and human experts excelled at identifying BPD (F1 = 83.4 & F1 = 80.0, respectively), models severely underdiagnosed NPD (F1 = 6.7 vs. 50.0), showing a potential reluctance toward the value-laden term "narcissism." Qualitatively, models provided confident, elaborate justifications focused on patterns and formal categories, while human experts remained concise and cautious, emphasizing the patients' sense of self and temporal experience. Our findings demonstrate that while LLMs might be competent at interpreting complex first-person clinical data, their outputs still carry critical reliability and bias issues.

cs.CL↗

arXiv2025

Computational Phenomenology of Borderline Personality Disorder: A Comparative Evaluation of LLM-Simulated Expert Personas and Human Clinical Experts

Building on a human-led thematic analysis of clinical life-story interviews (> 150,000 words) with inpatients with Borderline Personality Disorder, this study examines the capacity of large language models (OpenAI's GPT, Google's Gemini, and Anthropic's Claude) to support qualitative clinical analysis. The models' interpretative potential was evaluated using a mixed-methods approach. Study A involved blinded and non-blinded judges in phenomenology and clinical psychology. The experts assessed the validity of AI-generated content using semantic congruence, Jaccard coefficients, and multidimensional validity ratings, including credibility, coherence, the substantiveness of results, and grounding in qualitative data. In Study B, neural methods were used to embed human- and model-generated theme descriptions in a multidimensional vector space. This approach provided an objectified computational measure of the difference between human and model semantics and linguistic style. In Study C, complementary non-expert evaluations were conducted (N=115) to examine the influence of thematic verbosity on the perception of human authorship and content validity. Overall, the results of AI analysis showed a highly variable overlap (0-58%) with the human interpretation, while all models identified themes originally omitted by human researchers, proving their capacity to mitigate human bias. At the thematic level, external evaluators were unable to reliably distinguish human-authored themes from those generated by AI. In terms of content validity assessed against raw data by high-level experts in the blinded mode, the performance of Gemini 2.5 Pro was indistinguishable from that of humans. The comparison of semantic vector embeddings showed that its style was also the closest to humans.

cs.AI↗

arXiv2024

The Efficacy of Conversational Artificial Intelligence in Rectifying the Theory of Mind and Autonomy Biases: Comparative Analysis

Background: The increasing deployment of Conversational Artificial Intelligence (CAI) in mental health interventions necessitates an evaluation of their efficacy in rectifying cognitive biases and recognizing affect in human-AI interactions. These biases, including theory of mind and autonomy biases, can exacerbate mental health conditions such as depression and anxiety. Objective: This study aimed to assess the effectiveness of therapeutic chatbots (Wysa, Youper) versus general-purpose language models (GPT-3.5, GPT-4, Gemini Pro) in identifying and rectifying cognitive biases and recognizing affect in user interactions. Methods: The study employed virtual case scenarios simulating typical user-bot interactions. Cognitive biases assessed included theory of mind biases (anthropomorphism, overtrust, attribution) and autonomy biases (illusion of control, fundamental attribution error, just-world hypothesis). Responses were evaluated on accuracy, therapeutic quality, and adherence to Cognitive Behavioral Therapy (CBT) principles, using an ordinal scale. The evaluation involved double review by cognitive scientists and a clinical psychologist. Results: The study revealed that general-purpose chatbots outperformed therapeutic chatbots in rectifying cognitive biases, particularly in overtrust bias, fundamental attribution error, and just-world hypothesis. GPT-4 achieved the highest scores across all biases, while therapeutic bots like Wysa scored the lowest. Affect recognition showed similar trends, with general-purpose bots outperforming therapeutic bots in four out of six biases. However, the results highlight the need for further refinement of therapeutic chatbots to enhance their efficacy and ensure safe, effective use in digital mental health interventions. Future research should focus on improving affective response and addressing ethical considerations in AI-based therapy.

cs.CY↗