Patients With Personality: Realistic Patient Simulation through Controlled Diversity and Selective Disclosure

A new patient simulation framework called PatientsWithPersonality (PWP) generates realistic virtual patient responses by explicitly parameterizing personality traits, according to a preprint submitted in 2026. The system is designed to test clinical applications of large language models at scale without relying on costly human studies [1]. The framework is grounded in HEXACO, a six-dimensional personality space used to quantify and parameterize human behavioral traits [1]. This approach enables fine-grained control over conversational style, cooperativeness, and information disclosure within a unified framework [1]. In a clinician evaluation, PWP was judged nearly as realistic as recorded human actors and clearly ahead of prior simulators [1]. It was also flagged as "too informative" far less often than earlier systems, addressing a known problem where simulators overshare information unprompted [1]. Conditioning on HEXACO axes yields personas whose configured traits are recoverable by both clinicians and an autorater, and these personas span a substantially wider behavioral footprint than the closest baseline [1]. The researchers argue that the framework paves the way for more accurate and informative benchmarking of large language models, which are machine learning models designed for natural language processing tasks such as language generation [8]. The preprint was posted on arXiv, an open-access repository of electronic preprints that is moderated but not peer-reviewed [6]. As of November 2024, the repository was receiving about 24,000 articles per month and had surpassed two million total articles by the end of 2021 [6]. The paper appears under the Human-Computer Interaction category and was submitted on 13 May 2026 [1]. arXiv also hosts experimental community projects through its arXivLabs framework, which allows collaborators to develop and share new features directly on the site [4]. These projects, which include tools like the Bibliographic Explorer and CORE Recommender, must adhere to arXiv's values of openness, community, excellence, and user data privacy [4][5]. The Labs framework was formalized in 2020 to enable innovative collaborations with individuals and organizations [4].