arXiv:2605.30058v1 Announce Type: new 
Abstract: While LLM agents have demonstrated remarkable task-oriented abilities such as planning, reasoning, and action, few works have treated them as complete human personalities where emotional dimensions hold equal importance. In this paper, we introduce a novel benchmark to systematically assess whether LLM agents can simulate coherent, human-like psychology. Specifically, our benchmark constructs 11 diverse human characters grounded in orthogonal Big Five personality traits, with each profile deeply integrated with 1,000 structured autobiographical-style episodic memories distributed across theory-grounded developmental life stages. To rigorously evaluate the psychological manifestations of LLMs, we designed a curated suite of 64 decision-making scenarios, guided by the DIAMONDS taxonomy, a psychological framework that characterizes situations along eight dimensions: Duty, Intellect, Adversity, Mating, pOsitivity, Negativity, Deception, and Sociality. By subjecting agents to varying scenarios, the benchmark evaluates whether they can consolidate their innate personality traits and autobiographical memories to make behavioral decisions that are consistent with their specific psychological profiles. After systematic human validation and filtering, we obtained a benchmark consisting of 673 multiple-choice questions (MCQs). We believe this benchmark provides a principled and scalable testbed for studying human-like emotions, personality consistency, and value-consistent behavioural decision-making in LLM-based agents.

تم تقديم معيار جديد يسمى HEART-Bench لتقييم ما إذا كانت نماذج اللغة الكبيرة (LLM) يمكن أن تحاكي نفسية بشرية، مع التركيز على الأبعاد العاطفية إلى جانب القدرات الموجهة نحو المهام. يتضمن هذا المعيار 11 شخصية بشرية متنوعة تعتمد على السمات الشخصية الخمسة الكبرى، مع 1000 ذاكرة ذات طابع سيرة ذاتية منظمة و64 سيناريوًات لصنع القرار مستندة إلى تصنيف DIAMONDS.

Se ha introducido un nuevo benchmark llamado HEART-Bench para evaluar si los agentes de modelos de lenguaje (LLM) pueden simular una psicología humana, centrándose en las dimensiones emocionales junto con las habilidades orientadas a tareas. Este benchmark presenta 11 personajes humanos diversos basados en los cinco grandes rasgos de personalidad, cada uno con 1,000 recuerdos autobiográficos estructurados y 64 escenarios de toma de decisiones guiados por la taxonomía DIAMONDS.

Un nouveau benchmark appelé HEART-Bench a été introduit pour évaluer si les agents de modèles de langage (LLM) peuvent simuler une psychologie humaine, en mettant l'accent sur les dimensions émotionnelles aux côtés des capacités orientées vers les tâches. Ce benchmark présente 11 personnages humains diversifiés basés sur les Big Five traits de personnalité, chacun avec 1 000 souvenirs autobiographiques structurés et 64 scénarios de prise de décision guidés par la taxonomie DIAMONDS.

A new benchmark called HEART-Bench has been introduced to evaluate whether large language model (LLM) agents can simulate human-like psychology, focusing on emotional dimensions alongside task-oriented abilities. This benchmark features 11 diverse human characters based on the Big Five personality traits, each with 1,000 structured autobiographical memories and 64 decision-making scenarios guided by the DIAMONDS taxonomy.