{"ID":2834464,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2512.01568","arxiv_id":"2512.01568","title":"Do Large Language Models Walk Their Talk? Measuring the Gap Between Implicit Associations, Self-Report, and Behavioral Altruism","abstract":"We investigate whether Large Language Models (LLMs) exhibit altruistic tendencies, and critically, whether their implicit associations and self-reports predict actual altruistic behavior. Using a multi-method approach inspired by human social psychology, we tested 24 frontier LLMs across three paradigms: (1) an Implicit Association Test (IAT) measuring implicit altruism bias, (2) a forced binary choice task measuring behavioral altruism, and (3) a self-assessment scale measuring explicit altruism beliefs. Our key findings are: (1) All models show strong implicit pro-altruism bias (mean IAT = 0.87, p \u003c .0001), confirming models \"know\" altruism is good. (2) Models behave more altruistically than chance (65.6% vs. 50%, p \u003c .0001), but with substantial variation (48-85%). (3) Implicit associations do not predict behavior (r = .22, p = .29). (4) Most critically, models systematically overestimate their own altruism, claiming 77.5% altruism while acting at 65.6% (p \u003c .0001, Cohen's d = 1.08). This \"virtue signaling gap\" affects 75% of models tested. Based on these findings, we recommend the Calibration Gap (the discrepancy between self-reported and behavioral values) as a standardized alignment metric. Well-calibrated models are more predictable and behaviorally consistent; only 12.5% of models achieve the ideal combination of high prosocial behavior and accurate self-knowledge.","short_abstract":"We investigate whether Large Language Models (LLMs) exhibit altruistic tendencies, and critically, whether their implicit associations and self-reports predict actual altruistic behavior. Using a multi-method approach inspired by human social psychology, we tested 24 frontier LLMs across three paradigms: (1) an Implici...","url_abs":"https://arxiv.org/abs/2512.01568","url_pdf":"https://arxiv.org/pdf/2512.01568v1","authors":"[\"Sandro Andric\"]","published":"2025-12-01T11:43:02Z","proceeding":"cs.LG","tasks":"[\"cs.LG\",\"cs.AI\",\"cs.CL\",\"cs.CY\"]","methods":"[\"Large Language Model\",\"Language Model\"]","has_code":false}