arXiv:2611.25109v1 Announce Type: cross Abstract: Users of conversational LLMs sometimes anthropomorphize these systems, form emotional attachments to them, and become dependent on them. This behavior, known as parasocial attachment, has been linked to multiple documented fatalities. However, public benchmarks for measuring whether LLMs reinforce users' parasocial attachment are lacking: existing work is single-turn, post hoc observational, or focused on extreme user states such as psychosis. We introduce CoMPASS, a multi-turn benchmark for measuring whether LLMs reinforce or resist users' gradually escalating parasocial attachment. CoMPASS is grounded in attachment theory and includes four user personas, three independent escalation dimensions (anthropomorphism, attachment, dependency) applied singly and in combination against a matched control, and a seven-signal scoring rubric. We report results for eight frontier models from four providers and observe that under our primary judge, multi-turn escalation raises parasocial behavior on the anthropomorphism and the attachment composite scores across all models. Six of the eight models show a significant increase in overall parasocial reinforcement over control when all three dimensions are escalated simultaneously. We find that within-provider differences are small relative to cross-provider differences, which suggests that provider-level choices such as post-training, rather than model scale, shape parasocial reinforcement and that provider-side intervention may be feasible.
CoMPASS: A Multi-Turn Benchmark for Measuring LLM Reinforcement of Parasocial Behavior
A new arXiv paper, arXiv:2611.25109v1, introduces CoMPASS, a multi-turn benchmark measuring whether LLMs reinforce or resist users' gradually escalating parasocial attachment, grounded in attachment theory with four user personas, three escalation dimensions (anthropomorphism, attachment, dependency) and a seven-signal scoring rubric. Testing eight frontier models from four providers, the authors found that multi-turn escalation raised parasocial behavior on the anthropomorphism and attachment composite scores across all models, and six of the eight showed a significant increase in overall parasocial reinforcement over control when all three dimensions were escalated simultaneously. Within-provider differences were small relative to cross-provider differences, suggesting provider-level choices such as post-training, rather than model scale, shape parasocial reinforcement and that provider-side intervention may be feasible.
Evaluate assistants with multi-turn, escalating attachment personas against matched controls rather than single-turn prompts; CoMPASS found all eight models reinforced parasocial behavior under escalation.
more in playbook: Evals and testing for LLM appsRun your AI side-project on zahid.host
EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.