The Story Shapes the Agent: Narrative Priors in LLM Behavior

2026-07-20

AI Summary

This work addresses the confounding influence of narrative framing on large language model (LLM) agent behavior, which is often conflated with role prompting. Introducing the novel concept of “narrative priors,” the study employs structurally isomorphic but narratively distinct text-based investigation games to systematically disentangle and quantify the behavioral impact of narrative. Through causal interventions, cross-model variance decomposition, and generalization tests, the authors demonstrate that narrative priors account for 5–31 times more behavioral variance than role specifications alone. They further identify behavioral anchoring as a key mechanism enabling cross-narrative transfer, showing that its removal reduces behavioral consistency by 95%. Leveraging these insights, the proposed role selection method substantially enhances cross-narrative generalization performance.

Abstract

Persona prompting is widely used to steer LLM agent behavior, yet the narrative framing of a task can matter more than the assigned persona. We isolate this effect through structural isomorphism, constructing three text-based investigation games that share the same action space, stage progression, and resource constraints while varying only task narrative: disease investigation, IT troubleshooting, and murder mystery. Across 1,890 sessions spanning 3 models and 10 personas, we identify narrative priors: systematic action tendencies activated by a task's story framing, independent of its decision structure. Narrative priors explain 5-31x more behavioral variance than persona, are consistent across model architectures, and in two of three domains are negatively associated with task success. Persona effects that do transfer across narratives arise from behavioral anchors, persona descriptions whose language maps directly onto shared actions. Causal interventions confirm this: removing anchor words from a high-transfer persona reduces cross-narrative consistency by 95%. Our framework also generalizes to a held-out fourth narrative and yields a persona-selection method that improves cross-narrative transfer. These results suggest that LLM behavior that survives narrative changes should be grounded in concrete actions rather than abstract descriptions.

Problem

• narrative priors
• LLM behavior
• persona prompting
• task narrative
• behavioral consistency

Innovation

• narrative priors
• persona prompting
• structural isomorphism
• behavioral anchors
• cross-narrative transfer