Deterministic agent system improves historical simulation with large language models
Deterministic Narrative Gravity: Bridging Theory of Mind and Historical Simulation through ABM-LLM Architectures
Multiagent SystemsArtificial IntelligenceComputation and Language
Summary
One challenge in simulating historical social scenarios using large language AI models is that their random responses make it hard to repeat the exact same story or result. The authors created a system that uses smaller smart agents to control the AI models precisely and make the simulations repeatable and consistent. Their system also helps models with at least 12 billion parameters to produce richer, more varied storylines that avoid repetitive or aggressive patterns. This approach offers an open and verifiable way to simulate historical narratives with AI and can be run locally.
What this means in practice
- •For virtual archaeology developers: Build reproducible and transparent historical scenario simulations using deterministic agent-Language Model architectures.
- •For interactive narrative designers: Create dynamic story scenarios that sustain open-ended narratives without repetitive or closed dialogue patterns by integrating agent scaffolding.
Authors
Alicia Colmenero-Fernández, Lidia Ortega Alvarado
Abstract
Historical social simulation faces a fundamental reproducibility challenge: the stochastic nature of large language models (LLMs) introduces a digital butterfly effect that prevents exact replication across multi-turn runs. Furthermore, the inability to control low-level server parameters in commercial LLM hosting environments severely limits reproducibility. To overcome this opacity, we present a deterministic multi-agent framework coupling LLMs to an Agent-Based Modelling (ABM) pipeline grounded in Theory of Mind (ToM) and Cognitive Archaeology, and validate exact artifact identity within local, pinned execution environments using SHA-256 hashing. Run locally at temperature zero with clean per-call model reloads, the system injects differentiated cognitive states per turn through dissonance-driven behavioural modes and gated internal voices. Evaluating seven model configurations spanning 4B to 31B parameters against an ablated baseline reveals clear operational boundaries. Small models ($\leq$8B) suffer from context capacity overload, producing repetitive, specular dialogue loops. However, at $\geq$12B parameters, the ABM scaffolding breaks the monotonous response patterns seen in ablation. Crucially, these scaffolded models exhibit spontaneous diegetic expansion: they coherently expand scenario lore through tactically motivated, anachronism-free details rather than escalating into aggression. By resisting the structural pull toward premature scene closure, scaffolded sessions sustain an open terminal Narrative Vector ($V_N = \mathrm{Open}$ in 12/21 overall; 9/9 in official $\geq$12B configurations) whereas unscaffolded ablation never does (0/21). This framework provides a transparent, auditable benchmark architecture for virtual archaeology, with the full pipeline released as open-source software to enable independent replication.