Search papers, labs, and topics across Lattice.
EvoSpark is introduced as a framework to address the challenges of stochasticity, social memory stacking, and narrative-spatial dissonance in long-horizon narrative evolution within LLM-based multi-agent systems. It uses a Stratified Narrative Memory with a Role Socio-Evolutionary Base to resolve conflicting relational states and a Generative Mise-en-Sc\`ene mechanism to enforce Role-Location-Plot alignment. Experiments show EvoSpark outperforms baselines in generating coherent narratives by integrating an Emergent Character Grounding Protocol to transform stochastic events into persistent characters.
LLMs can now sustain logically coherent, long-horizon narratives in multi-agent systems, overcoming issues like social memory stacking and narrative-spatial dissonance that plague existing approaches.
Realizing endogenous narrative evolution in LLM-based multi-agent systems is hindered by the inherent stochasticity of generative emergence. In particular, long-horizon simulations suffer from social memory stacking, where conflicting relational states accumulate without resolution, and narrative-spatial dissonance, where spatial logic detaches from the evolving plot. To bridge this gap, we propose EvoSpark, a framework specifically designed to sustain logically coherent long-horizon narratives within Endogenous Interactive Agent Societies. To ensure consistency, the Stratified Narrative Memory employs a Role Socio-Evolutionary Base as living cognition, dynamically metabolizing experiences to resolve historical conflicts. Complementarily, Generative Mise-en-Sc\`ene mechanism enforces Role-Location-Plot alignment, synchronizing character presence with the narrative flow. Underpinning these is the Unified Narrative Operation Engine, which integrates an Emergent Character Grounding Protocol to transform stochastic sparking into persistent characters. This engine establishes a substrate that expands a minimal premise into an open-ended, evolving story world. Experiments demonstrate that EvoSpark significantly outperforms baselines across diverse paradigms, enabling the sustained generation of expressive and coherent narrative experiences.
