Enforcing Narrative Reliability and Epistemic Pacing in LLM-Driven Detective Games via Structured Knowledge Trees
A Structured Knowledge Tree architecture coupled with a tri-agent LLM pipeline for controlling dialogue in an open-ended interrogation game and demonstrates that the structured architecture reduces critical hallucinations by 64.78% and entirely prevents premature narrative disclosure.
Parsa Rahmati, Richard Zhao
· 0 citations