State-Inference-Based Prompting for Natural Language Trading with Game NPCs
This addresses the problem of unreliable NPC interactions for game developers and players, though it appears incremental as it builds on existing prompting methods for a specific domain.
The paper tackled the problem of rule violations in LLM-based game trading systems by developing State-Inference-Based Prompting (SIBP), which achieved >97% state compliance, >95% referencing accuracy, and 99.7% calculation precision in evaluations across 100 trading dialogues.
Large Language Models enable dynamic game interactions but struggle with rule-governed trading systems. Current implementations suffer from rule violations, such as item hallucinations and calculation errors, that erode player trust. Here, State-Inference-Based Prompting (SIBP) enables reliable trading through autonomous dialogue state inference and context-specific rule adherence. The approach decomposes trading into six states within a unified prompt framework, implementing context-aware item referencing and placeholder-based price calculations. Evaluation across 100 trading dialogues demonstrates >97% state compliance, >95% referencing accuracy, and 99.7% calculation precision. SIBP maintains computational efficiency while outperforming baseline approaches, establishing a practical foundation for trustworthy NPC interactions in commercial games.