ReactiveGWM: Steering NPC in Reactive Game World Models

2026-05-14

AI Summary

This work addresses the limitation of existing game world models, which treat non-player characters (NPCs) as static background elements and fail to capture dynamic interactions triggered by player actions. To overcome this, the authors propose ReactiveGWM, the first game-agnostic framework for learning interactive behavior representations. By injecting lightweight additive biases derived from player actions into a diffusion model and employing cross-attention modules to disentangle player control from NPC policy responses, ReactiveGWM enables zero-shot transfer across games without retraining. The method seamlessly integrates into pre-trained world models of diverse games and demonstrates strong performance in the Street Fighter series, generating robust, prompt-aligned NPC behaviors while preserving fine-grained player controllability.

Abstract

Current game world models simulate environments from a subjective, player-centric perspective. However, by treating the Non-Player Character (NPC) merely as background pixels, these models cannot capture interactions between the player and NPC. In that sense, they act as passive video renderers rather than real simulation engines, lacking the physical understanding needed to model action-induced NPC reactivities. We introduce ReactiveGWM, a reactive game world model that synthesizes dynamic interactions between the player and NPC. Instead of entangling all interaction dynamics, ReactiveGWM explicitly decouples player controls from NPC behaviors. Player actions are injected into the diffusion backbone via a lightweight additive bias, while high-level NPC responses (e.g., Offense, Control, Defense) are grounded through cross-attention modules. Crucially, these modules learn a game-agnostic representation of interactive logic. This enables zero-shot strategy transfer: our learned modules can be plugged directly into off-the-shelf, unannotated world models of different games. This instantly unlocks steerable NPC interactions without any domain-specific retraining. Evaluated on two Street Fighter games, ReactiveGWM maintains fine-grain player controllability while achieving robust, prompt-aligned NPC strategy adherence, paving the way for scalable, strategy-rich interaction with the NPC.

Problem

• game world models
• Non-Player Character
• player-NPC interaction
• reactive behavior
• interactive simulation

Innovation

• Reactive Game World Model
• NPC Interaction
• Decoupled Control
• Cross-Attention
• Zero-Shot Transfer