What-If Analysis of Large Language Models: Explore the Game World Using Proactive Thinking

2025-09-05

AI Summary

Current large language models (LLMs) excel at passive information processing but lack systematic counterfactual reasoning capabilities—specifically, the ability to proactively anticipate how future states would evolve under hypothetical interventions—limiting their applicability in high-stakes, dynamic domains such as strategic planning and risk assessment. To address this, we propose WiA-LLM, the first framework that tightly integrates structured what-if analysis with a reinforcement learning architecture, endowing LLMs with the capacity to actively model multi-step causal interventions and forecast state evolution. Our method employs environment-feedback-driven dynamic simulation to enable prospective consequence prediction for complex decisions. Evaluated in the *Honor of Kings* gaming environment, WiA-LLM achieves a 74.2% state-change prediction accuracy—double that of baseline methods—and demonstrates marked superiority on challenging long-horizon reasoning tasks. This work represents a paradigm shift from passive response to proactive anticipation in LLM-based decision-making.

Technology Category

• Humans and AI: Human-Aware Planning and Behavior Prediction
• Machine Learning: Large Multimodal Models (LMMs)
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Large language models (LLMs) excel at processing information reactively but lack the ability to systemically explore hypothetical futures. They cannot ask, "what if we take this action? how will it affect the final outcome" and forecast its potential consequences before acting. This critical gap limits their utility in dynamic, high-stakes scenarios like strategic planning, risk assessment, and real-time decision making. To bridge this gap, we propose WiA-LLM, a new paradigm that equips LLMs with proactive thinking capabilities. Our approach integrates What-If Analysis (WIA), a systematic approach for evaluating hypothetical scenarios by changing input variables. By leveraging environmental feedback via reinforcement learning, WiA-LLM moves beyond reactive thinking. It dynamically simulates the outcomes of each potential action, enabling the model to anticipate future states rather than merely react to the present conditions. We validate WiA-LLM in Honor of Kings (HoK), a complex multiplayer game environment characterized by rapid state changes and intricate interactions. The game's real-time state changes require precise multi-step consequence prediction, making it an ideal testbed for our approach. Experimental results demonstrate WiA-LLM achieves a remarkable 74.2% accuracy in forecasting game-state changes (up to two times gain over baselines). The model shows particularly significant gains in high-difficulty scenarios where accurate foresight is critical. To our knowledge, this is the first work to formally explore and integrate what-if analysis capabilities within LLMs. WiA-LLM represents a fundamental advance toward proactive reasoning in LLMs, providing a scalable framework for robust decision-making in dynamic environments with broad implications for strategic applications.

Problem

• LLMs lack proactive thinking for hypothetical futures
• Need systemic exploration of potential actions and outcomes
• Limits utility in dynamic high-stakes decision making scenarios

Innovation

• Integrates What-If Analysis with LLMs for proactive scenario evaluation
• Uses reinforcement learning for dynamic outcome simulation and feedback
• Validated in complex game environments for multi-step prediction