Think Thrice Before You Speak: Dual knowledge-enhanced Theory-of-Mind Reasoning for Persuasive Agents

2026-05-21

AI Summary

This study addresses the challenge that current large language models struggle to accurately reason about others’ mental states in persuasive dialogues, leading to fragmented representations and unstable inference. To tackle this, the work introduces the belief-desire-intention (BDI) cognitive architecture into persuasion modeling for the first time, constructs ToM-BPD—a large-scale, fine-grained annotated dataset—and proposes TTBYS, a knowledge-enhanced three-stage stepwise reasoning framework. TTBYS integrates explicit and implicit priors to jointly model beliefs, desires, and persuasion strategies across multi-turn dialogues. Experiments based on Qwen3-8B demonstrate that the proposed approach outperforms GPT-5 by 1.20%, 22.80%, and 16.97% in predicting desires, beliefs, and strategies, respectively, substantially improving reasoning consistency and interpretability.

Abstract

Persuasive dialogue requires reasoning about others' latent mental states, a capability known as Theory of Mind (ToM). However, due to reliance on simple prompting strategies and insufficient ToM knowledge, existing LLMs often fail to capture the intrinsic dependencies among mental states, leading to fragmented representations and unstable reasoning. To address these challenges, we introduce the ToM-based Persuasive Dialogue (ToM-PD) task, grounded in the Belief-Desire-Intention (BDI) framework, which explicitly models the sequential dependencies among mental states in multi-turn dialogues. To facilitate research on this task, we construct a large-scale annotated dataset, ToM-based Broad Persuasive Dialogues (ToM-BPD), capturing fine-grained mental states and corresponding persuasive strategies. We further propose Think Thrice Before You Speak (TTBYS), a knowledge-enhanced stepwise reasoning framework that leverages both explicit and implicit prior experiences to improve LLMs' inference of desires, beliefs, and persuasive strategies. Experimental results demonstrate that Qwen3-8B equipped with TTBYS outperforms GPT-5 by 1.20%, 22.80%, and 16.97% in predicting desires, beliefs, and persuasive strategies, respectively. Case studies further show that our approach enhances interpretability and consistency in reasoning.

Problem

• Theory of Mind
• Persuasive Dialogue
• Mental State Reasoning
• Belief-Desire-Intention
• Large Language Models

Innovation

• Theory of Mind
• Persuasive Dialogue
• BDI Framework
• Knowledge-enhanced Reasoning
• Stepwise Inference