Constrained Reasoning Chains for Enhancing Theory-of-Mind in Large Language Models

TL;DR

This paper introduces CCoToM, a zero-shot prompting method that enhances large language models' theory-of-mind reasoning by constructing explicit, constrained reasoning chains, improving performance in complex and non-narrative contexts.

Contribution

The paper proposes CCoToM, a novel zero-shot prompting technique that leverages domain knowledge and causal relations to improve ToM reasoning in LLMs, including non-narrative scenarios.

Findings

CCoToM outperforms previous methods across multiple datasets and models.

It effectively handles both narrative and non-narrative contexts.

Extensive experiments validate its superior performance.

Abstract

Theory-of-Mind (ToM) ability possessed by Large Language Models (LLMs) has been shown to be limited. Most existing methods for improving ToM in LLMs adopt zero-shot prompting, and they face challenges including poor performance in complex ToM reasoning tasks and an inability to handle non-narrative contexts. We propose a zero-shot prompting method named Constrained Chain-of-ToM (CCoToM) that leverages domain knowledge and the causal relations between ToM dimensions to address these limitations. Specifically, CCoToM guides LLMs to construct explicit reasoning chains by first prompting LLMs to infer related ToM dimensions (e.g., belief). Afterward, CCoToM prompts LLMs to infer the queried ToM dimension based on the generated related ToM dimensions and corresponding causal relations. Additionally, CCoToM adaptively imposes constraints on prompts to introduce inductive biases and improve consistency between ToM dimensions. Besides narratives, CCoToM can also handle non-narrative contexts like conversations. Extensive experiments show that CCoToM consistently outperforms previous state-of-the-art methods by large margins across all LLMs and datasets used. We also conduct in-depth analyses to gain deeper insights into CCoToM. We have made our code publicly available.