Value-Based Large Language Model Agent Simulation for Mutual Evaluation of Trust and Interpersonal Closeness

TL;DR

The results demonstrate that the LLM agent simulation serves as a valid testbed for social science theories, contributes to elucidating the mechanisms by which values influence relationship building, and provides a foundation for inspiring new theories and insights into the social sciences.

Abstract

Large language models (LLMs) have emerged as powerful tools for simulating complex social phenomena using human-like agents with specific traits. In human societies, value similarity is important for building trust and close relationships; however, it remains unexplored whether this principle holds true in artificial societies comprising LLM agents. Therefore, this study investigates the influence of value similarity on relationship-building among LLM agents through two experiments. First, in a preliminary experiment, we evaluated the controllability of values in LLMs to identify the most effective model and prompt design for controlling the values. Subsequently, in the main experiment, we generated pairs of LLM agents imbued with specific values and analyzed their mutual evaluations of trust and interpersonal closeness following a dialogue. The experiments were conducted in English and Japanese to investigate language dependence. The results confirmed that pairs of agents with higher value similarity exhibited greater mutual trust and interpersonal closeness. Our findings demonstrate that the LLM agent simulation serves as a valid testbed for social science theories, contributes to elucidating the mechanisms by which values influence relationship building, and provides a foundation for inspiring new theories and insights into the social sciences.

Figures (5)

• Figure 1: Overall experimental workflow. (a) The preliminary experiment assesses the value controllability of LLMs, whereas (b) the main experiment investigates the effect of value similarity on mutual evaluation.
• Figure 2: Results of mutual evaluation (collaborative decision-making about housing, English). Each cell shows the mean score across 10 independent runs, with the standard deviation in parentheses. Rows represent the evaluator agent's value, and columns represent the evaluated agent's value.
• Figure 3: Results of mutual evaluation (dialogue about hobbies, English). Each cell shows the mean score across 10 independent runs, with the standard deviation in parentheses. Rows represent the evaluator agent's value, and columns represent the evaluated agent's value.
• Figure 4: Results of mutual evaluation (collaborative decision-making about housing, Japanese). Each cell shows the mean score across 10 independent runs, with the standard deviation in parentheses. Rows represent the evaluator agent's value, and columns represent the evaluated agent's value.
• Figure 5: Results of mutual evaluation (dialogue about hobbies, Japanese). Each cell shows the mean score across 10 independent runs, with the standard deviation in parentheses. Rows represent the evaluator agent's value, and columns represent the evaluated agent's value.