October 15, 2025Open Access

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

View Full Paper

Ask AI

Bookmark

Pairs of LLM agents with higher value similarity showed increased mutual trust and interpersonal closeness.
The preliminary experiment focused on evaluating how to control values in LLMs for effective relationship-building.
The study included language-dependent analysis, conducted in both English and Japanese for broader insights.
Findings suggest that LLM agent simulation can inform social science theories regarding relationship dynamics.

Authors

YSYuki SakamotoThe University of Osaka TUTakahisa Uchida HIHiroshi IshiguroAdvanced Telecommunications Research Institute International

Discussion

No takes yet. Share an insight, caveat, or question.

Implication

Experiments reveal that LLM agents with similar values demonstrate greater trust and interpersonal closeness, highlighting the importance of value similarity.

Cite This Study

Sakamoto et al. (2025) studied this question.

synapsesocial.com/papers/68ef858cc6a308ba063554a1 https://doi.org/10.48550/arxiv.2507.11979

View Full Paper

Ask AI

Bookmark

Implication

Experiments reveal that LLM agents with similar values demonstrate greater trust and interpersonal closeness, highlighting the importance of value similarity.

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

Key Points

Discussion

Implication

Cite This Study

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

Key Points

Discussion

Implication

Cite This Study