SocialEval: Evaluating Social Intelligence of Large Language Models

2025-06-01

AI Summary

This study addresses the lack of systematic evaluation of social intelligence (SI) in large language models (LLMs). We introduce SocialEval, the first script-driven, bilingual SI benchmark. Methodologically, we design human-authored, narrative “World Tree” scripts and propose a dual-dimensional evaluation paradigm—outcome-oriented (goal achievement) and process-oriented (interpersonal skill deployment)—complemented by behavioral trajectory analysis, neural activation mapping, and representational geometric analysis for fine-grained capability attribution and neuro-mechanistic investigation. Key contributions include: (1) the first empirical demonstration that LLMs significantly underperform humans in SI, exhibiting excessive prosociality that frequently compromises goal attainment; and (2) discovery of functionally specialized, capability-specific representational subspaces in LLMs—paralleling the modular organization of human social cognitive networks.

Technology Category

• Cognitive Modeling & Cognitive Systems: Social Cognition And Interaction
• Machine Learning: Large Multimodal Models (LMMs)
• Humans and AI: Emotional Intelligence

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

LLMs exhibit promising Social Intelligence (SI) in modeling human behavior, raising the need to evaluate LLMs' SI and their discrepancy with humans. SI equips humans with interpersonal abilities to behave wisely in navigating social interactions to achieve social goals. This presents an operational evaluation paradigm: outcome-oriented goal achievement evaluation and process-oriented interpersonal ability evaluation, which existing work fails to address. To this end, we propose SocialEval, a script-based bilingual SI benchmark, integrating outcome- and process-oriented evaluation by manually crafting narrative scripts. Each script is structured as a world tree that contains plot lines driven by interpersonal ability, providing a comprehensive view of how LLMs navigate social interactions. Experiments show that LLMs fall behind humans on both SI evaluations, exhibit prosociality, and prefer more positive social behaviors, even if they lead to goal failure. Analysis of LLMs' formed representation space and neuronal activations reveals that LLMs have developed ability-specific functional partitions akin to the human brain.

Problem

• Evaluating Social Intelligence (SI) in Large Language Models (LLMs)
• Assessing discrepancy between LLMs and humans in social interactions
• Developing outcome- and process-oriented SI evaluation benchmarks

Innovation

• Script-based bilingual SI benchmark
• Outcome- and process-oriented evaluation integration
• Ability-specific functional partitions analysis