Saved in:
Bibliographic Details
Main Authors: Peng, Tai-Quan, Tian, Yuan, Liang, Songsong, Deng, Dazhen, Wu, Yingcai
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2603.20911
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911534615101440
author Peng, Tai-Quan
Tian, Yuan
Liang, Songsong
Deng, Dazhen
Wu, Yingcai
author_facet Peng, Tai-Quan
Tian, Yuan
Liang, Songsong
Deng, Dazhen
Wu, Yingcai
contents Large language models make agent-based simulation more behaviorally expressive, but they also sharpen a basic methodological tension: fluent, human-like output is not, by itself, evidence for theory. We evaluate what an LLM-driven simulation can credibly support using information engagement on social media as a test case. In a Weibo-like environment, we manipulate information load and descriptive norms, while allowing popularity cues (cumulative likes and Sina Weibo-style cumulative reshares) to evolve endogenously. We then ask whether simulated behavior changes in theoretically interpretable ways under these controlled variations, rather than merely producing plausible-looking traces. Engagement responds systematically to information load and descriptive norms, and sensitivity to popularity cues varies across contexts, indicating conditionality rather than rigid prompt compliance. We discuss methodological implications for simulation-based communication research, including multi-condition stress tests, explicit no-norm baselines because default prompts are not blank controls, and design choices that preserve endogenous feedback loops when studying bandwagon dynamics.
format Preprint
id arxiv_https___arxiv_org_abs_2603_20911
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Do LLM-Driven Agents Exhibit Engagement Mechanisms? Controlled Tests of Information Load, Descriptive Norms, and Popularity Cues
Peng, Tai-Quan
Tian, Yuan
Liang, Songsong
Deng, Dazhen
Wu, Yingcai
Artificial Intelligence
Computers and Society
Large language models make agent-based simulation more behaviorally expressive, but they also sharpen a basic methodological tension: fluent, human-like output is not, by itself, evidence for theory. We evaluate what an LLM-driven simulation can credibly support using information engagement on social media as a test case. In a Weibo-like environment, we manipulate information load and descriptive norms, while allowing popularity cues (cumulative likes and Sina Weibo-style cumulative reshares) to evolve endogenously. We then ask whether simulated behavior changes in theoretically interpretable ways under these controlled variations, rather than merely producing plausible-looking traces. Engagement responds systematically to information load and descriptive norms, and sensitivity to popularity cues varies across contexts, indicating conditionality rather than rigid prompt compliance. We discuss methodological implications for simulation-based communication research, including multi-condition stress tests, explicit no-norm baselines because default prompts are not blank controls, and design choices that preserve endogenous feedback loops when studying bandwagon dynamics.
title Do LLM-Driven Agents Exhibit Engagement Mechanisms? Controlled Tests of Information Load, Descriptive Norms, and Popularity Cues
topic Artificial Intelligence
Computers and Society
url https://arxiv.org/abs/2603.20911