When Extremists Enter LLM Societies: Investigating Extreme Behavior in Multi-Agent Systems

Key points

• We study this regularity in controlled LLM-based multi-agent systems (LLM-MAS).1
• We first analyzed three properties of extremists in sociology and model the LLM extremist based on LLM-native perspective.1
• We introduce EXAM, an EXtremist Agent Modeling framework that capable of configuring extremist agents for specific tasks and precisely controlling their population proportion, spatial distribution, interaction protocols, and measurement interfaces.1
• Across four scenarios: congressional election, financial crisis, epidemic public opinion, and iterated prisoner's dilemma, we observe that small persistent tail groups produce measurable, backbone- and scenario-dependent collective shifts.1

Sources (1)

• [1]When Extremists Enter LLM Societies: Investigating Extreme Behavior in Multi-Agent Systems
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 14, 09:27 AM

  Large language model (LLM) agents increasingly serve as computational actors in multi-agent social simulations.

  We study this regularity in controlled LLM-based multi-agent systems (LLM-MAS).

Extractive summary: sentences quoted from the sources.