Simulating Ideological Deadlock: Persona Stability and Conflict Resolution in LLM-Based Multi-Agent Frameworks

Abstract excerpt

<title>Abstract</title> <p> The integration of Large Language Models (LLMs) into Agent-Based Modeling (ABM) has revolutionized the capacity to simulate complex sociopolitical dynamics. However, state-of-the-art models are fundamentally aligned via safety protocols (e.g., RLHF) to prioritize helpfulness and consensus, severely limiting their utility in simulating intractable, zero-sum human conflict. To address t...

Topics

Open a Topic to create a Post that cites this publication.

• Uncategorized literature

Identifiers and source

Literature Corpus work

f0b00ef9-c9d3-50e9-8c85-b7f628894d31

DOI

10.21203/rs.3.rs-9414531/v1