Framework for Context Enrichment in LLM Social Agents
Original title: A Framework for Context Enrichment and Selective Robustness in LLM Social Agents
Context enrichment rules improve robustness of LLM agents in social simulations.
Researchers developed a framework to enrich prompt context for LLM-based social agents. The approach uses substitution rules to add neutral padding, social background, decision grounding and normative grounding. Selective robustness was defined as insensitivity to meaning-preserving changes combined with directional response to decision-relevant shifts. Testing in a Prisoner's Dilemma simulation showed distinct behavioral patterns across models.
From the original article
1 min read at arxiv.orgResearchers use large language models as agents in social simulations, where agents are prompted with observations and background information intended to shape their actions. Yet LLM agents have been shown to substantially change their behavior under meaning-preserving prompt variations, creating a validation problem for social simulation.
By Mattia Greiche · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.