When Does Personality Composition Matter for Multi-Agent LLM Teams?
arXiv:2606.27443v1 Announce Type: new Abstract: Personality prompting shapes how large language models communicate, yet whether these behavioral shifts affect objective task outcomes remains under-explored. Prior work shows that agents prompted with low agreeableness produce adversarial language, while those prompted with high agreeableness become cooperative, but the relationship between communication style and task performance has not been systematically examined across multiple domains. In th
Opening excerpt (first ~120 words) tap to expand
Computer Science > Artificial Intelligence arXiv:2606.27443 (cs) [Submitted on 25 Jun 2026] Title:When Does Personality Composition Matter for Multi-Agent LLM Teams? Authors:Aryan Keluskar, Amrita Bhattacharjee, Huan Liu View a PDF of the paper titled When Does Personality Composition Matter for Multi-Agent LLM Teams?, by Aryan Keluskar and 2 other authors View PDF HTML (experimental) Abstract:Personality prompting shapes how large language models communicate, yet whether these behavioral shifts affect objective task outcomes remains under-explored.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv.org.
Discussion
0 commentsMore from arXiv.org
