Metraipolis › archive › dc_7185cfa73368

Measuring Model-Specific Alignment Drift in LLM-Based Opinion Dynamics — arXiv:2611.03980

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.03980
License (as archived)
CC0-1.0
Archived
2026-11-04T18:33:45.745Z
SHA-256 of the archived text
e1c517ecb3e3fe92bfe13e415d883b6d64df4ff0623481eb7d29bddfe2ccedf4
Stable URL of this page
https://metraipolis.com/documento/dc_7185cfa73368

How to cite

Measuring Model-Specific Alignment Drift in LLM-Based Opinion Dynamics — arXiv:2611.03980 — arXiv (en), https://arxiv.org/abs/2611.03980 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7185cfa73368 (Metraipolis archive dc_7185cfa73368, sha256 e1c517ecb3e3fe92bfe13e415d883b6d64df4ff0623481eb7d29bddfe2ccedf4)
BibTeX
@misc{dc7185cfa73368,
  title = {Measuring Model-Specific Alignment Drift in LLM-Based Opinion Dynamics — arXiv:2611.03980},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.03980},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7185cfa73368; sha256 e1c517ecb3e3fe92bfe13e415d883b6d64df4ff0623481eb7d29bddfe2ccedf4},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.03980. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Large language model (LLM) agents are increasingly used as synthetic participants in opinion-dynamics experiments, often assuming that an agent's behaviour is fully determined by its prompt. We test this assumption with a framework in which LLM agents discuss a contested policy issue over repeated rounds while their stance is quantified by a fixed psychometric instrument independent of any dynamics model. Across six LLMs, committed personas are intrinsically stable: in each of the five models tested without peer input, every agent keeps its assigned side for 20 rounds. The observed loss of alignment under discussion is therefore interaction-induced, we call it alignment drift, and model-specific: gpt-4.1 keeps opinion clusters apart, claude-sonnet-5 holds them partially, and gpt-5.6-sol and claude-opus-5 abandon them. Under selective attention, one LLM maintains separated opinion clusters under stance-homophilous attention. Qualitative profiles persist at doubled population and horizon and on a second issue. Such heterogeneity makes the LLM choice a first-class validity variable for social simulation.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI) ; Physics and Society (physics.soc-ph).
Pagina del lavoro: https://arxiv.org/abs/2611.03980 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-11-04T18:32:37.255Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).