Metraipolis › archive › dc_0350cc323dd5
From Deceptive Outputs to Deceptive Mechanisms: A Causal Framework for Language-Model Deception Research — arXiv:2609.04166
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.04166
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-17T01:11:35.903Z
- SHA-256 of the archived text
2d428c16bca166dc938882ec6ff9a09d2eb4e15636fbb564d0e8dee676f58986- Stable URL of this page
https://metraipolis.com/documento/dc_0350cc323dd5
How to cite
From Deceptive Outputs to Deceptive Mechanisms: A Causal Framework for Language-Model Deception Research — arXiv:2609.04166 — arXiv (en), https://arxiv.org/abs/2609.04166 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_0350cc323dd5 (Metraipolis archive dc_0350cc323dd5, sha256 2d428c16bca166dc938882ec6ff9a09d2eb4e15636fbb564d0e8dee676f58986)
BibTeX
@misc{dc0350cc323dd5,
title = {From Deceptive Outputs to Deceptive Mechanisms: A Causal Framework for Language-Model Deception Research — arXiv:2609.04166},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.04166},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_0350cc323dd5; sha256 2d428c16bca166dc938882ec6ff9a09d2eb4e15636fbb564d0e8dee676f58986},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.04166. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Research and news coverage of language-model deception increasingly attributes human-like mental-state concepts to language models. Such claims can blur the distinction between behavior that looks deceptive and a mechanism that is actually deceptive. We introduce a causal taxonomy separating prior commitment from retrospective report, model preference from realized output, false preference from sensitivity to the utility of misleading a recipient, and deceptive behavior from the provenance of the objective or strategy producing it. We test these distinctions in two open-weight model families. Across controlled guessing-game and stock-trading experiments, we find that deceptive-looking behavior can arise without the corresponding proposed mechanism, while other interventions provide direct evidence that recipient information state can causally affect deceptive preference. These results show that deceptive behavior can provide evidence for a deceptive mechanism. But even evidence for such a mechanism does not establish model agency in the deception. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-04 in cs.AI). Categorie arXiv: Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.04166 — 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 2026-10-17T00:28:02.915Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-04?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).