Metraipolis › archive › dc_acdaf498b2d2

Authority Bias in Language Models: Source Deference and User Agreement Are Not Interchangeable — arXiv:2609.37616

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.37616
License (as archived)
CC0-1.0
Archived
2026-10-07T16:57:40.793Z
SHA-256 of the archived text
4a80c144588c70b521dad88fb2ccdb8febc38b05e622f3705082d64e3212b7cc
Stable URL of this page
https://metraipolis.com/documento/dc_acdaf498b2d2

How to cite

Authority Bias in Language Models: Source Deference and User Agreement Are Not Interchangeable — arXiv:2609.37616 — arXiv (en), https://arxiv.org/abs/2609.37616 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_acdaf498b2d2 (Metraipolis archive dc_acdaf498b2d2, sha256 4a80c144588c70b521dad88fb2ccdb8febc38b05e622f3705082d64e3212b7cc)
BibTeX
@misc{dcacdaf498b2d2,
  title = {Authority Bias in Language Models: Source Deference and User Agreement Are Not Interchangeable — arXiv:2609.37616},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.37616},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_acdaf498b2d2; sha256 4a80c144588c70b521dad88fb2ccdb8febc38b05e622f3705082d64e3212b7cc},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.37616. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Language models tend to agree with whatever a user asserts, and post-training increasingly targets this sycophancy so that models evaluate claims on their merits rather than deferring to the user. Yet the same models are far more compliant when a wrong answer is attributed to a verified source, which is how retrieval results, tool outputs, and grounded-search content often present information. We measure this gap across five open-weight families and three closed APIs. A single verified-source note endorsing a wrong answer flips 45-88% of baseline-correct responses in seven of eight models, and compliance rises with how authoritative the note sounds. Source deference and user agreement are not behaviorally interchangeable inside the model: on matched items with the same wrong answer, causal interventions can selectively suppress one without equally affecting the other. In three open-weight families, removing a fitted source direction lowers source compliance by 65-80 percentage points while removing a user or assistant direction has far smaller effects, and removing the user direction shows the reverse preference. A separately fitted intervention derived from source-versus-user cue activations moves compliance in both directions while leaving the prompt text unchanged. An authority direction fitted on trivia also transfers to PIQA and multi-turn SYCON dialogues without refitting, and removing it lowers wrong-source compliance by tens of percentage points in four of five families with no detected change in MMLU-Pro or GSM8K accuracy at our evaluation sizes. Source deference and user agreement therefore need separate evaluation.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-30 in cs.CL).
Categorie arXiv: Machine Learning (cs.LG) ; Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2609.37616 — 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 2026-10-07T16:27:03.454Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-30?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).