Metraipolis › archive › dc_72af59aba509

Not All Answers Are Contextually Persuadable: Inference Dynamics in Large Language Models under Contextual Influence — arXiv:2610.04791

Source
arXiv
Original
https://arxiv.org/abs/2610.04791
License (as archived)
CC0-1.0
Archived
2026-10-06T05:24:47.705Z
SHA-256 of the archived text
e27262ea0cf55126da993cb63e403ff0ae3de65c83db455bbd9430cafc80b416
Stable URL of this page
https://metraipolis.com/documento/dc_72af59aba509

How to cite

Not All Answers Are Contextually Persuadable: Inference Dynamics in Large Language Models under Contextual Influence — arXiv:2610.04791 — arXiv, https://arxiv.org/abs/2610.04791 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_72af59aba509 (Metraipolis archive dc_72af59aba509, sha256 e27262ea0cf55126da993cb63e403ff0ae3de65c83db455bbd9430cafc80b416)
BibTeX
@misc{dc72af59aba509,
  title = {Not All Answers Are Contextually Persuadable: Inference Dynamics in Large Language Models under Contextual Influence — arXiv:2610.04791},
  howpublished = {arXiv},
  url = {https://arxiv.org/abs/2610.04791},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_72af59aba509; sha256 e27262ea0cf55126da993cb63e403ff0ae3de65c83db455bbd9430cafc80b416},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv, https://arxiv.org/abs/2610.04791. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

At the core of modern prompting techniques is contextual sensitivity, the ability of large language models to adapt their predictions based on inference-time context. Despite its central role, inference behavior under strong contextual influence remains poorly understood, particularly at the level of internal inference dynamics. We introduce a theoretical framework for analyzing contextual influence through inference dynamics, enabling quantitative characterization of inference behavior beyond output-level answer changes. Our analysis shows that inference dynamics do not exhibit unbounded drift under repeated contextual assertions. Instead, predictive representations converge to stable, query-dependent regimes that fundamentally constrain whether contextual signals can alter a model's prediction. This leads to a surprising finding: Repeated contextual assertions do not act as accumulating evidence during inference and may therefore fail to alter a model's prediction even under unbounded repetition, while in other cases a prediction change becomes inevitable. We empirically validate our theoretical predictions, demonstrating strong alignment between theory and observed inference behavior. These contributions offer a principled pathway toward characterizing the limits of contextual influence during inference, providing practical implications for model development.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.04791 — 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 2026-10-06T05:24:05.515Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).