Metraipolis › archive › dc_06ce272f9ffc

Entertain, then Cancel: How Thinking Makes Reasoning Models Read Some and Or Literally — arXiv:2610.21204

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.21204
License (as archived)
CC0-1.0
Archived
2026-10-16T09:02:49.343Z
SHA-256 of the archived text
9e45743cfcbfcdff150f88abc92924bfa80c7c31a6ddb913e8237dabecb09cb6
Stable URL of this page
https://metraipolis.com/documento/dc_06ce272f9ffc

How to cite

Entertain, then Cancel: How Thinking Makes Reasoning Models Read Some and Or Literally — arXiv:2610.21204 — arXiv (en), https://arxiv.org/abs/2610.21204 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_06ce272f9ffc (Metraipolis archive dc_06ce272f9ffc, sha256 9e45743cfcbfcdff150f88abc92924bfa80c7c31a6ddb913e8237dabecb09cb6)
BibTeX
@misc{dc06ce272f9ffc,
  title = {Entertain, then Cancel: How Thinking Makes Reasoning Models Read Some and Or Literally — arXiv:2610.21204},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.21204},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_06ce272f9ffc; sha256 9e45743cfcbfcdff150f88abc92924bfa80c7c31a6ddb913e8237dabecb09cb6},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.21204. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

When a speaker says ``some students passed'', listeners infer that not all of them did. Reasoning models now let users decide how long a model thinks before it answers, but we do not know what extra thinking does to such everyday inferences. We study this directly: we keep a model's weights fixed, change only how much it thinks, and record which reading it chooses. A second tool cuts the model's own reasoning at successive points and asks for an answer, showing where the reading changes. Thinking makes models read logical words as logic: it cancels the implicatures of \emph{some} (\emph{not all}) and \emph{or} (\emph{not both}) in most tested open reasoning models and in GPT-5, while context-driven inferences such as indirect answers improve. The trace shows why: models first entertain the enriched reading and cancel it later. Answering from that early reading, an exploratory inference-time remedy, recovers these implicatures without losing literal accuracy. Evaluations of reasoning models should measure how they interpret, not only how accurate they are.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI) ; Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2610.21204 — 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 2026-10-16T08:44:17.718Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).