Metraipolis › archive › dc_b032df2bcb75

Not One Belief: Answer Payloads, Persistence, and Dissociated Confidence Signals in Conversational Question Answering — arXiv:2610.35694

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.35694
License (as archived): CC0-1.0
Archived: 2026-10-27T08:12:29.321Z
SHA-256 of the archived text: 4c9bf1cce81ab68cbd29ea91e2917f965a078a8e37134ba7786a8f06b334871e
Stable URL of this page: https://metraipolis.com/documento/dc_b032df2bcb75

How to cite

Not One Belief: Answer Payloads, Persistence, and Dissociated Confidence Signals in Conversational Question Answering — arXiv:2610.35694 — arXiv (en), https://arxiv.org/abs/2610.35694 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_b032df2bcb75 (Metraipolis archive dc_b032df2bcb75, sha256 4c9bf1cce81ab68cbd29ea91e2917f965a078a8e37134ba7786a8f06b334871e)

BibTeX

@misc{dcb032df2bcb75,
  title = {Not One Belief: Answer Payloads, Persistence, and Dissociated Confidence Signals in Conversational Question Answering — arXiv:2610.35694},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.35694},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_b032df2bcb75; sha256 4c9bf1cce81ab68cbd29ea91e2917f965a078a8e37134ba7786a8f06b334871e},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.35694. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Conversational QA systems should fix errors when users push back and signal which answers to trust. We split each follow-up into its speech act (re-ask, agree, or challenge) and its payload (an answer string, correct or not), and measure four confidence signals on the same transcripts: verbalized confidence, length-normalized likelihood, a frozen hidden-state probe, and P(True). Answers change when the user supplies an answer string and almost never under the speech act alone: in every retained full-size cell, the correct answer converts at least four times as many wrong answers as re-asking or agreement. The four signals disagree. On Qwen2.5-7B the frozen probe rises with answer persistence rather than truth; this holds on held-out questions, with Llama-3-8B (cleaned answers) as support, but not on four other models. Among 81 Qwen2.5-7B errors that survive one re-ask, re-asking, a fresh chat, or abstain repair almost none, while a gold answer repairs 75-95%. In two preregistered replications, every repair rate without an answer string stays below 0.15; a larger post-hoc rerun excludes 0.15 except for the empty elicit. With BM25 TriviaQA evidence, sentences that contain the answer repair 67-76% of persistent errors versus at most 12% when they do not. An end-to-end contain-gate cuts harm on TriviaQA, but a deployable gate that does not know the gold answer repairs few errors.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2610.35694 — 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 2026-10-27T07:56:00.276Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).