Metraipolis › archive › dc_bb069b59e6d1
The Answer Is Not the Argument — arXiv:2609.00264
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.00264
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-17T06:31:45.876Z
- SHA-256 of the archived text
45fe484ba014e124db73c1b22e0b250f7e4d437b38dc18c52c1755cc6ea9ca16- Stable URL of this page
https://metraipolis.com/documento/dc_bb069b59e6d1
How to cite
The Answer Is Not the Argument — arXiv:2609.00264 — arXiv (en), https://arxiv.org/abs/2609.00264 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_bb069b59e6d1 (Metraipolis archive dc_bb069b59e6d1, sha256 45fe484ba014e124db73c1b22e0b250f7e4d437b38dc18c52c1755cc6ea9ca16)
BibTeX
@misc{dcbb069b59e6d1,
title = {The Answer Is Not the Argument — arXiv:2609.00264},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.00264},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_bb069b59e6d1; sha256 45fe484ba014e124db73c1b22e0b250f7e4d437b38dc18c52c1755cc6ea9ca16},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.00264. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Chain-of-thought monitoring is proposed for AI oversight, yet evaluations often provide monitors with a trusted reference answer. We ask whether answer access improves verification of the reasoning or mainly supplies information about its conclusion. We collected 237 naturally generated, step-numbered solutions to 79 Humanity's Last Exam physics questions and independently labelled final-answer correctness and the first false step. Eight LLM monitors evaluated the traces with varying access to the reference answer. Certification raised mean balanced accuracy from 0.637 to 0.796, but its effect on error detection depended strongly on the conclusion: recall increased by +0.299 on wrong-answer traces, while there was no evidence of improvement on correct-answer traces containing a reasoning error (-0.083, 95% CI [-0.196, +0.030]). We then held the reasoning trace fixed in a seven-monitor certificate-congruence intervention. Replacing the true certificate with the trace's own incorrect conclusion reduced flagging by 0.659 (95% CI [0.602, 0.711]) and left flagging 0.389 below the answer-blind level. Conversely, a conflicting false certificate increased flagging of clean traces by 0.580, with 82.9% of newly flagged cases assigning the alleged error to an interior reasoning step. Trusted-answer access can therefore make monitoring appear substantially stronger because aggregate performance combines independent reasoning verification with a powerful certificate-conclusion consistency signal. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-02 in cs.AI). Categorie arXiv: Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.00264 — 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 2026-10-17T06:15:12.579Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-02?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).