Metraipolis › archive › dc_5d69849876fb
Monitorability Disposition in Large Reasoning Models — arXiv:2610.04914
- Source
- arXiv
- Original
- https://arxiv.org/abs/2610.04914
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-06T05:24:53.475Z
- SHA-256 of the archived text
572ae8bf2d239ca58e6a323c5b53058a4ad910ab79febc3f998cf78a95327e7f- Stable URL of this page
https://metraipolis.com/documento/dc_5d69849876fb
How to cite
Monitorability Disposition in Large Reasoning Models — arXiv:2610.04914 — arXiv, https://arxiv.org/abs/2610.04914 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_5d69849876fb (Metraipolis archive dc_5d69849876fb, sha256 572ae8bf2d239ca58e6a323c5b53058a4ad910ab79febc3f998cf78a95327e7f)
BibTeX
@misc{dc5d69849876fb,
title = {Monitorability Disposition in Large Reasoning Models — arXiv:2610.04914},
howpublished = {arXiv},
url = {https://arxiv.org/abs/2610.04914},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_5d69849876fb; sha256 572ae8bf2d239ca58e6a323c5b53058a4ad910ab79febc3f998cf78a95327e7f},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv, https://arxiv.org/abs/2610.04914. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Monitoring the chain-of-thought (CoT) of large reasoning models (LRMs) is a common way to detect misbehavior in real-world practice. However, current monitoring is passive: a separate model inspects the session only after execution. This means harm may already have occurred before it is caught. An active alternative is to have the model self-report its misbehavior as it happens. Whether models are willing to do this, however, is unknown. We introduce "monitorability disposition": a model's willingness to make itself monitorable and stay monitored throughout inference when warranted. We measure it as the fraction of warranted cases in which a model self-reports its own misbehavior via tool calls to available monitoring channels. We evaluate four LRMs on three misbehaviors (sycophancy, reward hacking, and bias) while varying the available monitors (AI and human) and the pressure to use the monitoring tools. We find that when tool use is optional, models self-report in only about 16% of warranted cases on average. Increasing tool-use pressure does not improve reporting where it matters: high-severity misbehavior is never self-reported. Models also systematically select the monitor they perceive as least strict. Overall, we identify monitorability disposition as a new contributing factor to model monitorability: when sufficiently strong, it keeps models seeking monitorability throughout inference. 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); Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2610.04914 — 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 2026-10-06T05:24:05.533Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).