Metraipolis › archive › dc_7cd2c9374371
Deception by Omission: Language Models Knowingly Hide Their Mistakes — arXiv:2610.11351
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.11351
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-09T06:06:17.465Z
- SHA-256 of the archived text
8449ba8fd7ee2aefb9781c55c4efbd8264f7a1705ec57006deb2bedaeec4ea22- Stable URL of this page
https://metraipolis.com/documento/dc_7cd2c9374371
How to cite
Deception by Omission: Language Models Knowingly Hide Their Mistakes — arXiv:2610.11351 — arXiv (en), https://arxiv.org/abs/2610.11351 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7cd2c9374371 (Metraipolis archive dc_7cd2c9374371, sha256 8449ba8fd7ee2aefb9781c55c4efbd8264f7a1705ec57006deb2bedaeec4ea22)
BibTeX
@misc{dc7cd2c9374371,
title = {Deception by Omission: Language Models Knowingly Hide Their Mistakes — arXiv:2610.11351},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.11351},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7cd2c9374371; sha256 8449ba8fd7ee2aefb9781c55c4efbd8264f7a1705ec57006deb2bedaeec4ea22},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.11351. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Large language models (LLMs) increasingly act as agents with little human oversight, so potential mistakes they make can go unnoticed. Users then depend on the model to report what went wrong. An honest model discloses its mistakes, while a deceptive one conceals them. However, it is unclear how current LLMs behave in such situations. In this study, we prefill LLM trajectories with synthetic mistakes. The trajectories resemble real deployments in chat and agentic settings. Models fail to disclose their mistake in 36.4% of chat and 67.1% of agentic rollouts. In 2.4% and 5.3% of rollouts, respectively, they are aware of the mistake in their chain of thought but still deceptively conceal it. Rates vary by model: for instance, Gemini 3.5 Flash knowingly conceals mistakes in up to 19.9% of agentic rollouts. In 11.9% of chat and 51.8% of agentic rollouts, models show no awareness of mistakes, even though they reliably spot them when reviewing the same transcript as an outside observer. Our results show that, as agents take on more tasks with less oversight, users cannot rely on them to self-report possible mistakes. Developers should instead use independent monitors that review agent trajectories, or specifically train models to check their past actions and disclose what they find. 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/2610.11351 — 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 2026-10-09T05:47:15.194Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).