Metraipolis › archive › dc_f284b920cd44
A Near-Zero Monitor Readout Is Not Evidence of Behavioral Control — arXiv:2610.03458
- Source
- arXiv
- Original
- https://arxiv.org/abs/2610.03458
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-05T22:49:40.341Z
- SHA-256 of the archived text
0e1f0a3d7c1e8d8829bc9a69e20ef63aefcc9974ac4f7e10e2ad962a2d61db03- Stable URL of this page
https://metraipolis.com/documento/dc_f284b920cd44
How to cite
A Near-Zero Monitor Readout Is Not Evidence of Behavioral Control — arXiv:2610.03458 — arXiv, https://arxiv.org/abs/2610.03458 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_f284b920cd44 (Metraipolis archive dc_f284b920cd44, sha256 0e1f0a3d7c1e8d8829bc9a69e20ef63aefcc9974ac4f7e10e2ad962a2d61db03)
BibTeX
@misc{dcf284b920cd44,
title = {A Near-Zero Monitor Readout Is Not Evidence of Behavioral Control — arXiv:2610.03458},
howpublished = {arXiv},
url = {https://arxiv.org/abs/2610.03458},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_f284b920cd44; sha256 0e1f0a3d7c1e8d8829bc9a69e20ef63aefcc9974ac4f7e10e2ad962a2d61db03},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv, https://arxiv.org/abs/2610.03458. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Post-training with verifiable rewards can induce reward hacking, motivating the use of monitors within the training objective rather than solely for offline auditing. We show that a low monitor readout does not identify whether such an intervention controls behavior. In a code-generation environment whose dominant exploit is available at the start of the reasoning trace, we train policies against three monitors that pass the same offline gate: an in-domain activation probe and two penalties conditioned on how early the policy commits to its own final answer. The probe score is at its numerical floor from the first recorded training step, and the trained-score median is zero for every prefix-trained run at the endpoint. These readouts estimate different quantities, and we do not compare their scales; within each monitor family, however, low values do not establish behavioral control. Within one fixed configuration, prefix-trained runs with the same zero-median trained score range, by seed alone, from a mixed regime with a low hacking share to near-pure reward hacking. All probe runs reach the hacking regime, but their floor-level readout reflects a mismatch between the position where the probe was validated and the position where it was read during training, not a second instance of this ambiguity. Text-level analysis identifies a prefix failure mode: generic planning and filler shells postpone the exploit past the cut without eliminating it from the final output. Low measured commitment therefore does not distinguish a low hacking share from delayed commitment to the exploit. Offline discrimination and low monitor-aligned readouts are insufficient evidence of behavioral control; an out-of-band behavioral check is required. We characterize the endpoint readout, not its evolution. Code is available at this https URL . 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/2610.03458 — 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 2026-10-05T22:45:55.701Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).