Metraipolis › archive › dc_1206aa1314cd
Shaping & Monitoring Models through their Internal Representations — arXiv:2611.07071
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2611.07071
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-06T14:42:10.396Z
- SHA-256 of the archived text
38695ee5a45fed82fa53027f1ec475dc8fa617854e237758dab0eafadb9f49f8- Stable URL of this page
https://metraipolis.com/documento/dc_1206aa1314cd
How to cite
Shaping & Monitoring Models through their Internal Representations — arXiv:2611.07071 — arXiv (en), https://arxiv.org/abs/2611.07071 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_1206aa1314cd (Metraipolis archive dc_1206aa1314cd, sha256 38695ee5a45fed82fa53027f1ec475dc8fa617854e237758dab0eafadb9f49f8)
BibTeX
@misc{dc1206aa1314cd,
title = {Shaping & Monitoring Models through their Internal Representations — arXiv:2611.07071},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2611.07071},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_1206aa1314cd; sha256 38695ee5a45fed82fa53027f1ec475dc8fa617854e237758dab0eafadb9f49f8},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.07071. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
We propose Mechanistic Shaping, a unified framework for shaping models through their internal representations. Mechanistic Shaping is applied during training so that models follow specifications (rules) about which concepts they express, measured via their outputs or internal representations (e.g., no harmful content), and also which concepts they rely on, i.e., what they use to make their decisions (e.g., not a candidate's gender when screening a résumé). Across multiple domains (vision, medical imaging, language), scales (up to 32B parameters), and training paradigms (classification, SFT, DPO, GRPO), we find that Mechanistic Shaping substantially reduces targeted behaviors while preserving task performance. Since training against a representation can also modify it, we ask whether shaped concepts remain monitorable, a question that depends on the class of monitors considered. We therefore introduce a hierarchy of such classes, which separates a concept that escapes one probe from a concept that is truly gone. We characterize three geometric mechanisms of circumvention: the concept can drift away from the monitor, be erased from the layer, or persist in higher-order statistics as a nonlinear encoding, which we explain through a product-code construction. Although other mechanisms may arise in different settings or at larger scales, these three account for all of our observations, and each can be detected by a matching monitor. In practice, we find that shaped concepts largely remain monitorable: refitted linear probes recover the concept in most settings, and the original probe remains reliable whenever gradients do not flow through it. Taken together, these results point to the potential of using interpretability as a tool for training, and call for a correspondingly rigorous science of interpretability-based training. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2611.07071 — 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 2026-11-06T14:18:42.666Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).