Metraipolis › archive › dc_ddbcadff7788

In-Context Neurofeedback: Can LLMs Control Their Internal Representations through Privileged Access? — arXiv:2609.00904

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.00904
License (as archived): CC0-1.0
Archived: 2026-10-17T06:32:16.713Z
SHA-256 of the archived text: 114558b7308141a02f5ccb28aa006bed7dcd0201aa34f060734a6e47ed172945
Stable URL of this page: https://metraipolis.com/documento/dc_ddbcadff7788

How to cite

In-Context Neurofeedback: Can LLMs Control Their Internal Representations through Privileged Access? — arXiv:2609.00904 — arXiv (en), https://arxiv.org/abs/2609.00904 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ddbcadff7788 (Metraipolis archive dc_ddbcadff7788, sha256 114558b7308141a02f5ccb28aa006bed7dcd0201aa34f060734a6e47ed172945)

BibTeX

@misc{dcddbcadff7788,
  title = {In-Context Neurofeedback: Can LLMs Control Their Internal Representations through Privileged Access? — arXiv:2609.00904},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.00904},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ddbcadff7788; sha256 114558b7308141a02f5ccb28aa006bed7dcd0201aa34f060734a6e47ed172945},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.00904. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Whether large language models (LLMs) can control their own internal representations matters for both machine metacognition and AI safety. A recent study applied neurofeedback to LLMs and claimed that they can control their internal representations. However, the reported control may rely on superficial mechanisms rather than genuine internal access because the control targets in that study are not privileged, meaning that a third party can infer them from the prompt. We redesign the neurofeedback paradigm for LLMs so that the control target satisfies the privileged access requirement, which is closer to neurofeedback experiments in human cognitive neuroscience. Under this stricter setting, the models do not demonstrate reliable control over privileged internal representations, suggesting that previously reported control cannot exclude the possibility that it relies on superficial mechanisms. Our results indicate that rigorous assessments of metacognition in LLMs require evaluation methods that demand privileged access.

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/2609.00904 — 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 2026-10-17T06:15:12.660Z 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).