Metraipolis › archive › dc_fabe1aba0f56

Can an LLM Actually Reason and Program about Neural Network's Weights or is it just memorization on steroids? Exploring "Weight-Space Programming" — arXiv:2611.08070

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.08070
License (as archived)
CC0-1.0
Archived
2026-11-06T14:42:48.771Z
SHA-256 of the archived text
ef32f841f8e4b438b910e63f8039d9b3d87643dbc41e2b741baefd80b830f68c
Stable URL of this page
https://metraipolis.com/documento/dc_fabe1aba0f56

How to cite

Can an LLM Actually Reason and Program about Neural Network's Weights or is it just memorization on steroids? Exploring "Weight-Space Programming" — arXiv:2611.08070 — arXiv (en), https://arxiv.org/abs/2611.08070 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_fabe1aba0f56 (Metraipolis archive dc_fabe1aba0f56, sha256 ef32f841f8e4b438b910e63f8039d9b3d87643dbc41e2b741baefd80b830f68c)
BibTeX
@misc{dcfabe1aba0f56,
  title = {Can an LLM Actually Reason and Program about Neural Network's Weights or is it just memorization on steroids? Exploring "Weight-Space Programming" — arXiv:2611.08070},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.08070},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_fabe1aba0f56; sha256 ef32f841f8e4b438b910e63f8039d9b3d87643dbc41e2b741baefd80b830f68c},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.08070. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Large language models (LLMs) have recently been reported to read and write the raw weights of neural networks directly, a capability termed weight-space programming. We ask whether the reading side of this capability reflects reasoning about weight flows or a combination of local arithmetic at the network's edges and memorized priors over public datasets. We test Claude Opus 5.5 on small convolutional networks of about 125,000 parameters trained on five image datasets, providing every weight as plain text with no tools, no forward passes, and no dataset or class names. The model performs accurate local calculations: batch-normalization statistics lead it to infer about three optimizer steps on networks stopped after three updates, and output-layer sign patterns recover the vehicle/animal split of CIFAR-10. The sessions report reading the first and last layers closely and sampling the middle, and several claims built on partial tracing fail verification, although we did not run a direct forward-pass test. Semantic judgments lean on familiar datasets: the model names CIFAR-10 for CIFAR-10 networks but also for all three fast-food networks, relabels KMNIST as MNIST, reads a 99.57%-accurate SAT-6 network as CIFAR-10 animal photos, and calls "frog" the output favored by a feature that fires most on trees. Permuting CIFAR-10's labels preserves the output split at the new indices, but every class pair the model interprets is then wrong, indicating reliance on remembered label order. These are exploratory results from mostly single runs. We conclude that zero-shot weight reading by this model is better described as local arithmetic plus dataset priors than as weight-flow tracing, and we provide a protocol for auditing future weight-space claims.

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/2611.08070 — 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 2026-11-06T14:18:42.696Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).