Metraipolis › archive › dc_445b289a13e8
Looped Transformers under the Jacobian Lens: Does the Global Workspace Survive Recurrence? — arXiv:2609.01924
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.01924
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-17T03:25:41.578Z
- SHA-256 of the archived text
26f0356853fb2d637f0f9ec79783bbc26e31bc45290b184fee998344e2012429- Stable URL of this page
https://metraipolis.com/documento/dc_445b289a13e8
How to cite
Looped Transformers under the Jacobian Lens: Does the Global Workspace Survive Recurrence? — arXiv:2609.01924 — arXiv (en), https://arxiv.org/abs/2609.01924 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_445b289a13e8 (Metraipolis archive dc_445b289a13e8, sha256 26f0356853fb2d637f0f9ec79783bbc26e31bc45290b184fee998344e2012429)
BibTeX
@misc{dc445b289a13e8,
title = {Looped Transformers under the Jacobian Lens: Does the Global Workspace Survive Recurrence? — arXiv:2609.01924},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.01924},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_445b289a13e8; sha256 26f0356853fb2d637f0f9ec79783bbc26e31bc45290b184fee998344e2012429},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.01924. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Recent work identifies a mid-depth band of verbalisable, causally potent representations in a standard feedforward transformer --- a functional analogue of a global workspace. Whether the same workspace functionality emerges when depth is implemented through recurrence rather than a stack of distinct layers remains unknown. Looped and depth-recurrent transformers provide a direct test of this question because they reuse the same weights across depth. We extend the Jacobian lens to iterated architectures using a virtual-unrolling adapter. We apply the full workspace suite --- lens fitting, readout, and eleven causal experiment families --- to Ouro-2.6B (48 layers looped 4 times, deeply supervised) and Huginn-0125 (a 4-layer core recurred 16 times, trained for latent reasoning), using Qwen3.6-27B (64 untied layers) as the standard baseline. We find that a workspace forms in the iterated part of each architecture, but that recurrence changes how it can be accessed. Ouro reconstructs workspace content in every loop, and linear transport cannot carry that content across loop boundaries; writes and ablations must therefore span every remaining loop. Huginn carries content forward across all sixteen recurrences, while reads, writes, and ablations act only within a sliding window of roughly two recurrences. Whether newly injected content can be verbalised tracks explicit per-iteration supervision; whether existing content can be steered does not. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-03 in cs.AI). Categorie arXiv: Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.01924 — 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 2026-10-17T03:13:12.408Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-03?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).