Metraipolis › archive › dc_8fe07689df55
Beyond Depth Truncation: Controlled Evaluation of Depth Utilization in Recursive Language Models — arXiv:2609.19934
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.19934
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-10T03:20:49.854Z
- SHA-256 of the archived text
ffb785c2df563c077dc9b769f8073fa78d43e399950b03799e8310764fa04c7f- Stable URL of this page
https://metraipolis.com/documento/dc_8fe07689df55
How to cite
Beyond Depth Truncation: Controlled Evaluation of Depth Utilization in Recursive Language Models — arXiv:2609.19934 — arXiv (en), https://arxiv.org/abs/2609.19934 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_8fe07689df55 (Metraipolis archive dc_8fe07689df55, sha256 ffb785c2df563c077dc9b769f8073fa78d43e399950b03799e8310764fa04c7f)
BibTeX
@misc{dc8fe07689df55,
title = {Beyond Depth Truncation: Controlled Evaluation of Depth Utilization in Recursive Language Models — arXiv:2609.19934},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.19934},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_8fe07689df55; sha256 ffb785c2df563c077dc9b769f8073fa78d43e399950b03799e8310764fa04c7f},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.19934. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Depth-recurrent language models iteratively apply a small layer stack, decoupling per-token compute from distinct parameter count. To determine whether such a model genuinely utilizes its depth, both recurrence and layer-pruning literatures rely on a shared evaluation: truncating depth at inference time, plotting quality against retained depth fraction, and reading off the slope. While cheap and training-free, this metric suffers from an unexamined flaw: it extracts a single scalar from an intervention that alters multiple model properties simultaneously. Depth truncation concurrently reduces the number of block applications, decreases the volume of distinct computation performed, and pushes the readout head onto an out-of-distribution residual stream. The observed slope conflates all three factors, yet is conventionally interpreted as reflecting solely the second. We propose the Depth Control Protocol (DCP), a diagnostic suite that disentangles these three quantities. DCP comprises three positive controls that isolate each factor while varying the others, a negative control applying the identical interventions to dense transformers to ensure the effect is not an artifact of the measurement protocol, and a controlled training intervention to verify causality. The linchpin control, running the full budget of block applications while executing only a single distinct iteration, is strictly realizable only in depth-wise weight-sharing architectures, since in a dense network repeating a layer yields an entirely different model rather than the same model in an alternative configuration. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-18 in cs.AI). Categorie arXiv: Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.19934 — 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 2026-10-10T03:04:13.669Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-18?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).