Skip to content
AI News HubLIVE
Original source2 min read

Dense Supervision Is Not Enough: The Readout Blind Spot in Looped Language Models

Summary

This paper shows that in looped language models, dense per-loop cross-entropy supervises only the variables exposed by the readout, not all variables active in the recurrent transition. Hidden-state scale provides a concrete failure mode: scale-invariant readouts such as RMSNorm and LayerNorm hide radial scale from the loss, while pre-norm residual recurrence continues to carry and update that scale. Thus dense supervision can train early exits without controlling recurrent scale. The design rule is to either make scale visible to a loss or remove it from the loop; scale-controlled variants achieve lower perplexity at matched inference-depth operating points.

SourcearXiv Machine LearningAuthor: Rituraj Sharma, Tu Vu
Dense Supervision Is Not Enough: The Readout Blind Spot in Looped Language Models
Report an error

The correction channel is not available yet. You can copy the article reference below for later.

Correction instructions
Read article

[2606.24898] Dense Supervision Is Not Enough: The Readout Blind Spot in Looped Language Models

[Submitted on 12 Jun 2026]

Title:Dense Supervision Is Not Enough: The Readout Blind Spot in Looped Language Models

View a PDF of the paper titled Dense Supervision Is Not Enough: The Readout Blind Spot in Looped Language Models, by Rituraj Sharma and 1 other authors

View PDF HTML (experimental)

Abstract:Looped language models turn hidden states into runtime state: each state is decoded for prediction and fed back into future computation. This creates a basic supervision question: which state variables does cross-entropy actually control? We show that dense per-loop cross-entropy controls the variables exposed by the readout, not every variable active in the recurrent transition. Hidden-state scale gives a concrete failure mode. Scale-invariant readouts such as RMSNorm and LayerNorm hide radial scale from the immediate cross-entropy loss, while pre-norm residual recurrence continues to carry and update that same scale. Thus per-loop loss can make early exits usable without controlling recurrent scale. In 44M and 129M looped transformers without inter-loop normalization, per-loop cross-entropy through RMSNorm readouts still drives final hidden-state norms into the thousands or tens of thousands. Scale-visible readouts and explicit norm penalties keep norms in the tens, and scale-removing recurrence is the complementary architectural fix. The resulting design rule is simple: dense supervision trains exits; recurrent scale control requires either making scale visible to a loss or removing it from the loop. Consistent with this rule, scale-controlled variants achieve lower perplexity at matched inference-depth operating points in our variable-depth benchmarks.

Subjects:

Machine Learning (cs.LG); Artificial Intelligence (cs.AI)

Cite as: arXiv:2606.24898 [cs.LG]

(or arXiv:2606.24898v1 [cs.LG] for this version)

https://doi.org/10.48550/arXiv.2606.24898

arXiv-issued DOI via DataCite

Submission history

From: Rituraj Sharma [view email] [v1] Fri, 12 Jun 2026 13:23:18 UTC (1,411 KB)

Full-text links:

Access Paper:

View a PDF of the paper titled Dense Supervision Is Not Enough: The Readout Blind Spot in Looped Language Models, by Rituraj Sharma and 1 other authors

View PDF

HTML (experimental)

TeX Source

view license

Current browse context:

cs.LG

new | recent | 2026-06

Change to browse by:

cs cs.AI

References & Citations

NASA ADS

Google Scholar

Semantic Scholar

Loading...

Data provided by:

Bibliographic Tools

Bibliographic and Citation Tools

Bibliographic Explorer Toggle

Bibliographic Explorer (What is the Explorer?)

Connected Papers Toggle

Connected Papers (What is Connected Papers?)

Litmaps Toggle

Litmaps (What is Litmaps?)

scite.ai Toggle

scite Smart Citations (What are Smart Citations?)

Code, Data, Media

Code, Data and Media Associated with this Article

alphaXiv Toggle

alphaXiv (What is alphaXiv?)

Links to Code Toggle

CatalyzeX Code Finder for Papers (What is CatalyzeX?)

DagsHub Toggle

DagsHub (What is DagsHub?)

GotitPub Toggle

Gotit.pub (What is GotitPub?)

Huggingface Toggle

Hugging Face (What is Huggingface?)

ScienceCast Toggle

ScienceCast (What is ScienceCast?)

Demos

Demos

Replicate Toggle

Replicate (What is Replicate?)

Spaces Toggle

Hugging Face Spaces (What is Spaces?)

Spaces Toggle

TXYZ.AI (What is TXYZ.AI?)

Related Papers

Recommenders and Search Tools

Link to Influence Flower

Influence Flower (What are Influence Flowers?)

Core recommender toggle

CORE Recommender (What is CORE?)

IArxiv recommender toggle

IArxiv Recommender (What is IArxiv?)

Author

Venue

Institution

Topic

About arXivLabs

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)

Key points and analysis

Article intelligence

EngineersAdvanced

Key points

  • Per-loop cross-entropy controls variables exposed by the readout, not every variable active in the recurrent transition.
  • In 44M and 129M looped transformers without inter-loop normalization, per-loop cross-entropy through RMSNorm readouts drives final hidden-state norms into the thousands or tens of thousands.
  • Design rule: dense supervision trains exits; recurrent scale control requires making scale visible to a loss or removing it from the loop.
  • Scale-controlled variants achieve lower perplexity at matched inference-depth operating points in variable-depth benchmarks.

Highlights and analysis are generated automatically and may contain errors. Check the original source.