onóna: Twin Embeddings with Tied Unembeddings
The paper introduces onóna, a two-table parameterization for Transformer language models that preserves input-output weight sharing while increasing output expressiveness.
Source-based brief · Editor review not recorded · See checks and limits ↓
One primary source; other links are context, not independent corroboration.
- Primary sourceHugging Face
- Discovery linkHugging Face
- Linked in sourceView arXiv page
- Linked in sourceView PDF
The paper page is linked. Peer review and findings have not been independently checked by AINewsIQ.
Open full article view →- Original paper+45
- Fetched source page+30
- No editor link review+0
- Original non-discussion link+5
This measures the stored source trail and editorial steps. It does not certify accuracy, peer review, or independent corroboration. Popularity adds no points.
View full scoring →Read the scoring rubric →What the source says
The paper introduces onóna, a two-table parameterization for Transformer language models that preserves input-output weight sharing while increasing output expressiveness. The input uses the average of two embedding tables, and the output head combines the softmax distributions from those tables via a gate conditioned on the current input token. Under matched token budgets across Dense, Gated DeltaNet hybrid, mixture-of-experts, and Attention Residual architectures at 142M, 330M, and 1B scales, the authors report lower validation loss and perplexity in 11 of 12 comparisons and higher average downstream accuracy in 10 of 12. The cost is one extra embedding table and one extra vocabulary projection.
Claim explorer
Attribution, source and conditionsSource-stated points from the stored brief. Claim-level assessments have not been recorded.
Source claims not independently checked; results come from the authors' own evaluations at three scales and four architectures.
This connects a source, an attributed statement and its qualifications. It does not establish that the claim is correct.
Questions worth asking
Open questions for further reportingWhat remains unconfirmed?
Source claims not independently checked; results come from the authors' own evaluations at three scales and four architectures.
What material was available for this brief?
The recorded brief basis is fetched source page. These points are AINewsIQ paraphrases, not quotations or independent confirmations.
onóna uses two embedding tables: the Transformer input is their average, and the output head mixes their softmax distributions.
Source: Hugging Face ↗Stored source-based brief · key point 1 · AINewsIQ paraphraseA gate conditioned only on the current input token sets the mixture weight, and its values can be precomputed for inference.
Source: Hugging Face ↗Stored source-based brief · key point 2 · AINewsIQ paraphraseAcross 12 matched-budget comparisons, onóna lowers final validation loss and external perplexity in 11, with the 1B MoE model as the loss exception.
Source: Hugging Face ↗Stored source-based brief · key point 3 · AINewsIQ paraphraseAverage accuracy over eight downstream tasks improves in 10 of 12 comparisons; the stated cost is one extra embedding table and one extra vocabulary projection.
Source: Hugging Face ↗Stored source-based brief · key point 4 · AINewsIQ paraphraseLimits from the stored brief
Source claims not independently checked; results come from the authors' own evaluations at three scales and four architectures.
This is a source-based abstract, not independent validation of the source's claims.
Evidence & sources
These links show where the signal came from and what the source page linked to. A linked page is context, not independent corroboration.
Related archive coverage
These relationships record a citation or topic overlap, not confirmation of the finding.
The paper page is linked. Peer review and findings have not been independently checked by AINewsIQ.
Timeline & updates
- Signal dated by source or discovery feed
This timestamp places the signal in the radar. It may be the discovery post date rather than the primary page's first publication date.
- Shared brief stored
AINewsIQ prepared one abstract from fetched source page.
Brief history · 1 retained version
These are changes to AINewsIQ’s own publication. Earlier history is unavailable unless recorded here. Complete retained versions are available to the editor.
- Version 1 · published
Source-based brief stored or updated. This does not establish that the publisher changed its page.
Automated record
Source observations · 1 recent captures
A changed capture means the extracted text differed. It can reflect page furniture or extraction changes; it does not establish a substantive correction. Source captures remain private.
Retained capture; no earlier comparison available.
New engagement snapshots do not mean the source itself changed. Source observations and AINewsIQ brief revisions are separate records.
Why this may matter
Weight tying is a common way to cut parameter count in language models. If the reported gains hold, a modest added cost could make the tied output head more expressive without giving up input-output sharing.
Limits and open questions
Source claims not independently checked; results come from the authors' own evaluations at three scales and four architectures.
Organization-specific financial, security, and operational effects require the organization's own context and review.

Join the discussion
Have a different perspective or a question about the evidence? Discussion is free with a verified email account. Every comment is reviewed before publication.
No published comments yet.
New commenters: 2 submissions per 24 hours, at least 5 minutes apart, until 3 are approved. Discussion guidelines and moderation →