Papers › Universal Transformers
Universal Transformers
Mostafa Dehghani, Stephan Gouws, Oriol Vinyals, Jakob Uszkoreit, Łukasz Kaiser
Recurrent neural networks (RNNs) sequentially process data by updating their state with each new data point, and have long been the de facto choice for sequence modeling tasks. However, their inherently sequential computation makes them slow to train. Feed-forward and convolutional architectures have recently been shown to achieve superior results on some sequence modeling tasks such as machine translation, with the added advantage that they concurrently process all inputs in the sequence, leading to easy parallelization and faster training times. Despite these successes, however, popular feed-forward sequence models like the Transformer fail to generalize in many simple tasks that recurrent models handle with ease, e.g. copying strings or even simple logical inference when the string or formula lengths exceed those observed at training time. We propose the Universal Transformer (UT), a parallel-in-time self-attentive recurrent sequence model which can be cast as a generalization of the Transformer model and which addresses these issues. UTs combine the parallelizability and global receptive field of feed-forward sequence models like the Transformer with the recurrent inductive bias of RNNs. We also add a dynamic per-position halting mechanism and find that it improves accuracy on several tasks. In contrast to the standard Transformer, under certain assumptions, UTs can be shown to be Turing-complete. Our experiments show that UTs outperform standard Transformers on a wide range of algorithmic and language understanding tasks, including the challenging LAMBADA language modeling task where UTs achieve a new state of the art, and machine translation where UTs achieve a 0.9 BLEU improvement over Transformers on the WMT14 En-De dataset.
In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.
Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.
For agents, on Syntology's MCP service (how to connect):
get_harvested_code_for_paper(arxiv_id="1807.03819")get_code(code_sha256="…")with a sample'scode_sha256from the list belowget_citation_path(paper_1="1807.03819", paper_2="…")with another paper's arXiv id or titleget_concepts_for_paper(arxiv_id="1807.03819")
Code
No sample from the archive's official repository (tensorflow/tensor2tensor) is linked to this paper; the 25 samples below come from 5 community repositories. Syntology harvested code from tensorflow/tensor2tensor for 3 other papers.
tensorflow/tensor2tensor: Syntology's graph links these samples to Tensor2Tensor for Neural Machine Translation (arXiv:1803.07416; that paper's own run record: 30 ran (of which 0 constructed an object rather than computing a result; 29 with no instrument failure: 0 honoured, 0 violated, 29 with no contract checked; 1 where Syntology's instrument failed) · 18 unverified), Model-Based Reinforcement Learning for Atari (arXiv:1903.00374; that paper's own run record: 16 ran (of which 8 constructed an object rather than computing a result; 10 with no instrument failure: 0 honoured, 0 violated, 10 with no contract checked; 6 where Syntology's instrument failed) · 5 unverified) and Simplify-then-Translate: Automatic Preprocessing for Black-Box Machine Translation (arXiv:2005.11197; that paper's own run record: 2 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 2 where Syntology's instrument failed) · 2 unverified); nothing checks that they implement this paper's method.
Syntology 25 code samples harvested from 5 repositories linked to this paper: 17 ran (of which 8 constructed an object rather than computing a result; 17 with no instrument failure: 1 honoured, 0 violated, 16 with no contract checked; 0 where Syntology's instrument failed) · 8 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code; here there are none.
By repository: community (archive-listed): 25 samples from 5 repositories, 17 ran (17 with no instrument failure; 0 where Syntology's instrument failed). The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.
No sample is linked to this paper from 2 of its repositories that Syntology harvested for other papers:
- tensorflow/tensor2tensor, official repository: harvested for 3 other papers, listed under the repository
- kpot/keras-transformer, community (archive-listed): harvested for another paper, listed under the repository
Levels written between 2026-09-22 and 2026-09-26 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.
Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.
A paper named beside a repository with no sample linked to this paper is shown with that paper's own run record, not this paper's: “ran” means executed on a synthesized input, not that the code is correct, and “where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Papers are listed in arXiv-id order, at most three per repository.
Code Syntology ran Syntology
25 samples harvested: 17 ran (of which 8 constructed an object rather than computing a result; 17 with no instrument failure: 1 honoured, 0 violated, 16 with no contract checked; 0 where Syntology's instrument failed) · 8 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page has not changed since , the first build that kept a record date for it; when this build read Syntology's graph is in the build record. That date is not when the samples ran.
Licence: 24 of the 25 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.
Harvested from 5 repositories linked to this paper, official or community; the list below is grouped by repository, each group headed by its name and what it is to the paper. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.
Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (how to connect); click an identity to copy that call.
Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of six reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.
A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity. A census tag appears when Syntology's census ran implementations with that name on one shared input per bucket: it gives how many were compared, from how many papers, and how many distinct outputs they produced, and names this sample's bucket when this sample was one of them.
82a2fb5195404bdd · report
fixture: signature stories:Tensor/2/int, questions:Tensor/2/int, outputs:Tensor/2/int, stories_vocab:SimpleVocab/0/?, questions_vocab:SimpleVocab/0/?, targets_vocab:SimpleVocab/0/?, targets:Tensor/2/int, identity_output:bool/0/bool · returned tuple · level recorded 2026-09-26 48da21be9c4b565a · report
fixture: signature state:Tensor/3/float, inputs:Tensor/3/float, fn:callable/0/?, time_enc:Tensor/3/float, pos_enc:Tensor/3/float, max_hop:int/0/int, encoder_output:NoneType/0/? · returned tuple · level recorded 2026-09-24 df7ba9e22f4021de · report
fixture: signature pondering:EagerTensor/2/float, halt_threshold:float/0/float · returned SymbolicTensor · level recorded 2026-09-24 ab44c54bef771f0e · report
fixture: returned EmbeddingSharedWeights · level recorded 2026-09-22 81f546780efe0c63 · report
fixture: returned FeedForwardNetwork · level recorded 2026-09-22 8f3fd62209af26f0 · report
fixture: returned LayerWrapper · level recorded 2026-09-22 97703c2875660e0b · report
fixture: returned MultiheadAttention · level recorded 2026-09-22 72219f63ad9d256a · report
fixture: signature input:EagerTensor/3/float, bias:EagerTensor/4/float, cache:NoneType/0/? · returned EagerTensor · level recorded 2026-09-24 785dd04c3d9161a4 · report
9d2c61e66b4346a7 · report
d65227646941963e · report
b3026a17c1914957 · report
50973ad904b644b9 · report
fixture: signature x:Tensor/2/int · returned Tensor · level recorded 2026-09-22 census: 13 implementations from 9 papers, 9 distinct outputs across 5 buckets, one shared input per bucket model/recurrent.py ran no licence file found · pointer only ·
55f808c7c4cab374 · report
fixture: signature x:Tensor/3/float · returned Tensor · level recorded 2026-09-24 1e5ecdd2d7834bc5 · report
fixture: signature x:Tensor/2/int, e_mask:Tensor/2/bool · returned Tensor · level recorded 2026-09-24 ac79741037ef4e2b · report
fixture: signature x:Tensor/2/int, memory:Tensor/3/float, e_mask:Tensor/2/bool, d_mask:Tensor/2/float · returned Tensor · level recorded 2026-09-22 58801803f9aacb41 · report
fixture: signature length:int/0/int, channels:int/0/int, min_timescale:float/0/float, max_timescale:float/0/float · returned Tensor · level recorded 2026-09-22 6dd3fd72476b96f4 · report
d939293e6a574006 · report
848001379d12e45b · report
8187fac026534417 · report
fixture: returned LayerNormalization · level recorded 2026-09-22 99fc6b1e929c1a86 · report
fixture: returned MultiHeadSelfAttention · level recorded 2026-09-22 d297901c2cb19187 · report
fixture: returned TransformerTransition · level recorded 2026-09-22 1eb7bb2c7019e81d · report
fixture: returned _BaseMultiHeadAttention · level recorded 2026-09-22 bfa58c8b92ee3953 · report
Method as code Syntology
What Syntology's graph records as this paper's claimed method, lifted into code: each assembly names its bricks (functions and classes), each with its own role and its own verification level. An assembly is not a check that the code implements the method correctly or reproduces the paper's results. Levels are each brick's own verification level in Syntology's graph (level 0: no recorded run), listed brick by brick and never combined into one level; “ran” means the brick executed on a synthesized input, not that its output is correct, and the instrument figure counts failures of Syntology's instrument, not of the code. Where a model read the code against the paper's claim, each assembly's label verdict is one model's judgement, kept in this page's JSON twin; it is not calibrated, so no card prints it. Siblings are assemblies lifted from this paper in the same batch for the same method: alternative readings kept side by side; none replaces or supersedes another.
4 assemblies on this paper: 4 harvested, 0 generated, listed in 1 groups (siblings lifted in one batch form one group).
4 sibling assemblies for Universal Transformer (harvested), lifted in the same batch on 2026-09-09: alternative readings kept side by side; none replaces or supersedes another.
harvested Assembly for Universal Transformer (graph method: Universal Transformer). 1 brick; individually on a fixture: 1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified. The assembly has not been run end-to-end.
one of 4 siblings
- ACT_basic entry · level 1 · ran ·
48da21be9c4b565a
harvested Assembly for Universal Transformer (graph method: Universal Transformer). 10 bricks; individually on a fixture: 6 ran (of which 4 constructed an object rather than computing a result; 6 with no instrument failure: 0 honoured, 0 violated, 6 with no contract checked; 0 where Syntology's instrument failed) · 4 unverified. The assembly has not been run end-to-end.
one of 4 siblings
- UniversalTransformer entry · level 0 · unverified ·
b3026a17c1914957 - DecoderStack helper · level 0 · unverified ·
785dd04c3d9161a4 - EmbeddingSharedWeights helper · level 1 · ran ·
ab44c54bef771f0e - EncoderStack helper · level 0 · unverified ·
9d2c61e66b4346a7 - ACT helper · level 1 · ran ·
df7ba9e22f4021de - FeedForwardNetwork helper · level 1 · ran ·
81f546780efe0c63 - LayerNormalization helper · level 0 · unverified ·
d65227646941963e - LayerWrapper helper · level 1 · ran ·
8f3fd62209af26f0 - MultiheadAttention helper · level 1 · ran ·
97703c2875660e0b - SelfAttention helper · level 1 · ran ·
72219f63ad9d256a
harvested Assembly for Universal Transformer (graph method: Universal Transformer). 5 bricks; individually on a fixture: 4 ran (of which 4 constructed an object rather than computing a result; 4 with no instrument failure: 0 honoured, 0 violated, 4 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified; every one of the 4 samples that ran constructed an object rather than computing a result. The assembly has not been run end-to-end.
one of 4 siblings
- TransformerBlock entry · level 0 · unverified ·
bfa58c8b92ee3953 - LayerNormalization model · level 1 · ran ·
8187fac026534417 - MultiHeadSelfAttention model · level 1 · ran ·
99fc6b1e929c1a86 - TransformerTransition model · level 1 · ran ·
d297901c2cb19187 - _BaseMultiHeadAttention helper · level 1 · ran ·
1eb7bb2c7019e81d
harvested Assembly for Universal Transformer (graph method: Universal Transformer). 9 bricks; individually on a fixture: 6 ran (of which 0 constructed an object rather than computing a result; 5 with no instrument failure: 1 honoured, 0 violated, 4 with no contract checked; 1 where Syntology's instrument failed) · 3 unverified. The assembly has not been run end-to-end.
one of 4 siblings
- RecurrentTransformer entry · level 0 · unverified ·
848001379d12e45b - ModelBase helper · level 0 · unverified ·
6dd3fd72476b96f4 - RecurrentDecoder model · level 0 · unverified ·
d939293e6a574006 - RecurrentEncoder model · level 1 · ran ·
1e5ecdd2d7834bc5 - StandardDecoder model · level 1 · ran ·
ac79741037ef4e2b - Embeddings model · level 1 · ran ·
50973ad904b644b9 - generate_signal helper · level 1 · ran · honoured contract ·
58801803f9aacb41 - clones helper · level 1 · ran · our draft was wrong ·
7b4317c9bb6b15bc - PositionalEncoding helper · level 1 · ran ·
55f808c7c4cab374
Each brick ends with its code_sha256, Syntology's identity for that exact code; click one to copy the get_code call that returns it. For agents, on Syntology's MCP service: assembly(arxiv_id="1807.03819") returns these assemblies with each brick's role and wiring (how to connect).
Tasks
Results from the paper archive 2025-07-28
| Task | Dataset | Model | Metric | Value | Rank at snapshot | Leaderboard | Report |
|---|---|---|---|---|---|---|---|
| Language Modelling | LAMBADA | Universal Transformer (w/ dynamic halting) | Accuracy | 56.25 | #30 of 37 | Archive leaderboard | report |
| Machine Translation | WMT2014 English-German | universal transformer base | BLEU score | 28.9 | #37 of 91 | Archive leaderboard | report |
Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.
Methods
Introduced by this paper: Universal Transformer
Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections