CounterpediaExplore knowledge. Follow the sources.
Research entry · Not admitted

Attention Is All You Need

Built from a research dossier; admission and independent checking of this account are not supplied in this view. Source attribution does not establish independent corroboration. Read the full dossier · Inspect

Attention Is All You Need is a 2017 paper by eight authors that proposed the Transformer, a sequence architecture without recurrence or convolution, using attention and feed-forward layers. Its inspected arXiv versions report translation and English constituency-parsing experiments. The proceedings PDF contains the translation results but does not contain that parsing experiment. Seven selected arXiv PDFs have different byte digests; that does not establish seven distinct scientific revisions.

The reported translation scores depend on the representation and passage. For English-to-French, arXiv v7’s abstract and Table 2 report 41.8 BLEU while its results prose reports 41.0. The proceedings web abstract reports 41.1, and its PDF reports 41.0. Earlier versions contain further internal differences. A citation therefore needs the edition, representation, passage and task or model conditions, preserving disagreement without choosing an unsupported winning number. Public reviews record individual judgments; later achievements belong with the works that report them.

Sources behind this account

All 8 source candidates →

Showing 4 of 8; the full list follows the account.

Carried research candidates; listing does not admit a Source or establish support.

Contents and source readers

8 research source candidates. Only recorded, eligible public Source destinations appear above.

Scope

The 2017 paper by Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Łukasz Kaiser and Illia Polosukhin, in its arXiv editions (1706.03762, v1 12 June 2017 through v7 2 August 2023) and its NeurIPS 2017 proceedings edition (Advances in Neural Information Processing Systems 30, paper ID 3058), together with the public NeurIPS reviews; the specific quantitative and architectural claims each edition makes (an architecture without recurrence or convolution, built from attention and feed-forward layers); and the gap between those claims and what the paper is commonly cited as having shown.

This dossier is not a history of the Transformer, a survey of later models, an account of the authors' later careers, or a judgment about which edition is "correct".

Key distinctions

  1. "The paper" names several editions and representations. All seven arXiv PDFs were acquired, their embedded v1–v7 labels read, and their SHA-256 digests found to differ. The proceedings PDF and its web abstract are separate representations. A version label or rounded listing size is not a digest; the actual byte comparison is recorded under S02. Different binary hashes do not, by themselves, establish different scientific claims: normalized extracted v6/v7 text differed only in the embedded version/date in this inspection.
  2. "The Transformer got 41.8 BLEU on English-to-French" needs a representation and passage. In arXiv v7, the abstract and Table 2 report 41.8, while §6.1 prose reports 41.0. The proceedings HTML abstract reports 41.1; its PDF reports 41.0. Earlier PDFs add further differences: v2 has 41.2 in the abstract, 41.0 in Table 2 and 41.17 in §6.1; v3 has 41.0, 41.0 and 41.17. The complete selected-PDF readings are in S02. These are attributed statements, not an adjudication that every number represents an identical experimental configuration.
  3. "28.4 BLEU on English-to-German" is also edition-dependent. The proceedings abstract page reports 27.5; arXiv v7 reports 28.4.
  4. arXiv is not the venue. The paper was refereed at NeurIPS 2017; its reviews are public. The arXiv record is the version history, not the peer-reviewed edition.
Show 6 more
  1. "Introduced attention" is wrong. The paper's background cites earlier attention mechanisms, including Bahdanau et al. (S07). Its architectural claim removes recurrence and convolution from the sequence model while retaining attention and feed-forward layers; the title does not erase those other components.
  2. "Proved attention is all you need" is a title, not a result, and "attention only" is not the architecture. The Transformer removes recurrence and convolution from the sequence model; it retains feed-forward layers, residual connections, layer normalization, embeddings and positional encodings alongside attention. The inspected arXiv versions contain two translation tasks and an English constituency-parsing task; the inspected proceedings PDF does not contain that parsing experiment. Its own text leaves restricted attention for very long sequences and other modalities to future work (PC-02 AIAYN-C06, C07).
  3. "The paper introduced GPT / BERT / LLMs" is anachronism. Those are later works that used the architecture. The paper claims translation quality, parallelizability and training cost.
  4. "Eight Google authors" is approximately right and the footnote is more careful. The paper lists equal contribution with random order and per-author contributions; two authors are marked as having done the work while at Google (one University of Toronto affiliation is listed). Read the footnote in the captured edition rather than paraphrasing it.
  5. "Most-cited paper of the century" is a dated, source-specific ranking. Citation counts and rank lists change by database and date; any such sentence needs its source and date (see S08, OPTIONAL).
  6. Training cost is version- and configuration-specific. The v7 abstract associates 3.5 days on eight GPUs with the big English-to-French model; the PDF specifies P100 hardware. The v2 abstract instead says 4.5 days. Preserve the edition and experiment rather than describing one training-time number as a property of Transformers (S02).

Analysis

A citation to the paper does not select one result. In the retained arXiv v7 PDF, the abstract and Table 2 report 41.8 English-to-French BLEU while §6.1 prose reports 41.0. The proceedings HTML abstract and its PDF also differ. The reader can identify which passage supports a precise attribution without deciding that all passages describe identical experimental configurations or choosing an unsupported corrected value.

The inspected representations also delimit the experimental account. All seven retained arXiv PDFs include a constituency-parsing experiment; the inspected proceedings PDF does not. This establishes a difference between the selected documents, not what every reviewer saw or which edition a later citer used. Distinct binary hashes establish distinct selected byte objects, not seven different scientific findings.

The paper’s architectural proposal, earlier attention work, individual reviewer judgments and later engineering work remain separate. Earlier work is cited by the paper; the public reviews contain both favorable assessments and questions. A reviewer’s statement about reimplementation is not an independently verified reproduction record. Later success on other tasks needs the works that report it, rather than being attributed retrospectively to the 2017 experiments.

Temporal bounds

The selected arXiv editions range from v1 on 12 June 2017 through v7 on 2 August 2023. The proceedings are NeurIPS/NIPS 2017. The seven PDF versions, v1/v7 listings, proceedings PDF/HTML and public review page were retained and read on 6 November 2026; exact passage comparisons were additionally recorded on 7 November. This 11 November candidate makes no new source-retrieval observation.

Version dates and retrieval dates answer different questions. The proceedings HTML reading is a 2026 observation of a proceedings page, not proof of its publication-day 2017 contents. The original unsupported historical-HTML inference remains withdrawn. A dated v5 PDF cannot supply a missing historical HTML snapshot.

The separate PC-02 design dossier remains research_artifact_only and is not superseded or assigned new claim identity here. This publication candidate proposes a distinct Wave 2 dossier target, pending owner choice of route and disposition. Existing Stage AIAYN articles and their architectural HTML bindings remain separate. Later works and optional citation rankings retain their own dates; no current citation ranking or universal scaling conclusion is asserted.

Questions in the research record

Questions the research dossier keeps separate, each with the positions and limits it records for the institutions involved. This is research analysis, not admitted findings.

Different English-to-French BLEU statements across and within representations

The retained v7 listing, PDF abstract and PDF Table 2 report 41.8; the same PDF's §6.1 prose reports 41.0. The proceedings HTML reports 41.1; its PDF abstract, Table 2 and §6.1 report 41.0. Earlier PDFs show additional differences, including v2's 41.2/41.0/41.17 and v3's 41.0/41.0/41.17 at those three locations (S02). v5 is the first of the seven inspected PDFs whose abstract and table report 41.8. That is not proof of when a historical HTML page changed. Each statement carries its representation, passage and model/task context; the reason for a difference and which result should prevail are not established by reading the numbers.

28.4 vs 27.5 (English-to-German BLEU)

S01: 28.4. S03 abstract page: 27.5 with a 165M-parameter single model. Same qualification.

"Introduced attention" vs "based solely on attention"

S07 predates the paper and is cited by it. The paper's novelty claim is architectural (no recurrence or convolution), not the invention of attention.

Reviewer caveats vs later reputation

S04 records individual reviewers' open points (long sequences, hyperparameter sensitivity, the long-range-dependency argument) beside favorable judgments. Later reputation treats the architecture as general; the paper deferred that, and its reviewers asked for it.

Proceedings PDF vs proceedings web abstract

Two objects at one venue carry different abstract text (confirmed: page 27.5 and 41.1; PDF 28.4 and 41.0). "The NeurIPS version" is itself two surfaces; a citation that names the venue without the surface is incomplete.

Title vs claim

"All you need" is a title. The inspected arXiv PDFs report two translation tasks and one English constituency-parsing task; the inspected proceedings PDF reports the translation experiments and does not contain that parsing experiment (S02/S03). Keep the experiment attached to the representation that includes it.

Training cost vs generality

In v7, the 3.5-day statement attaches to the big English-to-French model and the PDF names P100 hardware; v2's abstract instead reports 4.5 days. Quote the selected experiment, hardware and edition. Neither value is a general property of the architecture (S02).

Questions and evidence limits

Provisional claim analyses from the research dossier: a proposed wording, its candidate sources, and what those sources can and cannot establish. They are research interpretation, not admitted claims, findings or decisions. Where the dossier writes “Refusal branch” or “Reopening condition”, it proposes when a broader formulation should be withheld or reconsidered; no refusal or standing decision has been made.

Edition chronology and representation identity

Provisional reader wording

The selected arXiv record lists seven versions of Attention Is All You Need, from 12 June 2017 through 2 August 2023; the work also has a NeurIPS 2017 proceedings edition with separate HTML and PDF representations.

Candidate research source cards, not new bindings

What this analysis can establish

The retained version listings and embedded PDF labels identify the selected edition dates. All seven PDFs have different SHA-256 digests; the proceedings PDF and HTML are separately identified bytes. The authorship set is the eight authors named in Part A.

What it cannot establish / reason to qualify a stronger claim

Different hashes do not establish seven substantive scientific revisions. The recorded normalized v6/v7 text comparison found only their embedded version/date changed at that comparison scope. No publication-day state of the later-retrieved HTML or edition used by a later citer is inferred.

Standing and next check

Keep the exact selected representation and date role. No PC-02 or Stage identity is transferred to these local candidates. No governed Claim, SourceEdition, Claim-to-SourceEdition relationship, Check outcome or standing is created by this provisional analysis.

BLEU values by passage and representation

Provisional reader wording

In arXiv v7, the English-to-French abstract and Table 2 report 41.8 BLEU while §6.1 prose reports 41.0; the proceedings HTML abstract reports 41.1 while its PDF abstract, table and prose report 41.0.

Candidate research source cards, not new bindings

What this analysis can establish

For all seven selected arXiv PDFs, the recorded comparison uses abstract physical p. 1 and Table 2’s Transformer (big) row/§6.1 prose on p. 8. In abstract/table/prose order: v1 41.0/41.0/41.0; v2 41.2/41.0/41.17; v3 41.0/41.0/41.17; v4 41.0/41.0/41.0; v5–v7 41.8/41.8/41.0. The proceedings PDF is 41.0/41.0/41.0 at those locations; its separately retained HTML abstract is 41.1. These are attributed WMT 2014 English-to-French passage readings.

What it cannot establish / reason to qualify a stronger claim

The values alone do not establish identical experiment/model conditions, an author-issued correction or which value should prevail. English-to-German is a different task: the PDF abstracts report 28.4 while the proceedings HTML reports 27.5. A model description in one passage is not silently transferred to every other passage.

Standing and next check

The selected local comparison is complete at its stated scope. Preserve representation, internal location and task/model qualifiers; leave corrected-result and experimental-equivalence inferences unresolved. No governed Claim, SourceEdition, Claim-to-SourceEdition relationship, Check outcome or standing is created by this provisional analysis.

Individual public reviews

Provisional reader wording

The official proceedings review page contains three reviewer texts; Reviewer 1 reports reimplementation and calls the paper a clear accept, while the reviews also record questions and caveats.

Candidate research source cards, not new bindings

What this analysis can establish

The retained public review page supports attribution of a particular reviewer’s judgment and stated activity. The existing reading records learning-rate sensitivity and other questions about the architecture, evaluation and presentation alongside favorable assessments.

What it cannot establish / reason to qualify a stronger claim

Three reviews are not a collective certification of every claim. A reviewer’s statement about reimplementation is not a separately inspected independent-reproduction record, and the review page does not establish that every reviewer saw the exact selected proceedings PDF.

Standing and next check

Retain each speaker’s attribution. Do not promote reported reimplementation into an independently verified result. No governed Claim, SourceEdition, Claim-to-SourceEdition relationship, Check outcome or standing is created by this provisional analysis.

Architectural and experimental scope

Provisional reader wording

The arXiv PDFs describe a sequence architecture without recurrence or convolution while retaining attention, feed-forward layers and other components; their selected experiments include WMT 2014 translation and English constituency parsing.

Candidate research source cards, not new bindings

What this analysis can establish

The inspected arXiv versions carry the parsing experiment (v1 physical p. 8; v2–v7 physical p. 9), while the inspected proceedings PDF does not contain it. The paper’s own discussion defers restricted attention for very long sequences to future work. The earlier Bahdanau preprint is separately retained prior work cited by the paper.

What it cannot establish / reason to qualify a stronger claim

The title does not mean attention is the only component or that the paper invented attention. The experiments do not establish performance on every task or modality. Absence of parsing in the inspected proceedings PDF is a representation finding, not an account of every document in the review process.

Standing and next check

Preserve the distinction between prior work, the selected paper’s assertions and demonstrated tasks, and later claims. PC-02 references remain navigation, not claim identity or standing. No governed Claim, SourceEdition, Claim-to-SourceEdition relationship, Check outcome or standing is created by this provisional analysis.

Later reputation and unsupported generalization

Provisional reader wording

“Attention Is All You Need proved that Transformers are all you need for AI” goes beyond what the selected paper representations and public reviews establish.

Candidate research source cards, not new bindings

What this analysis can establish

The selected record permits a bounded account of the proposed architecture, reported translation/parsing tasks at their actual representations, and reviewer judgments. The selected FlashAttention v1 preprint is an example of separately dated later work on attention’s memory cost; its findings belong to that work.

What it cannot establish / reason to qualify a stronger claim

The compressed formulation is an editorial question, not a sourced universal claim. Neither the title nor later reputation proves general scaling, every later modality or present citation rankings. The selected later preprint is not automatically its subsequent proceedings edition.

Standing and next check

Attach later results to their own works and editions. Omit unsourced popularity/ranking claims and any universal experimental conclusion from the proposed lead. No governed Claim, SourceEdition, Claim-to-SourceEdition relationship, Check outcome or standing is created by this provisional analysis.

Why not?

Research account dated 2026-11-06. Descriptions of current or pending states refer to the recorded research date.

BEST "WHY NOT?" SPECIMEN

Popular formulation

"The Transformer paper reported 41.8 BLEU on English-to-French."

Why not

Qualify. The v7 abstract and Table 2 do report 41.8, as does its listing; that documentary statement is not made false by a different number elsewhere. Its §6.1 prose reports 41.0. The proceedings HTML reports 41.1, and its PDF abstract, table and prose report 41.0 (S01–S03). State which passage is being cited and preserve the disagreement rather than asserting one unqualified score for every representation.

So: the paper reports X needs which edition, which representation, which internal passage, and which task/model conditions.

Reopening condition

The local read-and-hash comparison is complete for the selected seven arXiv PDFs and the proceedings PDF/HTML; its readings are in S02/S03. A precise sentence can now identify the passage that reports 41.8 and acknowledge the internal disagreement. A stronger claim that it is the corrected result, or that another value describes an identical experiment, needs additional author-issued or experimental evidence. Governed acquisition and any public display authorization remain separate from this local comparison.

Origin of this passage

Research dossier · E. BEST "WHY NOT?" SPECIMEN · lines 414–429

Artifact SHA-256: 3de7c01814508106566e47543326a72240b4703fc169b5d2f39a4b12b8b570d0

Attributed research text; this locator does not establish source support or standing.

SECOND "WHY NOT?" — WHAT WAS ESTABLISHED

Popular formulation

"The paper introduced attention."

Why not

The paper cites earlier attention work (S07); its proposed sequence architecture removes recurrence and convolution while retaining attention and feed-forward layers. Reviewer 1 regarded the combination, rather than striking novelty in each underlying technique, as the achievement (S04). That is the reviewer's judgment, not an independent reproduction result.

Counterpedia reveal The paper's own background section and its reviewers already draw the line the popular sentence erases.

Origin of this passage

Research dossier · F. SECOND "WHY NOT?" — WHAT WAS ESTABLISHED · lines 435–444

Artifact SHA-256: 3de7c01814508106566e47543326a72240b4703fc169b5d2f39a4b12b8b570d0

Attributed research text; this locator does not establish source support or standing.

THIRD "WHY NOT?" — LATER SUCCESS

Popular formulation

"The 2017 paper showed Transformers scale to everything."

Why not

The inspected arXiv PDFs report two translation directions and an English constituency-parsing task, and defer restricted attention for very long sequences and other modalities to future work (S02; PC-02 AIAYN-C06, C07). The inspected proceedings PDF does not contain the parsing experiment. Later claims about scaling or other modalities require their own works; S06 is one selected example of later work on attention's memory cost, not evidence for every later generalization. Attribute each result to the work and representation that reports it.

Origin of this passage

Research dossier · G. THIRD "WHY NOT?" — LATER SUCCESS · lines 450–456

Artifact SHA-256: 3de7c01814508106566e47543326a72240b4703fc169b5d2f39a4b12b8b570d0

Attributed research text; this locator does not establish source support or standing.

Sources

8 documentary source candidates · show the source list

Sources

8 research source candidates are carried by this entry, as authored in its research dossier. Each is a research candidate, not an admitted Counterpedia Source.

The dossier records an issuer for 5 of 8 and a date for 0; where it labels no issuer, the card says so. Addresses are shown as recorded. A recorded address is not a Counterpedia capture and does not establish retrieval, authenticity or support for any claim. A card links to a Counterpedia source page only where an accepted record relates a capture of an address recorded for that document to that public source; listing a candidate implies no other association.

Sources by recorded issuer

Select a candidate below to mark it for reading. Selection never changes the entry or its conclusions.

All 8 sources selected.

  1. arXiv abstract record and submission history (PC-02 PAPER-S01)

    Issuer
    arXiv (cs.CL; cs.LG)

    Open: Landing

    More about this source · 1 recorded address
    Role of this document
    repository metadata for version labels and submission dates, not a byte digest or peer-review record. Exact retained PDF bytes are identified separately under S02.
    Exact title
    Attention Is All You Need
    Landing
    https://arxiv.org/abs/1706.03762
    Technical details
    Research handle in the dossier
    AIAYN-S01
    Research notes
    Dossier notes on this source (opens in a new tab)
    Counterpedia source link
    None: no capture in Counterpedia’s index of captures with accepted source relationships has the exact address recorded for this document.

  2. Exact arXiv PDF editions (PC-02 PAPER-S02)

    Issuer
    Not labelled in the dossier.

    Open: Exact PDF locators

    More about this source · 2 recorded addresses
    Role of this document
    author-posted PDF editions. All seven were retained, parsed and checked against their embedded version labels on 2026-11-06. Their distinct hashes establish different selected byte objects, not seven different scientific results or accepted governed identities.
    Exact title
    Attention Is All You Need, per version
    Exact PDF locators
    https://arxiv.org/pdf/1706.03762v1
    Exact PDF locators
    throughhttps://arxiv.org/pdf/1706.03762v7
    Technical details
    Research handle in the dossier
    AIAYN-S02
    Research notes
    Dossier notes on this source (opens in a new tab)
    Counterpedia source link
    None: no capture in Counterpedia’s index of captures with accepted source relationships has the exact addresses recorded for this document.

  3. NeurIPS 2017 proceedings record (PC-02 PAPER-S03)

    Issuer
    Advances in Neural Information Processing Systems 30 (NIPS 2017), Curran Associates

    Open: Landing inspected

    More about this source · 2 recorded addresses
    Role of this document
    official proceedings representations of the refereed work. The acquired HTML abstract reports 27.5/41.1; the acquired 11-page PDF reports 28.4/41.0 in its abstract and Table 2, and 41.0 in §6.1 prose. Unlike all seven inspected arXiv PDFs, this proceedings PDF does not contain the parsing experiment. That absence alone establishes nothing about every document seen by reviewers. Cite the representation and passage.
    Exact title
    Attention is All you Need
    Landing inspected
    https://papers.nips.cc/paper_files/paper/2017/hash/3f5ee243547dee91fbd053c1c4a845aa-Abstract.html
    Exact PDF inspected
    https://papers.nips.cc/paper/2017/file/3f5ee243547dee91fbd053c1c4a845aa-Paper.pdf
    Technical details
    Research handle in the dossier
    AIAYN-S03
    Research notes
    Dossier notes on this source (opens in a new tab)
    Counterpedia source link
    None: no capture in Counterpedia’s index of captures with accepted source relationships has the exact addresses recorded for this document.

  4. NeurIPS 2017 public reviews (PC-02 PAPER-S04)

    Issuer
    NeurIPS proceedings (public review record)

    Open: Landing inspected

    More about this source · 1 recorded address
    Role of this document
    three reviewer texts published with the proceedings; each is one reviewer's judgment and request, not a collective certification of the paper's claims. Reviewer 1 reports reimplementing the work and calls it "a clear accept" while noting the learning-rate schedule "seems to really matter" (a reviewer's statement about reimplementation, not a separate reproduction record); Reviewer 2 notes the architecture advances state of the art by about 1 BLEU and asks about dropout and a dangling reference; a reviewer notes the lack of mathematical definition of multi-head attention and asks for evaluation on long sentences and for stronger validation of the long-range-dependency argument.
    Exact title
    Reviews for paper ID 3058, Attention is All you Need
    Landing inspected
    https://papers.neurips.cc/paper_files/paper/2017/file/3f5ee243547dee91fbd053c1c4a845aa-Reviews.html
    Technical details
    Research handle in the dossier
    AIAYN-S04
    Research notes
    Dossier notes on this source (opens in a new tab)
    Counterpedia source link
    None: no capture in Counterpedia’s index of captures with accepted source relationships has the exact address recorded for this document.

  5. Google Research publication page

    Issuer
    Google Research (publications)

    Open: Landing

    More about this source · 1 recorded address
    Role of this document
    institutional republication page carrying an abstract (reads 28.4 BLEU English-to-German and 41.0 English-to-French as retrieved on 2026-11-06, which matches the proceedings PDF abstract, not the arXiv v7 listing).
    Exact title
    Attention is All You Need
    Landing
    https://research.google/pubs/pub46201/
    Technical details
    Research handle in the dossier
    AIAYN-S05
    Research notes
    Dossier notes on this source (opens in a new tab)
    Counterpedia source link
    None: no capture in Counterpedia’s index of captures with accepted source relationships has the exact address recorded for this document.

  6. FlashAttention (PC-02 PAPER-S05)

    Issuer
    arXiv

    Open: Landing inspected

    More about this source · 2 recorded addresses
    Role of this document
    later engineering work on attention's memory cost; included as an example of what later work added that readers fold back into the 2017 paper.
    Exact title
    FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness
    Authors
    Tri Dao et al.
    Landing inspected
    https://arxiv.org/abs/2205.14135v1
    Exact PDF inspected
    https://arxiv.org/pdf/2205.14135v1
    Technical details
    Research handle in the dossier
    AIAYN-S06
    Research notes
    Dossier notes on this source (opens in a new tab)
    Counterpedia source link
    None: no capture in Counterpedia’s index of captures with accepted source relationships has the exact addresses recorded for this document.

  7. Bahdanau, Cho, Bengio (2014) attention mechanism

    Issuer
    Not labelled in the dossier.

    Open: Landing inspected

    More about this source · 2 recorded addresses
    Role of this document
    prior work the paper cites; evidence against the "introduced attention" reading.
    Exact title
    Neural Machine Translation by Jointly Learning to Align and Translate
    Landing inspected
    https://arxiv.org/abs/1409.0473v1
    Exact PDF inspected
    https://arxiv.org/pdf/1409.0473v1
    Technical details
    Research handle in the dossier
    AIAYN-S07
    Research notes
    Dossier notes on this source (opens in a new tab)
    Counterpedia source link
    None: no capture in Counterpedia’s index of captures with accepted source relationships has the exact addresses recorded for this document.

  8. Citation-count and ranking surfaces — OPTIONAL

    Issuer
    Not labelled in the dossier.

    Open: Landing

    More about this source · 1 recorded address
    Role of this document
    tertiary and database-dependent; usable only with source and date attached.
    Landing
    https://en.wikipedia.org/wiki/Attention_Is_All_You_Need
    Technical details
    Research handle in the dossier
    AIAYN-S08
    Research notes
    Dossier notes on this source (opens in a new tab)
    Counterpedia source link
    None: no capture in Counterpedia’s index of captures with accepted source relationships has the exact address recorded for this document.
Work with these sources

Explore a question

Start with a question, read the sources, and develop a working answer.

Your notes stay on this page until you leave or reload it. Download an editable copy to continue later. They are not saved to your account.

What would you like to understand?

2. Read with your question in mind

8 of 8 sources selected. Open the sources and look for the passages that address each part of your question.

Choose and read sources

Source pages open in a new tab so your question and notes stay here.

Compare selected sources (8)
Recorded source details. Missing fields are not inferred; selection does not establish support for a claim. Scroll across to compare sources.
DetailarXiv abstract record and submission history (PC-02 PAPER-S01)Exact arXiv PDF editions (PC-02 PAPER-S02)NeurIPS 2017 proceedings record (PC-02 PAPER-S03)NeurIPS 2017 public reviews (PC-02 PAPER-S04)Google Research publication pageFlashAttention (PC-02 PAPER-S05)Bahdanau, Cho, Bengio (2014) attention mechanismCitation-count and ranking surfaces — OPTIONAL
Recorded title
  • Exact titleAttention Is All You Need
  • Exact titleAttention Is All You Need, per version
  • Exact titleAttention is All you Need
  • Exact titleReviews for paper ID 3058, Attention is All you Need
  • Exact titleAttention is All You Need
  • Exact titleFlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness
  • Exact titleNeural Machine Translation by Jointly Learning to Align and Translate
Not supplied in this view.
Issuer / publisher
  • IssuerarXiv (cs.CL; cs.LG)
Not supplied in this view.
  • IssuerAdvances in Neural Information Processing Systems 30 (NIPS 2017), Curran Associates
  • IssuerNeurIPS proceedings (public review record)
  • IssuerGoogle Research (publications)
  • IssuerarXiv
Not supplied in this view.Not supplied in this view.
Documentary role
  • Authority / posturerepository metadata for version labels and submission dates, not a byte digest or peer-review record. Exact retained PDF bytes are identified separately under S02.
  • Authority / postureauthor-posted PDF editions. All seven were retained, parsed and checked against their embedded version labels on 2026-11-06. Their distinct hashes establish different selected byte objects, not seven different scientific results or accepted governed identities.
  • Authority / postureofficial proceedings representations of the refereed work. The acquired HTML abstract reports 27.5/41.1; the acquired 11-page PDF reports 28.4/41.0 in its abstract and Table 2, and 41.0 in §6.1 prose. Unlike all seven inspected arXiv PDFs, this proceedings PDF does not contain the parsing experiment. That absence alone establishes nothing about every document seen by reviewers. Cite the representation and passage.
  • Authority / posturethree reviewer texts published with the proceedings; each is one reviewer's judgment and request, not a collective certification of the paper's claims. Reviewer 1 reports reimplementing the work and calls it "a clear accept" while noting the learning-rate schedule "seems to really matter" (a reviewer's statement about reimplementation, not a separate reproduction record); Reviewer 2 notes the architecture advances state of the art by about 1 BLEU and asks about dropout and a dangling reference; a reviewer notes the lack of mathematical definition of multi-head attention and asks for evaluation on long sentences and for stronger validation of the long-range-dependency argument.
  • Authority / postureinstitutional republication page carrying an abstract (reads 28.4 BLEU English-to-German and 41.0 English-to-French as retrieved on 2026-11-06, which matches the proceedings PDF abstract, not the arXiv v7 listing).
  • Authority / posturelater engineering work on attention's memory cost; included as an example of what later work added that readers fold back into the 2017 paper.
  • Authority / postureprior work the paper cites; evidence against the "introduced attention" reading.
  • Authority / posturetertiary and database-dependent; usable only with source and date attached.
Recorded datesNot supplied in this view.Not supplied in this view.Not supplied in this view.Not supplied in this view.Not supplied in this view.Not supplied in this view.Not supplied in this view.Not supplied in this view.
Recorded identifiersNot supplied in this view.Not supplied in this view.Not supplied in this view.Not supplied in this view.Not supplied in this view.
  • AuthorsTri Dao et al.
Not supplied in this view.Not supplied in this view.
Recorded locators
Source status
  • research candidate
  • research candidate
  • research candidate
  • research candidate
  • research candidate
  • research candidate
  • research candidate
  • research candidate

The comparison shows recorded source details. It does not evaluate whether the sources answer your question.

Write what you learned. Name the source and passage behind each point, and distinguish what it says from your interpretation.

Which parts of your question remain unanswered? Record disagreements, missing material and your next step.

Keep your work and return to it

Downloads your question, notes and source selection as an editable file (.json). Reopen it on this entry to continue. It contains no source document files.

Reopening a file replaces the question, notes and selection on this page. Download your current work first if you want to keep both.

Plain text with the current selected source details and your notes. It contains no source document files and is not saved to your account.

Return to your question and notes

All 8 sources selected.

Return to the entry

Record provenance
research_artifact
{"pack_id":"demo-corpus-wave2","artifact_slug":"attention-is-all-you-need-demo-corpus-dossier-v0-3","artifact_sha256":"3de7c01814508106566e47543326a72240b4703fc169b5d2f39a4b12b8b570d0","source_href":"/research/demo-corpus-wave2/attention-is-all-you-need-demo-corpus-dossier-v0-3","source_kind":"dossier"}
Projection axes
Postureresearch
Source kindresearch_template
Source surfacenot supplied

This renderer preserves upstream distinctions; it does not mint identity, support, authenticity, admission, or truth.

INSPECTCounterpedia Check · this read modelInspect only the signals this renderer actually received
Record identifier
research:demo-corpus-wave2:attention-is-all-you-need-demo-corpus-dossier-v0-3
Posture
research
Source kind
research_template
Lifecycle
not recorded
Source footprint
not supplied to this presentation
Capture associations
not evaluated here
Unresolved capture identity
not evaluated here
Truth
not a finding of this renderer
Authenticity
not inferred from source or capture presence
Research profile
Demo Corpus
Artifact kind
dossier
Source candidates
8
CHECK is inspection, not a truth badge. Source presence ≠ support; capture association ≠ authenticity; rendering ≠ admission.