Papers › Train Short, Test Long: Attention with Linear Biases Enables Input Length Extrapolation

Train Short, Test Long: Attention with Linear Biases Enables Input Length Extrapolation

27 Aug 2021ICLR 2022 4arXiv:2108.12409archive 2025-07-28

Ofir Press, Noah A. Smith, Mike Lewis

Since the introduction of the transformer model by Vaswani et al. (2017), a fundamental question has yet to be answered: how does a model achieve extrapolation at inference time for sequences that are longer than it saw during training? We first show that extrapolation can be enabled by simply changing the position representation method, though we find that current methods do not allow for efficient extrapolation. We therefore introduce a simpler and more efficient position method, Attention with Linear Biases (ALiBi). ALiBi does not add positional embeddings to word embeddings; instead, it biases query-key attention scores with a penalty that is proportional to their distance. We show that this method trains a 1.3 billion parameter model on input sequences of length 1024 that extrapolates to input sequences of length 2048, achieving the same perplexity as a sinusoidal position embedding model trained on inputs of length 2048 but training 11% faster and using 11% less memory. ALiBi's inductive bias towards recency also leads it to outperform multiple strong position methods on the WikiText-103 benchmark.

PaperPDFConference PDFCodeCode Syntology ranMethod as code

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):

Code

Syntology 16 code samples harvested from 5 repositories linked to this paper: 16 ran (of which 6 constructed an object rather than computing a result; 13 with no instrument failure: 4 honoured, 0 violated, 9 with no contract checked; 3 where Syntology's instrument failed) · 0 unverified. Where Syntology's instrument failed: 2 where our contract draft was wrong; 1 where the contract check could not run; failures of Syntology's instrument, not of the code.

By repository: official repository: 4 samples from 1 repository, 4 ran (3 with no instrument failure; 1 where Syntology's instrument failed); community (archive-listed): 12 samples from 4 repositories, 12 ran (10 with no instrument failure; 2 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:

Levels written between 2026-09-22 and 2026-09-25 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

ofirpress/attention_with_linear_biases officialmentioned in paperpytorchMITSyntology: 4 samples, 4 ran (1 instrument-failed) · 0 unverified report
answerdotai/modernbert mentioned on GitHubpytorchApache-2.0Syntology: 5 samples, 5 ran (2 instrument-failed) · 0 unverified report
baichuan-inc/baichuan-13b mentioned on GitHubpytorchApache-2.0Syntology: not harvested report
brkkaya/a_linear_attention_with_biases mentioned on GitHubpytorchSyntology: 1 sample, 1 ran (0 instrument-failed) · 0 unverified report
jaketae/alibi mentioned on GitHubpytorchSyntology: 4 samples, 4 ran (0 instrument-failed) · 0 unverified report
lucidrains/PaLM-jax mentioned on GitHubjaxSyntology: 2 samples, 2 ran (0 instrument-failed) · 0 unverified report
vinairesearch/phogpt mentioned on GitHubpytorchBSD-3-ClauseSyntology: not harvested report
MindCode-4/code-6 mindsporeSyntology: not harvested report
MindSpore-scientific/code-3 mindsporeSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Coneheads: Hierarchy Aware Attention (arXiv:2306.00392; that paper's own run record: 7 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 1 violated, 0 with no contract checked; 6 where Syntology's instrument failed) · 0 unverified); nothing checks that they implement this paper's method. report
labmlai/annotated_deep_learning_paper_implementations pytorchSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Playing Atari with Deep Reinforcement Learning (arXiv:1312.5602; that paper's own run record: 64 ran (of which 24 constructed an object rather than computing a result; 46 with no instrument failure: 5 honoured, 0 violated, 41 with no contract checked; 18 where Syntology's instrument failed) · 53 unverified), Adam: A Method for Stochastic Optimization (arXiv:1412.6980; that paper's own run record: 27 ran (of which 7 constructed an object rather than computing a result; 23 with no instrument failure: 4 honoured, 2 violated, 17 with no contract checked; 4 where Syntology's instrument failed) · 10 unverified), Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift (arXiv:1502.03167; that paper's own run record: 19 ran (of which 5 constructed an object rather than computing a result; 14 with no instrument failure: 0 honoured, 0 violated, 14 with no contract checked; 5 where Syntology's instrument failed) · 2 unverified) and 12 more (every one in this page's JSON twin); nothing checks that they implement this paper's method. report

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

16 samples harvested: 16 ran (of which 6 constructed an object rather than computing a result; 13 with no instrument failure: 4 honoured, 0 violated, 9 with no contract checked; 3 where Syntology's instrument failed) · 0 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page last changed ; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

4ran · honoured contract
2ran · our draft was wrong
1ran; the contract check could not run
9ran

Licence: 5 of the 16 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”. For 1 sample that ran, the graph records a signature this page does not show: for a class the fixture built without calling, the harness called the object with inputs it made up itself, so the recorded signature is the harness's probe, not the fixture's arguments. “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.

ofirpress/attention_with_linear_biases official repository 4 samples, 4 ran (1 instrument-failed) · 0 unverified
GatedLinear fairseq/modules/downsampled_multihead_attention.py ran MIT recorded; this copy not marked cleared · pointer only · ca6cbbd64821d7fa · report fixture: signature in_features:int/0/int, out_features:int/0/int, dropout:float/0/float, bias:bool/0/bool · returned Sequential · level recorded 2026-09-25
Linear in 127 papers fairseq/modules/downsampled_multihead_attention.py ran MIT recorded; this copy not marked cleared · pointer only · 77d359c7acb72d6a · report fixture: signature in_features:int/0/int, out_features:int/0/int, dropout:float/0/float, bias:bool/0/bool · returned Linear · level recorded 2026-09-25
augmented_memory fairseq/models/speech_to_text/modules/augmented_memory_attention.py ran MIT recorded; this copy not marked cleared · pointer only · 1769a89f798d0556 · report fixture: signature klass:callable/0/? · returned type · level recorded 2026-09-25
safe_readline in 67 papers fairseq/binarizer.py ran · our draft was wrong MIT recorded; this copy not marked cleared · pointer only · 82bdf940665aeeb4 · report fixture: signature f:StringIO/0/? · returned str · level recorded 2026-09-24
answerdotai/modernbert community (archive-listed) 5 samples, 5 ran (2 instrument-failed) · 0 unverified
BertAlibiUnpadSelfAttention src/bert_layers/attention.py ran · metamorphic tier: deterministic Apache-2.0 (permissive) · 201cc87ee9341bc1 · report fixture: returned BertAlibiUnpadSelfAttention · level recorded 2026-09-22
IndexFirstAxis src/bert_layers/attention.py ran Apache-2.0 (permissive) · 06c2a55caabae8b8 · report fixture: returned IndexFirstAxis · level recorded 2026-09-22
IndexPutFirstAxis src/bert_layers/attention.py ran Apache-2.0 (permissive) · 1760a9fa00a33980 · report fixture: returned IndexPutFirstAxis · level recorded 2026-09-22
pad_input in 16 papers src/bert_layers/attention.py ran; the contract check could not run (RAISES) Apache-2.0 (permissive) · 802fa8542c8445f9 · report fixture: signature hidden_states:Tensor/2/float, indices:Tensor/1/int, batch:int/0/int, seqlen:int/0/int · returned Tensor · level recorded 2026-09-22
unpad_input_only in 7 papers src/bert_layers/attention.py ran · our draft was wrong fingerprinted Apache-2.0 (permissive) · 4e534e11984c06b5 · report fixture: signature hidden_states:Tensor/3/float, attention_mask:Tensor/2/int · returned Tensor · level recorded 2026-09-22
brkkaya/a_linear_attention_with_biases community (archive-listed) 1 sample, 1 ran (0 instrument-failed) · 0 unverified
AliBiAttention models/alibi_attn.py ran · metamorphic tier: deterministic fingerprinted no licence file found · pointer only · 12ef66be25ef338d · report fixture: returned AliBiAttention · level recorded 2026-09-22
jaketae/alibi community (archive-listed) 4 samples, 4 ran (0 instrument-failed) · 0 unverified
ALiBiConfig alibi/attention.py ran MIT (permissive) · 678123248e9a2d53 · report fixture: returned ALiBiConfig · level recorded 2026-09-22
ALiBiMultiHeadAttention alibi/attention.py ran MIT (permissive) · 8099385f6961bb81 · report fixture: signature x:Tensor/3/float · returned Tensor · level recorded 2026-09-24
get_alibi_slope alibi/attention.py ran · honoured contract MIT (permissive) · aa08d318b01ec8ef · report fixture: signature num_heads:int/0/int · returned Tensor · level recorded 2026-09-22
get_relative_positions in 8 papers alibi/attention.py ran · honoured contract MIT (permissive) · 69e643285942e484 · report fixture: signature seq_len:int/0/int · returned Tensor · level recorded 2026-09-22
lucidrains/PaLM-jax community (archive-listed) 2 samples, 2 ran (0 instrument-failed) · 0 unverified
calc_alibi_bias palm_jax/palm_lite.py ran · honoured contract MIT (permissive) · 076a228889dcba7c · report fixture: signature seq_len:int/0/int, heads:int/0/int · returned ndarray · level recorded 2026-09-22
get_alibi_slopes palm_jax/palm_lite.py ran · honoured contract fingerprinted MIT (permissive) · 0e4286158fe03dff · report fixture: signature heads:int/0/int · returned list · level recorded 2026-09-22

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 Attention with Linear Biases (ALiBi) (harvested), lifted in the same batch on 2026-09-09: alternative readings kept side by side; none replaces or supersedes another.

harvested Assembly for Attention with Linear Biases (ALiBi) (graph method: Attention with Linear Biases (ALiBi)). 1 brick; individually on a fixture: 1 ran (of which 1 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 one sample that ran constructed an object rather than computing a result. The assembly has not been run end-to-end.

one of 4 siblings

  1. AliBiAttention entry · level 1 · ran · 12ef66be25ef338d

harvested Assembly for Attention with Linear Biases (ALiBi) (graph method: Attention with Linear Biases (ALiBi)). 5 bricks; individually on a fixture: 5 ran (of which 3 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 2 where Syntology's instrument failed) · 0 unverified. The assembly has not been run end-to-end.

one of 4 siblings

  1. BertAlibiUnpadSelfAttention entry · level 1 · ran · 201cc87ee9341bc1
  2. pad_input helper · level 1 · ran; the contract check could not run (RAISES) · 802fa8542c8445f9
  3. unpad_input_only helper · level 1 · ran · our draft was wrong · 4e534e11984c06b5
  4. IndexFirstAxis helper · level 1 · ran · 06c2a55caabae8b8
  5. IndexPutFirstAxis helper · level 1 · ran · 1760a9fa00a33980

harvested Assembly for Attention with Linear Biases (ALiBi) (graph method: Attention with Linear Biases (ALiBi)). 2 bricks; individually on a fixture: 2 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 2 honoured, 0 violated, 0 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

  1. calc_alibi_bias entry · level 1 · ran · honoured contract · 076a228889dcba7c
  2. get_alibi_slopes helper · level 1 · ran · honoured contract · 0e4286158fe03dff

harvested Assembly for Attention with Linear Biases (ALiBi) (graph method: Attention with Linear Biases (ALiBi)). 4 bricks; individually on a fixture: 4 ran (of which 1 constructed an object rather than computing a result; 4 with no instrument failure: 2 honoured, 0 violated, 2 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

  1. ALiBiMultiHeadAttention entry · level 1 · ran · 8099385f6961bb81
  2. ALiBiConfig helper · level 1 · ran · 678123248e9a2d53
  3. get_alibi_slope helper · level 1 · ran · honoured contract · aa08d318b01ec8ef
  4. get_relative_positions helper · level 1 · ran · honoured contract · 69e643285942e484

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="2108.12409") returns these assemblies with each brick's role and wiring (how to connect).

Tasks

Inductive BiasPlaying the Game of 2048Word Embeddings

1 archive task tag without a task page not shown.

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

Introduced by this paper: ALiBi

ALiBi

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