Papers › Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context
Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context
Zihang Dai, Zhilin Yang, Yiming Yang, Jaime Carbonell, Quoc V. Le, Ruslan Salakhutdinov
Transformers have a potential of learning longer-term dependency, but are limited by a fixed-length context in the setting of language modeling. We propose a novel neural architecture Transformer-XL that enables learning dependency beyond a fixed length without disrupting temporal coherence. It consists of a segment-level recurrence mechanism and a novel positional encoding scheme. Our method not only enables capturing longer-term dependency, but also resolves the context fragmentation problem. As a result, Transformer-XL learns dependency that is 80% longer than RNNs and 450% longer than vanilla Transformers, achieves better performance on both short and long sequences, and is up to 1,800+ times faster than vanilla Transformers during evaluation. Notably, we improve the state-of-the-art results of bpc/perplexity to 0.99 on enwiki8, 1.08 on text8, 18.3 on WikiText-103, 21.8 on One Billion Word, and 54.5 on Penn Treebank (without finetuning). When trained only on WikiText-103, Transformer-XL manages to generate reasonably coherent, novel text articles with thousands of tokens. Our code, pretrained models, and hyperparameters are available in both Tensorflow and PyTorch.
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="1901.02860")get_code(code_sha256="…")with a sample'scode_sha256from the list belowget_citation_path(paper_1="1901.02860", paper_2="…")with another paper's arXiv id or titleget_concepts_for_paper(arxiv_id="1901.02860")
Code
Syntology 143 code samples: 141 from 20 repositories linked to this paper, 2 identical to code harvested from another paper's repository. Across all 143: 67 ran (of which 37 constructed an object rather than computing a result; 49 with no instrument failure: 4 honoured, 1 violated, 44 with no contract checked; 18 where Syntology's instrument failed) · 76 unverified. Where Syntology's instrument failed: 6 where our contract draft was wrong; 12 where the contract check could not run; failures of Syntology's instrument, not of the code.
By repository: official repository: 13 samples from 1 repository, 8 ran (7 with no instrument failure; 1 where Syntology's instrument failed); community (archive-listed): 128 samples from 19 repositories, 57 ran (41 with no instrument failure; 16 where Syntology's instrument failed); 2 identical to code first harvested elsewhere. 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 7 of its repositories that Syntology harvested for other papers:
- google-research/meliad, community (archive-listed): harvested for another paper, listed under the repository
- huggingface/transformers, community (archive-listed): harvested for 4 other papers, listed under the repository
- samwisegamjeee/pytorch-transformers, community (archive-listed): harvested for another paper, listed under the repository
- PaddlePaddle/PaddleNLP, community (archive-listed): harvested for another paper, listed under the repository
- jiean001/models_m, community (archive-listed): harvested for another paper, listed under the repository
- labmlai/annotated_deep_learning_paper_implementations, community (archive-listed): harvested for 15 other papers, listed under the repository
- lvyufeng/bert4ms, 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.
37 repositories listed; official and paper-mentioned ones first.
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
143 samples harvested: 67 ran (of which 37 constructed an object rather than computing a result; 49 with no instrument failure: 4 honoured, 1 violated, 44 with no contract checked; 18 where Syntology's instrument failed) · 76 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: 43 of the 143 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 20 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. 2 samples are listed last, in a group without a repository. “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 10 samples 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. 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.
1ae137f0cee15e33 · report
fixture: returned AdaptiveEmbedding · level recorded 2026-09-22 e8aac29a19d0c2e7 · report
fixture: returned DecoderLayer · level recorded 2026-09-22 220657028509cfb1 · report
fixture: returned LogUniformSampler · level recorded 2026-09-22 d99c9cc9af9de483 · report
fixture: returned MultiHeadAttn · level recorded 2026-09-22 c96fba039a607703 · report
fixture: returned PositionwiseFF · level recorded 2026-09-22 338c39719307318d · report
fixture: returned ProjectedAdaptiveLogSoftmax · level recorded 2026-09-22 fe28ba867add05c0 · report
fixture: returned RelMultiHeadAttn · level recorded 2026-09-22 93228a3eb5c4d179 · report
fixture: signature embedding:callable/0/?, bias:Tensor/1/float, labels:Tensor/2/int, inputs:Tensor/3/float, sampler:LogUniformSampler/0/? · returned Tensor · level recorded 2026-09-22 9ef2fc89678fefb2 · report
17ee18e7e701439f · report
a033a134e9f6452f · report
89afd0bff30e18a9 · report
fb613dd5bc21371c · report
4c8b264bb79a35a1 · report
fixture: returned RelMultiHeadAttention · level recorded 2026-09-22 499bf5fee7915c78 · report
fixture: signature query_len:int/0/int, key_len:int/0/int, dtype:DType/0/? · returned EagerTensor · level recorded 2026-09-22 8d59a6d5900428f5 · report
fixture: signature x:EagerTensor/4/float · returned EagerTensor · level recorded 2026-09-22 census: 5 implementations from 6 papers, 4 distinct outputs on one shared input par_model.py ran · honoured contract fingerprinted no licence file found · pointer only ·
2398c4a5c285bdc4 · report
fixture: signature max_position:int/0/int, d_embed:int/0/int, min_freq:float/0/float · returned EagerTensor · level recorded 2026-09-22 f7a746d73ab2f28c · report
fixture: signature d_model:int/0/int, d_fnn:int/0/int, activation:str/0/str · returned Sequential · level recorded 2026-09-22 edeef97857b7cdca · report
fixture: signature q:EagerTensor/4/float, R:EagerTensor/4/float, direction:str/0/str · returned EagerTensor · level recorded 2026-09-22 df4647a6d9f03c0e · report
fixture: signature x:EagerTensor/4/float · returned EagerTensor · level recorded 2026-09-22 87202268910cb496 · report
de3ab66ec62fc167 · report
1be1a875d12d10b3 · report
534b6317f069c394 · report
9d6735f22035b700 · report
fixture: signature text:str/0/str · returned str · level recorded 2026-09-24 6b2a8f1c93d0b071 · report
fixture: signature gpu:int/0/int, ps_dev:str/0/str · returned function · level recorded 2026-09-25 census: 6 implementations from 6 papers, 3 distinct outputs across 3 buckets, one shared input per bucket tf/data_utils.py ran Apache-2.0 (permissive) ·
36eed0b265968be7 · report
fixture: signature data:ndarray/1/int, batch_size:int/0/int, num_passes:int/0/int · returned ndarray · level recorded 2026-09-26 0e3f1e158da4c3b8 · report
e505aeb6374f0f1c · report
48246428164ddbc5 · report
f07d3c5f67c9a93d · report
d5a05e8b3b59f012 · report
8ca3ba583d69ea38 · report
3d6b054785523852 · report
69598ced7806d8c3 · report
aa1de448b0528802 · report
773651748c9618d6 · report
fixture: signature curr_out:EagerTensor/2/float, prev_mem:EagerTensor/2/float, mem_len:int/0/int · returned EagerTensor · level recorded 2026-09-22 9109ba9791e44324 · report
fixture: signature lookup_table:EagerTensor/2/float, x:EagerTensor/2/int, use_tpu:bool/0/bool · returned EagerTensor · level recorded 2026-09-22 d50e8679cde3a927 · report
fixture: signature pos_seq:EagerTensor/1/float, inv_freq:EagerTensor/1/float, bsz:int/0/int · returned EagerTensor · level recorded 2026-09-22 d37883dc58ff9b69 · report
fixture: signature x:EagerTensor/4/float · returned EagerTensor · level recorded 2026-09-22 99e35dc770639097 · report
8d07f1308bd24cbc · report
e77541c1b4b15830 · report
fixture: returned AdaptiveEmbedding · level recorded 2026-09-22 census: 6 implementations from 6 papers, 4 distinct outputs across 3 buckets, one shared input per bucket pytorch_pretrained_bert/modeling_transfo_xl.py ran · metamorphic tier: deterministic Apache-2.0 (permissive) ·
06ce2b899569db5b · report
fixture: returned PositionalEmbedding · level recorded 2026-09-22 a97a22ac6fb31922 · report
fixture: returned TransfoXLConfig · level recorded 2026-09-22 35ef111fc00c21b4 · report
fixture: signature model:callable/0/?, config:Config/0/? · returned dict · level recorded 2026-09-22 fc52846dc775edff · report
d4857f57ecefa3d5 · report
62b1b4ab79fdfc8b · report
752529d23a4b3b22 · report
3dafbc50941af43d · report
dc7c9e91c34d6ddf · report
fd8d58e6a1428332 · report
8361fa24d3e06e13 · report
532a59e61c71b7a8 · report
d7bd7d30264b3658 · report
6fee57012e68e835 · report
b4bb49d2de702f67 · report
1ebc2bf2c3cddc4d · report
census: 13 implementations from 12 papers, 10 distinct outputs across 4 buckets, one shared input per bucket model.py ran licence not identified · pointer only ·
e3217ee8bd314909 · report
fixture: returned Attention · level recorded 2026-09-22 census: 5 implementations from 5 papers, 3 distinct outputs across 2 buckets, one shared input per bucket model.py ran licence not identified · pointer only ·
548da299b3bc04e9 · report
fixture: returned Block · level recorded 2026-09-22 72ccb5501088c2b4 · report
fixture: returned GPT2Model · level recorded 2026-09-22 census: 5 implementations from 5 papers, 5 distinct outputs across 3 buckets, one shared input per bucket model.py ran licence not identified · pointer only ·
39480331817170e8 · report
fixture: returned LayerNorm · level recorded 2026-09-22 df5a6cf54dd0aca6 · report
fixture: returned MLP · level recorded 2026-09-22 5fb8aacfbced509d · report
fixture: returned PathLSTM · level recorded 2026-09-22 9a2f71717dcf3962 · report
9d5e914b24b8911a · report
fixture: signature curr_out:EagerTensor/2/float, prev_mem:EagerTensor/2/float, mem_len:int/0/int, reuse_len:int/0/int · returned EagerTensor · level recorded 2026-09-22 9bf0f6e6298f3fbe · report
fixture: signature pos_seq:EagerTensor/1/float, inv_freq:EagerTensor/1/float, bsz:int/0/int · returned EagerTensor · level recorded 2026-09-22 5f7c5b38aeb29bf5 · report
fixture: signature x:EagerTensor/4/float, klen:int/0/int · returned EagerTensor · level recorded 2026-09-22 8e76562720b0c921 · report
fixture: signature qlen:int/0/int, klen:int/0/int, d_model:int/0/int, clamp_len:int/0/int, attn_type:str/0/str, bi_data:bool/0/bool, bsz:int/0/int, dtype:DType/0/? · returned EagerTensor · level recorded 2026-09-22 5abf0267f0fc7514 · report
7d9e9736e6c57d11 · report
fixture: returned ProjectedAdaptiveLogSoftmax · level recorded 2026-09-22 a7430322070cedb5 · report
709b8a145b15f0ca · report
fixture: returned GemmaConfig · level recorded 2026-09-22 39c5be8f7b1c0490 · report
fixture: returned GemmaRotaryEmbedding · level recorded 2026-09-22 census: 15 implementations from 60 papers, 2 distinct outputs on one shared input this sample: bucket 1 of 1, arg 1: rank 4, kind float, dtype float32 src/gemma.py ran; the contract check could not run (RAISES) fingerprinted no licence file found · pointer only ·
4e762613ee907cdb · report
fixture: signature hidden_states:Tensor/4/float, n_rep:int/0/int · returned Tensor · level recorded 2026-09-22 census: 24 implementations from 96 papers, 4 distinct outputs across 2 buckets, one shared input per bucket this sample: bucket 1 of 2, arg 1: rank 2, kind float, dtype float32 src/gemma.py ran · honoured contract fingerprinted no licence file found · pointer only ·
cec833a99feb528c · report
fixture: signature x:Tensor/2/float · returned Tensor · level recorded 2026-09-22 e4e12108c18eb032 · report
de07f82e734e3e2c · report
census: 14 implementations from 15 papers, 8 distinct outputs across 7 buckets, one shared input per bucket src/gemma.py unverified no licence file found · pointer only ·
5c61a4c30ce59547 · report
5597fd8e1219b866 · report
fixture: returned MultiHeadedAttention · level recorded 2026-09-22 4977fa89216a56fa · report
fixture: returned RelPositionMultiHeadedAttention · level recorded 2026-09-22 84585e9be4407fc4 · report
fixture: signature inputs:Tensor/2/float · returned Tensor · level recorded 2026-09-22 032b78da6979837c · report
fixture: signature q:Tensor/3/float, k:Tensor/3/float, v:Tensor/3/float, pos_embedding:Tensor/3/float, mask:NoneType/0/? · returned Tensor · level recorded 2026-09-22 census: 12 implementations from 12 papers, 1 distinct output on one shared input this sample: bucket 1 of 1, arg 1: rank 2, kind float, dtype float32 src/model/transformer_xl.py ran · metamorphic tier: invariant fingerprinted Apache-2.0 (permissive) ·
845f84e3a319753d · report
fixture: signature x:Tensor/2/float · returned Tensor · level recorded 2026-09-22 6a748d270690866b · report
fixture: returned GatoInputBase · level recorded 2026-09-22 ad41eff0feadfd69 · report
fixture: returned ICTaskInput · level recorded 2026-09-22 7eb33bec45fe09f3 · report
fixture: returned NLPTaskInput · level recorded 2026-09-22 b54068382071f8c8 · report
fixture: returned PatchEmbeddings · level recorded 2026-09-22 census: 6 implementations from 6 papers, 4 distinct outputs across 3 buckets, one shared input per bucket src/model/transformer_xl.py ran · metamorphic tier: deterministic Apache-2.0 (permissive) ·
142586d370dc7c1e · report
fixture: returned PositionalEmbedding · level recorded 2026-09-22 eed8c286b97219b7 · report
fixture: returned PositionwiseFF · level recorded 2026-09-22 2e57dbd19f63d3f4 · report
fixture: returned RLTaskInput · level recorded 2026-09-22 2a023fc4bac8d229 · report
fixture: returned VQATaskInput · level recorded 2026-09-22 2a5f4bac2414eb0f · report
fixture: returned VisionEmbedding · level recorded 2026-09-22 9ec384a6bfe40952 · report
fdc345d0da45dd95 · report
a2315dc06271216b · report
bf75653cc30018e0 · report
c018d465b6253b94 · report
fixture: signature query:Tensor/3/float, key:Tensor/3/float, value:Tensor/3/float, pos_embedding:Tensor/3/float, mask:Tensor/3/bool · returned Tensor · level recorded 2026-09-24 274ae09f88dbb42a · report
census: 13 implementations from 9 papers, 9 distinct outputs across 5 buckets, one shared input per bucket transformer/transformerxl.py ran · metamorphic tier: deterministic fingerprinted MIT (permissive) ·
07bf8cf6e731d1bc · report
fixture: returned PositionalEncoding · level recorded 2026-09-22 c7b5eb55f15a37de · report
fixture: returned TransformerEmbedding · level recorded 2026-09-22 91e5af7605c83b9a · report
8a8691f05cb1b5da · report
fixture: returned AdaptiveEmbedding · level recorded 2026-09-22 e672e8a20905b62b · report
fixture: returned MultiHeadAttn · level recorded 2026-09-22 4088ee884194d7c4 · report
fixture: returned ProjectedAdaptiveLogSoftmax · level recorded 2026-09-22 8fa37260188b5fd8 · report
fixture: returned RelMultiHeadAttn · level recorded 2026-09-22 c614edf5b8c0d3b0 · report
5ffa84d26ba39532 · report
ea7c8500cbc85ff5 · report
cb862dd19b1f381a · report
5014b04d3358c101 · report
5e1b71bde76b69e7 · report
582150577e5bd769 · report
6b820cc9aac66bc1 · report
c92e4fc990bba2fe · report
fixture: signature norm_type:str/0/str, dim:int/0/int · returned type · level recorded 2026-09-22 c1184d27672a7461 · report
fixture: signature in_channels:int/0/int, out_channels:int/0/int, activation:callable/0/?, norm_type:str/0/str, use_dropout:bool/0/bool, dropout_probability:float/0/float · returned Sequential · level recorded 2026-09-22 a4f0d0fb0e45662a · report
fixture: signature layers:list/1/callable · returned Sequential · level recorded 2026-09-22 census: 13 implementations from 12 papers, 10 distinct outputs across 4 buckets, one shared input per bucket ding/torch_utils/network/transformer.py unverified Apache-2.0 (permissive) ·
fe3454282cf0b88e · report
b28585d20fefea46 · report
2 samples are identical to code Syntology harvested from another paper's repository; this paper's own copy was not located, so no repository is named and no licence is recorded.
78d91efd74e4d12a · report
fixture: returned Tensor · 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.
18 assemblies on this paper: 18 harvested, 0 generated, listed in 1 groups (siblings lifted in one batch form one group).
18 sibling assemblies for Transformer-XL (harvested), lifted in the same batch on 2026-09-09: alternative readings kept side by side; none replaces or supersedes another. Bricks printed for 12 of 18 sibling assemblies; the other 6 are listed with their counts; every brick is in this page's JSON twin.
harvested Assembly for Transformer-XL (graph method: Transformer-XL). 15 bricks; individually on a fixture: 11 ran (of which 9 constructed an object rather than computing a result; 11 with no instrument failure: 0 honoured, 1 violated, 10 with no contract checked; 0 where Syntology's instrument failed) · 4 unverified. The assembly has not been run end-to-end.
one of 18 siblings
- TransformerXL entry · level 0 · unverified ·
a2315dc06271216b - GatoInputBase helper · level 1 · ran ·
6a748d270690866b - ICTaskInput helper · level 1 · ran ·
ad41eff0feadfd69 - NLPTaskInput helper · level 1 · ran ·
7eb33bec45fe09f3 - PositionalEmbedding model · level 1 · ran ·
142586d370dc7c1e - RLTaskInput helper · level 1 · ran ·
2e57dbd19f63d3f4 - RelPartialLearnableDecoderLayer model · level 0 · unverified ·
9ec384a6bfe40952 - VQATaskInput helper · level 1 · ran ·
2a023fc4bac8d229 - VisionEmbedding helper · level 1 · ran ·
2a5f4bac2414eb0f - print_with_rank helper · level 0 · unverified ·
bf75653cc30018e0 - PositionwiseFF model · level 1 · ran ·
eed8c286b97219b7 - RelPartialLearnableMultiHeadAttn model · level 0 · unverified ·
fdc345d0da45dd95 - PatchEmbeddings helper · level 1 · ran ·
b54068382071f8c8 - to_2tuple helper · level 1 · ran · violated contract ·
4e978f9f7835b10a - GEGLU helper · level 1 · ran ·
845f84e3a319753d
harvested Assembly for Transformer-XL (graph method: Transformer-XL). 23 bricks; individually on a fixture: 6 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; 3 where Syntology's instrument failed) · 17 unverified. The assembly has not been run end-to-end.
one of 18 siblings
- TransfoXLModel entry · level 0 · unverified ·
532a59e61c71b7a8 - AdaptiveEmbedding model · level 1 · ran ·
e77541c1b4b15830 - DecoderLayer helper · level 0 · unverified ·
fc52846dc775edff - PositionalEmbedding helper · level 1 · ran ·
06ce2b899569db5b - RelLearnableDecoderLayer model · level 0 · unverified ·
752529d23a4b3b22 - RelPartialLearnableDecoderLayer model · level 0 · unverified ·
fd8d58e6a1428332 - TransfoXLPreTrainedModel helper · level 0 · unverified ·
d7bd7d30264b3658 - MultiHeadAttn helper · level 0 · unverified ·
d4857f57ecefa3d5 - PositionwiseFF helper · level 0 · unverified ·
62b1b4ab79fdfc8b - RelLearnableMultiHeadAttn model · level 0 · unverified ·
3dafbc50941af43d - RelPartialLearnableMultiHeadAttn model · level 0 · unverified ·
8361fa24d3e06e13 - TransfoXLConfig helper · level 1 · ran ·
a97a22ac6fb31922 - cached_path helper · level 0 · unverified ·
6fee57012e68e835 - load_tf_weights_in_transfo_xl helper · level 0 · unverified ·
1ebc2bf2c3cddc4d - RelMultiHeadAttn model · level 0 · unverified ·
dc7c9e91c34d6ddf - get_from_cache helper · level 0 · unverified ·
b4bb49d2de702f67 - build_tf_to_pytorch_map helper · level 1 · ran · our draft was wrong ·
35ef111fc00c21b4 - http_get helper · level 0 · unverified ·
18f49ada16168f38 - s3_etag helper · level 0 · unverified ·
51dd06ec77c8a8d1 - s3_get helper · level 0 · unverified ·
1079045e75efbf23 - url_to_filename helper · level 1 · ran · our draft was wrong ·
5eff22fa0a651276 - s3_request helper · level 1 · ran · our draft was wrong ·
7661139be003d514 - split_s3_path helper · level 0 · unverified ·
d3852f416bfe86c3
harvested Assembly for Transformer-XL (graph method: Transformer-XL). 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 18 siblings
- RelativeMultiHeadAttention entry · level 1 · ran ·
c018d465b6253b94
harvested Assembly for Transformer-XL (graph method: Transformer-XL). 14 bricks; individually on a fixture: 9 ran (of which 8 constructed an object rather than computing a result; 8 with no instrument failure: 0 honoured, 0 violated, 8 with no contract checked; 1 where Syntology's instrument failed) · 5 unverified. The assembly has not been run end-to-end.
one of 18 siblings
- MemTransformerLM entry · level 0 · unverified ·
9ef2fc89678fefb2 - AdaptiveEmbedding model · level 1 · ran ·
1ae137f0cee15e33 - DecoderLayer model · level 1 · ran ·
e8aac29a19d0c2e7 - LogUniformSampler helper · level 1 · ran ·
220657028509cfb1 - PositionalEmbedding helper · level 1 · ran ·
06ce2b899569db5b - ProjectedAdaptiveLogSoftmax helper · level 1 · ran ·
338c39719307318d - RelLearnableDecoderLayer model · level 0 · unverified ·
17ee18e7e701439f - RelPartialLearnableDecoderLayer model · level 0 · unverified ·
89afd0bff30e18a9 - sample_logits helper · level 1 · ran · our draft was wrong ·
93228a3eb5c4d179 - MultiHeadAttn helper · level 1 · ran ·
d99c9cc9af9de483 - PositionwiseFF helper · level 1 · ran ·
c96fba039a607703 - RelLearnableMultiHeadAttn model · level 0 · unverified ·
a033a134e9f6452f - RelPartialLearnableMultiHeadAttn model · level 0 · unverified ·
fb613dd5bc21371c - RelMultiHeadAttn model · level 1 · ran ·
fe28ba867add05c0
harvested Assembly for Transformer-XL (graph method: Transformer-XL). 12 bricks; individually on a fixture: 4 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; 4 where Syntology's instrument failed) · 8 unverified. The assembly has not been run end-to-end.
one of 18 siblings
- transformer entry · level 0 · unverified ·
d0f56431ab6c3b37 - _cache_mem model · level 1 · ran; the contract check could not run (RAISES) ·
773651748c9618d6 - _create_mask helper · level 0 · unverified ·
ed15761e38990647 - mask_adaptive_embedding_lookup helper · level 0 · unverified ·
99e35dc770639097 - mask_adaptive_logsoftmax helper · level 0 · unverified ·
fca4d6d03913ae24 - mul_adaptive_embedding_lookup helper · level 0 · unverified ·
8d07f1308bd24cbc - mul_adaptive_logsoftmax helper · level 0 · unverified ·
e344186184b46850 - positional_embedding model · level 1 · ran; the contract check could not run (RAISES) ·
d50e8679cde3a927 - positionwise_FF model · level 0 · unverified ·
3652bf001c221967 - rel_multihead_attn model · level 0 · unverified ·
b1f1afea05cd968b - embedding_lookup helper · level 1 · ran; the contract check could not run (RAISES) ·
9109ba9791e44324 - rel_shift model · level 1 · ran; the contract check could not run (RAISES) ·
d37883dc58ff9b69
harvested Assembly for Transformer-XL (graph method: Transformer-XL). 11 bricks; individually on a fixture: 7 ran (of which 1 constructed an object rather than computing a result; 3 with no instrument failure: 2 honoured, 0 violated, 1 with no contract checked; 4 where Syntology's instrument failed) · 4 unverified. The assembly has not been run end-to-end.
one of 18 siblings
- PARTransformerXL entry · level 0 · unverified ·
de3ab66ec62fc167 - AdaptiveSoftmax model · level 0 · unverified ·
87202268910cb496 - StochasticBlock model · level 0 · unverified ·
534b6317f069c394 - RelMultiHeadAttention model · level 1 · ran ·
4c8b264bb79a35a1 - RelaxedOneHot helper · level 0 · unverified ·
1be1a875d12d10b3 - positionwise_ffn model · level 1 · ran · our draft was wrong ·
f7a746d73ab2f28c - create_lookahead_mask helper · level 1 · ran · honoured contract ·
499bf5fee7915c78 - positional_encoding helper · level 1 · ran · honoured contract ·
2398c4a5c285bdc4 - relative_position_logits helper · level 1 · ran; the contract check could not run (RAISES) ·
edeef97857b7cdca - left_shift helper · level 1 · ran; the contract check could not run (RAISES) ·
8d59a6d5900428f5 - right_shift helper · level 1 · ran; the contract check could not run (RAISES) ·
df4647a6d9f03c0e
harvested Assembly for Transformer-XL (graph method: Transformer-XL). 2 bricks; individually on a fixture: 2 ran (of which 2 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified; every one of the 2 samples that ran constructed an object rather than computing a result. The assembly has not been run end-to-end.
one of 18 siblings
- RelPositionMultiHeadedAttention entry · level 1 · ran ·
4977fa89216a56fa - MultiHeadedAttention helper · level 1 · ran ·
5597fd8e1219b866
harvested Assembly for Transformer-XL (graph method: Transformer-XL). 7 bricks; individually on a fixture: 4 ran (of which 2 constructed an object rather than computing a result; 3 with no instrument failure: 1 honoured, 0 violated, 2 with no contract checked; 1 where Syntology's instrument failed) · 3 unverified. The assembly has not been run end-to-end.
one of 18 siblings
- GemmaInfiniAttention entry · level 0 · unverified ·
de07f82e734e3e2c - GemmaAttention helper · level 0 · unverified ·
e4e12108c18eb032 - GemmaConfig helper · level 1 · ran ·
709b8a145b15f0ca - GemmaRotaryEmbedding helper · level 1 · ran ·
39c5be8f7b1c0490 - apply_rotary_pos_emb helper · level 0 · unverified ·
5c61a4c30ce59547 - repeat_kv helper · level 1 · ran; the contract check could not run (RAISES) ·
4e762613ee907cdb - rotate_half helper · level 1 · ran · honoured contract ·
cec833a99feb528c
harvested Assembly for Transformer-XL (graph method: Transformer-XL). 6 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) · 2 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 18 siblings
- TransformerXL entry · level 0 · unverified ·
fa690903d5ea5d89 - TransformerEmbedding helper · level 1 · ran ·
c7b5eb55f15a37de - XLAttentionLayer model · level 0 · unverified ·
91e5af7605c83b9a - PositionalEncoding helper · level 1 · ran ·
07bf8cf6e731d1bc - TokenEmbedding helper · level 1 · ran ·
0840d53a3e87bbd0 - FeedForward helper · level 1 · ran ·
2893a3bddf619d82
harvested Assembly for Transformer-XL (graph method: Transformer-XL). 16 bricks; individually on a fixture: 8 ran (of which 7 constructed an object rather than computing a result; 7 with no instrument failure: 0 honoured, 0 violated, 7 with no contract checked; 1 where Syntology's instrument failed) · 8 unverified. The assembly has not been run end-to-end.
one of 18 siblings
- MemTransformerLM entry · level 0 · unverified ·
5ffa84d26ba39532 - AdaptiveEmbedding model · level 1 · ran ·
8a8691f05cb1b5da - DecoderLayer model · level 0 · unverified ·
c614edf5b8c0d3b0 - LogUniformSampler helper · level 1 · ran ·
220657028509cfb1 - PositionalEmbedding helper · level 1 · ran ·
06ce2b899569db5b - ProjectedAdaptiveLogSoftmax helper · level 1 · ran ·
4088ee884194d7c4 - RelLearnableDecoderLayer model · level 0 · unverified ·
ea7c8500cbc85ff5 - RelPartialLearnableDecoderLayer model · level 0 · unverified ·
5014b04d3358c101 - sample_logits helper · level 0 · unverified ·
582150577e5bd769 - MultiHeadAttn helper · level 1 · ran ·
e672e8a20905b62b - PositionwiseFF helper · level 1 · ran ·
c96fba039a607703 - OptionalParameterList helper · level 0 · unverified ·
55b4958b2a4e4373 - RelLearnableMultiHeadAttn model · level 0 · unverified ·
cb862dd19b1f381a - RelPartialLearnableMultiHeadAttn model · level 0 · unverified ·
5e1b71bde76b69e7 - RelMultiHeadAttn model · level 1 · ran ·
8fa37260188b5fd8 - add_and_scale helper · level 1 · ran; the contract check could not run (RAISES) ·
78d91efd74e4d12a
harvested Assembly for Transformer-XL (graph method: Transformer-XL). 14 bricks; individually on a fixture: 5 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; 3 where Syntology's instrument failed) · 9 unverified. The assembly has not been run end-to-end.
one of 18 siblings
- transformer_xl entry · level 0 · unverified ·
d077730400326171 - _cache_mem model · level 1 · ran; the contract check could not run (RAISES) ·
9d5e914b24b8911a - _create_mask helper · level 0 · unverified ·
c1f8a649a7a837d5 - embedding_lookup helper · level 0 · unverified ·
5abf0267f0fc7514 - positionwise_ffn model · level 0 · unverified ·
ddf11fc956cd28ce - rel_multihead_attn model · level 0 · unverified ·
6ccb8d28491109b6 - relative_positional_encoding model · level 1 · ran · honoured contract ·
8e76562720b0c921 - two_stream_rel_attn model · level 0 · unverified ·
2d935f29546db9cd - gelu helper · level 1 · ran · honoured contract ·
d08f324f950148de - head_projection model · level 0 · unverified ·
6804d62cb7cd8df3 - post_attention model · level 0 · unverified ·
cf208c8c20d52c1b - rel_attn_core model · level 0 · unverified ·
b517847a8c809203 - positional_embedding helper · level 1 · ran; the contract check could not run (RAISES) ·
9bf0f6e6298f3fbe - rel_shift model · level 1 · ran; the contract check could not run (RAISES) ·
5f7c5b38aeb29bf5
harvested Assembly for Transformer-XL (graph method: Transformer-XL). 2 bricks; 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) · 1 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 18 siblings
- RelativeMultiHeadAttention entry · level 0 · unverified ·
274ae09f88dbb42a - Linear helper · level 1 · ran ·
1f215b06ae5d8476
- MemTransformerLM · 7e8e94db34a4c27b · 14 bricks; individually on a fixture: 9 ran (of which 8 constructed an object rather than computing a result; 8 with no instrument failure: 0 honoured, 0 violated, 8 with no contract checked; 1 where Syntology's instrument failed) · 5 unverified · The assembly has not been run end-to-end · assemblies[12] in the JSON twin
- TransformerModel · b2b2f8242acdc1d5 · 9 bricks; individually on a fixture: 7 ran (of which 6 constructed an object rather than computing a result; 6 with no instrument failure: 0 honoured, 0 violated, 6 with no contract checked; 1 where Syntology's instrument failed) · 2 unverified · The assembly has not been run end-to-end · assemblies[13] in the JSON twin
- XLNet · e405f80080d3da6d · 1 brick; individually on a fixture: 0 ran · 1 unverified · The assembly has not been run end-to-end · assemblies[14] in the JSON twin
- RelativeMultiHeadAttention · ed79e3501287d964 · 2 bricks; individually on a fixture: 2 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 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 · assemblies[15] in the JSON twin
- TransformerLayer · f7629f5167ba1e0d · 5 bricks; individually on a fixture: 3 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; 3 where Syntology's instrument failed) · 2 unverified · The assembly has not been run end-to-end · assemblies[16] in the JSON twin
- MemTransformerLM · ff415c7b0c04a973 · 14 bricks; individually on a fixture: 8 ran (of which 7 constructed an object rather than computing a result; 7 with no instrument failure: 0 honoured, 0 violated, 7 with no contract checked; 1 where Syntology's instrument failed) · 6 unverified · The assembly has not been run end-to-end · assemblies[17] in the JSON twin
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="1901.02860") 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 | Hutter Prize | 24-layer Transformer-XL | Bit per Character (BPC) | 0.99 | #4 of 18 | Archive leaderboard | report |
| Language Modelling | Hutter Prize | 24-layer Transformer-XL | Number of params | 277M | #4 of 18 | Archive leaderboard | report |
| Language Modelling | Hutter Prize | 18-layer Transformer-XL | Bit per Character (BPC) | 1.03 | #7 of 18 | Archive leaderboard | report |
| Language Modelling | Hutter Prize | 18-layer Transformer-XL | Number of params | 88M | #7 of 18 | Archive leaderboard | report |
| Language Modelling | Hutter Prize | 12-layer Transformer-XL | Bit per Character (BPC) | 1.06 | #9 of 18 | Archive leaderboard | report |
| Language Modelling | Hutter Prize | 12-layer Transformer-XL | Number of params | 41M | #9 of 18 | Archive leaderboard | report |
| Language Modelling | One Billion Word | Transformer-XL Large | Number of params | 0.8B | #4 of 27 | Archive leaderboard | report |
| Language Modelling | One Billion Word | Transformer-XL Large | PPL | 21.8 | #4 of 27 | Archive leaderboard | report |
| Language Modelling | One Billion Word | Transformer-XL Base | Number of params | 0.46B | #8 of 27 | Archive leaderboard | report |
| Language Modelling | One Billion Word | Transformer-XL Base | PPL | 23.5 | #8 of 27 | Archive leaderboard | report |
| Language Modelling | Penn Treebank (Word Level) | Transformer-XL | Params | 24M | #22 of 43 | Archive leaderboard | report |
| Language Modelling | Penn Treebank (Word Level) | Transformer-XL | Test perplexity | 54.55 | #22 of 43 | Archive leaderboard | report |
| Language Modelling | Penn Treebank (Word Level) | Transformer-XL | Validation perplexity | 56.72 | #22 of 43 | Archive leaderboard | report |
| Language Modelling | Text8 | Transformer-XL - 24 layers | Bit per Character (BPC) | 1.08 | #5 of 24 | Archive leaderboard | report |
| Language Modelling | Text8 | Transformer-XL - 24 layers | Number of params | 277M | #5 of 24 | Archive leaderboard | report |
| Language Modelling | WikiText-103 | Transformer-XL Large | Number of params | 257M | #34 of 89 | Archive leaderboard | report |
| Language Modelling | WikiText-103 | Transformer-XL Large | Test perplexity | 18.3 | #34 of 89 | Archive leaderboard | report |
| Language Modelling | WikiText-103 | Transformer-XL Large | Validation perplexity | 18.2 | #34 of 89 | Archive leaderboard | report |
| Language Modelling | WikiText-103 | Transformer-XL Standard | Number of params | 151M | #54 of 89 | Archive leaderboard | report |
| Language Modelling | WikiText-103 | Transformer-XL Standard | Test perplexity | 24.0 | #54 of 89 | Archive leaderboard | report |
| Language Modelling | WikiText-103 | Transformer-XL Standard | Validation perplexity | 23.1 | #54 of 89 | Archive leaderboard | report |
| Language Modelling | enwik8 | Transformer-XL (24 layers) | Bit per Character (BPC) | 0.99 | #12 of 42 | Archive leaderboard | report |
| Language Modelling | enwik8 | Transformer-XL (24 layers) | Number of params | 277M | #12 of 42 | Archive leaderboard | report |
| Language Modelling | enwik8 | Transformer-XL (18 layers) | Bit per Character (BPC) | 1.03 | #23 of 42 | Archive leaderboard | report |
| Language Modelling | enwik8 | Transformer-XL (18 layers) | Number of params | 88M | #23 of 42 | Archive leaderboard | report |
| Language Modelling | enwik8 | Transformer-XL (12 layers) | Bit per Character (BPC) | 1.06 | #26 of 42 | Archive leaderboard | report |
| Language Modelling | enwik8 | Transformer-XL (12 layers) | Number of params | 41M | #26 of 42 | 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: Transformer-XL
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