Papers › Vision Transformers Need Registers

Vision Transformers Need Registers

28 Sep 2023arXiv:2309.16588archive 2025-07-28

Timothée Darcet, Maxime Oquab, Julien Mairal, Piotr Bojanowski

Transformers have recently emerged as a powerful tool for learning visual representations. In this paper, we identify and characterize artifacts in feature maps of both supervised and self-supervised ViT networks. The artifacts correspond to high-norm tokens appearing during inference primarily in low-informative background areas of images, that are repurposed for internal computations. We propose a simple yet effective solution based on providing additional tokens to the input sequence of the Vision Transformer to fill that role. We show that this solution fixes that problem entirely for both supervised and self-supervised models, sets a new state of the art for self-supervised visual models on dense visual prediction tasks, enables object discovery methods with larger models, and most importantly leads to smoother feature maps and attention maps for downstream visual processing.

PaperPDFCodeCode Syntology ran

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 20 code samples: 18 from 2 repositories linked to this paper, 2 identical to code harvested from another paper's repository. Across all 20: 15 ran (of which 0 constructed an object rather than computing a result; 13 with no instrument failure: 1 honoured, 1 violated, 11 with no contract checked; 2 where Syntology's instrument failed) · 5 unverified. Where Syntology's instrument failed: 2 where our contract draft was wrong; failures of Syntology's instrument, not of the code.

By repository: official repository: 3 samples from 1 repository, 3 ran (1 with no instrument failure; 2 where Syntology's instrument failed); community (archive-listed): 15 samples from 1 repository, 12 ran (12 with no instrument failure; 0 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 3 of its repositories that Syntology harvested for other papers:

Levels written between 2026-09-22 and 2026-09-24 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed. 3 of the 15 samples that ran have no recorded date.

facebookresearch/dinov2 officialmentioned in papermentioned on GitHubpytorchSyntology: 3 samples, 3 ran (2 instrument-failed) · 0 unverified report
borisdayma/clip-jax mentioned on GitHubjaxApache-2.0Syntology: 15 samples, 12 ran (0 instrument-failed) · 3 unverified report
locuslab/massive-activations mentioned on GitHubpytorchMITSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Massive Activations in Large Language Models (arXiv:2402.17762; that paper's own run record: 6 ran (of which 0 constructed an object rather than computing a result; 5 with no instrument failure: 1 honoured, 1 violated, 3 with no contract checked; 1 where Syntology's instrument failed) · 16 unverified); nothing checks that they implement this paper's method. report
nickjiang2378/test-time-registers mentioned on GitHubpytorchSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Vision Transformers Don't Need Trained Registers (arXiv:2506.08010; that paper's own run record: 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) · 1 unverified); nothing checks that they implement this paper's method. report
rwightman/pytorch-image-models pytorchSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to AdamP: Slowing Down the Slowdown for Momentum Optimizers on Scale-invariant Weights (arXiv:2006.08217; that paper's own run record: 4 ran (of which 1 constructed an object rather than computing a result; 2 with no instrument failure: 1 honoured, 0 violated, 1 with no contract checked; 2 where Syntology's instrument failed) · 0 unverified); nothing checks that they implement this paper's method. report
gitlab.com/birder/birder pytorchSyntology: not harvested 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

20 samples harvested: 15 ran (of which 0 constructed an object rather than computing a result; 13 with no instrument failure: 1 honoured, 1 violated, 11 with no contract checked; 2 where Syntology's instrument failed) · 5 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.

1ran · honoured contract
1ran · violated contract
2ran · our draft was wrong
11ran
5unverified

Licence: 2 of the 20 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 2 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”. “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.

facebookresearch/dinov2 official repository 3 samples, 3 ran (2 instrument-failed) · 0 unverified
convert_path_or_url_to_url dinov2/hub/backbones.py ran · our draft was wrong fingerprinted Apache-2.0 (permissive) · 7715dbecb1b8cfd6 · report fixture: signature path:str/0/str · returned str · level recorded, date not recorded
is_url in 30 papers dinov2/hub/backbones.py ran · violated contract Apache-2.0 (permissive) · 911ac52477a145cc · report fixture: signature path:str/0/str · returned bool · level recorded, date not recorded
named_apply in 32 papers dinov2/models/vision_transformer.py ran · our draft was wrong Apache-2.0 (permissive) · e7fcb6d9ac9deaf4 · report fixture: signature fn:callable/0/?, module:callable/0/?, name:str/0/str, depth_first:bool/0/bool, include_root:bool/0/bool · returned SimpleModel · level recorded, date not recorded
borisdayma/clip-jax community (archive-listed) 15 samples, 12 ran (0 instrument-failed) · 3 unverified
count_params in 80 papers clip_jax/utils.py ran Apache-2.0 (permissive) · 141f5bebbf4455dd · report fixture: signature pytree:dict/0/? · returned int · level recorded 2026-09-24
image_to_logits clip_jax/data.py ran Apache-2.0 (permissive) · d3f6068e5f1cc03f · report fixture: signature image:ndarray/3/int, mean:tuple/1/float, std:tuple/1/float, format:str/0/str · returned ndarray · level recorded 2026-09-24
logical_axis_rules clip_jax/partitions.py ran fingerprinted Apache-2.0 (permissive) · c8e657cf59894a37 · report fixture: signature activation_partitioning_dims:int/0/int, parameter_partitioning_dims:int/0/int · returned list · level recorded 2026-09-24
maybe_get_artifact clip_jax/wandb_utils.py ran fingerprinted Apache-2.0 (permissive) · 4be991be800e82bc · report fixture: signature pretrained_model_name_or_path:str/0/str, tmp_dir:str/0/str · returned str · level recorded 2026-09-24
norm in 120 papers
census: 23 implementations from 26 papers, 15 distinct outputs across 6 buckets, one shared input per bucket
clip_jax/modeling.py ran Apache-2.0 (permissive) · e89c437719b73776 · report fixture: signature use_rmsnorm:bool/0/bool · returned function · level recorded 2026-09-24
normalize in 638 papers
census: 70 implementations from 82 papers, 51 distinct outputs across 11 buckets, one shared input per bucket
clip_jax/modeling.py ran Apache-2.0 (permissive) · 9446a8d501beb52e · report fixture: signature x:ArrayImpl/2/float, eps:float/0/float, safe_norm:bool/0/bool · returned ArrayImpl · level recorded 2026-09-24
posemb_sincos_2d in 20 papers clip_jax/modeling.py ran · honoured contract Apache-2.0 (permissive) · 2d67a8a756fe7cf8 · report fixture: signature h:int/0/int, w:int/0/int, width:int/0/int, temperature:float/0/float, dtype:callable/0/? · returned ArrayImpl · level recorded 2026-09-22
sample_nucleus_topp_logits clip_jax/maxtext/inference_utils.py ran Apache-2.0 (permissive) · 9e0d5ce23e1ddbfb · report fixture: signature logits:ArrayImpl/2/float, nucleus_topp:float/0/float, temperature:float/0/float, rng:ArrayImpl/1/int · returned ArrayImpl · level recorded 2026-09-24
sample_topk_logits clip_jax/maxtext/inference_utils.py ran Apache-2.0 (permissive) · 213c7fbbc33f5eec · report fixture: signature logits:ArrayImpl/2/float, topk:int/0/int, temperature:float/0/float, rng:ArrayImpl/1/int · returned ArrayImpl · level recorded 2026-09-24
sampling in 26 papers clip_jax/maxtext/inference_utils.py ran Apache-2.0 (permissive) · 5d8736f0387a699d · report fixture: signature logits:ArrayImpl/2/float, rng:ArrayImpl/1/int, algorithm:str/0/str, topk:int/0/int, nucleus_topp:float/0/float, temperature:float/0/float · returned ArrayImpl · level recorded 2026-09-24
shift_tokens_left clip_jax/data.py ran Apache-2.0 (permissive) · fd9f7273b986bc74 · report fixture: signature logits:ndarray/2/float, pad_token_id:int/0/int · returned ndarray · level recorded 2026-09-24
with_logical_constraint clip_jax/utils.py ran Apache-2.0 (permissive) · b2d0cfef5bc878ce · report fixture: signature x:ArrayImpl/2/float, c:ArrayImpl/1/int, use_constraint:bool/0/bool · returned ArrayImpl · level recorded 2026-09-24
get_system_characteristics clip_jax/maxtext/accelerator_to_spec_map.py unverified Apache-2.0 (permissive) · 87c57735e68651cc · report
load_config in 389 papers clip_jax/utils.py unverified Apache-2.0 (permissive) · 831515af28a2a745 · report
logits_to_image clip_jax/data.py unverified Apache-2.0 (permissive) · e147d773ebc71aa4 · report
Repository not located 2 samples, 0 ran · 2 unverified
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.
has_ddp_wrapper in 6 papers repository not located unverified licence of this copy not recorded · 1cd2728a84872357 · report
remove_ddp_wrapper in 5 papers repository not located unverified licence of this copy not recorded · f565e9c6316d8c64 · report

Tasks

Object DiscoverySelf-Supervised Image Classification

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Self-Supervised Image Classification ImageNet DINOv2+reg (ViT-g/14) Number of Params 1100M #1 of 144 Archive leaderboard report
Self-Supervised Image Classification ImageNet DINOv2+reg (ViT-g/14) Top 1 Accuracy 87.1 #1 of 144 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

Absolute Position EncodingsAdamAttentionBPEDense ConnectionsDropoutLabel SmoothingLayer NormalizationLinear LayerMulti-Head AttentionPosition-Wise Feed-Forward LayerResidual ConnectionSoftmaxTransformerVision Transformer

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections