Papers › GAN Dissection: Visualizing and Understanding Generative Adversarial Networks

GAN Dissection: Visualizing and Understanding Generative Adversarial Networks

26 Nov 2018ICLR 2019 5arXiv:1811.10597archive 2025-07-28

David Bau, Jun-Yan Zhu, Hendrik Strobelt, Bolei Zhou, Joshua B. Tenenbaum, William T. Freeman, Antonio Torralba

Generative Adversarial Networks (GANs) have recently achieved impressive results for many real-world applications, and many GAN variants have emerged with improvements in sample quality and training stability. However, they have not been well visualized or understood. How does a GAN represent our visual world internally? What causes the artifacts in GAN results? How do architectural choices affect GAN learning? Answering such questions could enable us to develop new insights and better models. In this work, we present an analytic framework to visualize and understand GANs at the unit-, object-, and scene-level. We first identify a group of interpretable units that are closely related to object concepts using a segmentation-based network dissection method. Then, we quantify the causal effect of interpretable units by measuring the ability of interventions to control objects in the output. We examine the contextual relationship between these units and their surroundings by inserting the discovered object concepts into new images. We show several practical applications enabled by our framework, from comparing internal representations across different layers, models, and datasets, to improving GANs by locating and removing artifact-causing units, to interactively manipulating objects in a scene. We provide open source interpretation tools to help researchers and practitioners better understand their GAN models.

PaperPDFConference PDFCodeCode 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 34 code samples harvested from 5 repositories linked to this paper: 25 ran (of which 0 constructed an object rather than computing a result; 23 with no instrument failure: 0 honoured, 0 violated, 23 with no contract checked; 2 where Syntology's instrument failed) · 9 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: 11 samples from 1 repository, 10 ran (10 with no instrument failure; 0 where Syntology's instrument failed); community (archive-listed): 23 samples from 4 repositories, 15 ran (13 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.

Levels written between 2026-09-24 and 2026-09-26 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed. 2 of the 25 samples that ran have no recorded date.

CSAILVision/gandissect officialmentioned in papermentioned on GitHubpytorchMITSyntology: 11 samples, 10 ran (0 instrument-failed) · 1 unverified report
AllenChen0958/homework3-GAN-Dissection mentioned on GitHubpytorchMITSyntology: 2 samples, 2 ran (0 instrument-failed) · 0 unverified report
alexandonian/ganocracy mentioned on GitHubpytorchMITSyntology: 14 samples, 10 ran (2 instrument-failed) · 4 unverified report
cvfx-2019/homework3-GAN-Dissection mentioned on GitHubpytorchMITSyntology: not harvested report
freedombenLiu/gandissect mentioned on GitHubpytorchSyntology: 3 samples, 0 ran · 3 unverified report
mit-quest/bridge_gandissect mentioned on GitHubpytorchMITSyntology: not harvested report
sophiaalexander/GANDissect-for-High-Schoolers mentioned on GitHubpytorchMITSyntology: 4 samples, 3 ran (0 instrument-failed) · 1 unverified report
sophiaalexander/GANs-for-High-Schoolers mentioned on GitHubpytorchMITSyntology: 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.

Code Syntology ran Syntology

34 samples harvested: 25 ran (of which 0 constructed an object rather than computing a result; 23 with no instrument failure: 0 honoured, 0 violated, 23 with no contract checked; 2 where Syntology's instrument failed) · 9 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.

2ran · our draft was wrong
23ran
9unverified

Licence: 3 of the 34 samples are pointer only, meaning Syntology does not serve that copy's text. For 3 of them, this paper's copy is under a permissive licence and cleared, but the stored copy of the same code that Syntology serves is not; the licence label's tooltip gives Syntology's recorded reason. 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: an argument's recorded type or element kind is a printout of an object in Syntology's harness (such as a mock object, a class or a pandas printout), not the type of an input, so it does not describe the call's shape. “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.

CSAILVision/gandissect official repository 11 samples, 10 ran (0 instrument-failed) · 1 unverified
activation_surface netdissect/actviz.py ran MIT (permissive) · be1a5843b99a3274 · report fixture: signature data:ndarray/2/float, target_shape:tuple/1/int, source_shape:tuple/1/int, scale_offset:tuple/1/tuple, deg:int/0/int, pad:bool/0/bool · returned ndarray · level recorded 2026-09-25
activation_visualization netdissect/actviz.py ran MIT (permissive) · add93cb6d00227f2 · report fixture: signature image:ndarray/3/int, data:ndarray/2/float, level:float/0/float, alpha:float/0/float, source_shape:NoneType/0/?, crop:bool/0/bool, zoom:NoneType/0/?, border:int/0/int, negate:bool/0/bool, return_mask:bool/0/bool · returned ndarray · level recorded 2026-09-25
annotate_model_shapes netdissect/modelconfig.py ran MIT (permissive) · 6bc3f04e4a74356d · report fixture: signature model:callable/0/?, gen:bool/0/bool, imgsize:tuple/1/int · returned SimpleModel · level recorded 2026-09-25
autoimport_eval netdissect/autoeval.py ran MIT (permissive) · 4acf7e1c09b788e9 · report fixture: signature term:str/0/str · returned dict · level recorded 2026-09-25
build_numpy_category_map netdissect/broden.py ran MIT (permissive) · 2df0915cd60cd1be · report fixture: signature map_data:list/1/dict, key1:str/0/str, key2:str/0/str · returned list · level recorded 2026-09-25
create_temp_mmap_grid netdissect/dissection.py ran MIT (permissive) · 50f195ea6bd9b847 · report fixture: signature shape:tuple/1/int, dtype:str/0/str, fill:float/0/float · returned memmap · level recorded 2026-09-25
load_json in 344 papers netdissect/easydict.py ran MIT (permissive) · c4b1e3a489d47780 · report fixture: signature filename:str/0/str · returned EasyDict · level recorded 2026-09-25
mask_border netdissect/actviz.py ran MIT (permissive) · caca87125879a567 · report fixture: signature mask:ndarray/2/int, border:int/0/int · returned ndarray · level recorded 2026-09-25
scale_summary netdissect/aceoptimize.py ran MIT (permissive) · 62613471797cd12b · report fixture: signature scale:Tensor/1/float, lownums:int/0/int, highnums:int/0/int · returned str · level recorded 2026-09-25
shared_temp_mmap_grid netdissect/dissection.py ran MIT (permissive) · 29455122b233b9d8 · report fixture: returned memmap · level recorded 2026-09-25
draw_heatmap netdissect/aceoptimize.py unverified MIT (permissive) · 23f4b6af18f81f92 · report
AllenChen0958/homework3-GAN-Dissection community (archive-listed) 2 samples, 2 ran (0 instrument-failed) · 0 unverified
build_dense_label_array netdissect/broden.py ran MIT (permissive) · 1ca87d6074add59d · report fixture: signature label_data:list/1/dict, key:str/0/str, allow_none:bool/0/bool · returned list · level recorded 2026-09-24
index_has_any_data netdissect/broden.py ran MIT (permissive) · 130670bcfa93f2d8 · report fixture: signature row:dict/0/?, categories:list/1/str · returned bool · level recorded 2026-09-24
alexandonian/ganocracy community (archive-listed) 14 samples, 10 ran (2 instrument-failed) · 4 unverified
D_arch in 21 papers gan_training/ganocracy/models/biggan.py ran · our draft was wrong MIT (permissive) here; the stored copy is pointer only · 85eb95cce203dfc7 · report fixture: signature ch:int/0/int, attention:str/0/str, ksize:str/0/str, dilation:str/0/str · returned dict · level recorded, date not recorded
G_arch in 22 papers gan_training/ganocracy/models/biggan.py ran · our draft was wrong MIT (permissive) here; the stored copy is pointer only · e8c308c9283407d7 · report fixture: signature ch:int/0/int, attention:str/0/str, ksize:str/0/str, dilation:str/0/str · returned dict · level recorded, date not recorded
gram_schmidt in 32 papers gan_training/ganocracy/layers/norm.py ran MIT (permissive) · 9741295ab539fa46 · report fixture: signature x:Tensor/2/float, ys:list/1/Tensor · returned Tensor · level recorded 2026-09-25
hook_sizes gan_training/ganocracy/models/utils.py ran MIT (permissive) · 494bb0ccbe9e7198 · report fixture: signature G:callable/0/?, inputs:tuple/1/Tensor, verbose:bool/0/bool · returned tuple · level recorded 2026-09-26
interp in 22 papers gan_training/ganocracy/utils/visualizer.py ran MIT (permissive) · a4672b1fe29d691a · report fixture: signature x0:Tensor/2/float, x1:Tensor/2/float, num_midpoints:int/0/int, device:str/0/str · returned Tensor · level recorded 2026-09-26
power_iteration in 30 papers gan_training/ganocracy/layers/norm.py ran MIT (permissive) · c14a1db70dacc94a · report fixture: signature W:Tensor/2/float, u_:list/1/Tensor, update:bool/0/bool, eps:float/0/float · returned tuple · level recorded 2026-09-25
smooth_data in 5 papers gan_training/ganocracy/utils/visualizer.py ran fingerprinted MIT (permissive) · dcc34da31f1514dd · report fixture: signature data:ndarray/1/float, amount:float/0/float · returned ndarray · level recorded 2026-09-26
sqrt_newton_schulz in 14 papers gan_training/ganocracy/metrics/inception_score.py ran MIT (permissive) · 16aba3aaeb3655f3 · report fixture: signature A:Tensor/3/float, numIters:int/0/int, dtype:NoneType/0/? · returned Tensor · level recorded 2026-09-26
torch_cov in 14 papers gan_training/ganocracy/metrics/inception_score.py ran fingerprinted MIT (permissive) · 764abafe62f0d3e9 · report fixture: signature m:Tensor/2/float, rowvar:bool/0/bool · returned Tensor · level recorded 2026-09-26
truncated_z_sample gan_training/ganocracy/utils/visualizer.py ran MIT (permissive) · db2b0dc3665a41d1 · report fixture: signature batch_size:int/0/int, dim_z:int/0/int, truncation:float/0/float, seed:int/0/int, device:str/0/str · returned Tensor · level recorded 2026-09-26
numpy_calculate_frechet_distance in 8 papers gan_training/ganocracy/metrics/inception_score.py unverified MIT (permissive) · 7f2187cc95ef4c91 · report
proj in 38 papers gan_training/ganocracy/layers/norm.py unverified MIT (permissive) · cc36462ea199cca3 · report
validate in 313 papers gan_training/ganocracy/metrics/train_inception.py unverified MIT (permissive) · 65677fad42a182e2 · report
freedombenLiu/gandissect community (archive-listed) 3 samples, 0 ran · 3 unverified
get_layers in 20 papers netdissect/server.py unverified MIT (permissive) · a7ff4317d03c4c34 · report
get_rankings netdissect/server.py unverified MIT (permissive) · 25db45525a41991f · report
get_units netdissect/server.py unverified MIT (permissive) · ba5a60de38335521 · report
sophiaalexander/GANDissect-for-High-Schoolers community (archive-listed) 4 samples, 3 ran (0 instrument-failed) · 1 unverified
activation_visualization netdissect/actviz.py ran MIT (permissive) · 2813ec86a0e1c405 · report fixture: signature image:ndarray/3/float, data:ndarray/2/float, level:float/0/float, alpha:float/0/float, source_shape:tuple/1/int, target_shape:tuple/1/int, crop:bool/0/bool, zoom:NoneType/0/?, border:int/0/int, negate:bool/0/bool, return_mask:bool/0/bool, scale_offset:NoneType/0/?, deg:int/0/int, pad:bool/0/bool · returned ndarray · level recorded 2026-09-26
make_svg_bargraph netdissect/bargraph.py ran MIT (permissive) · 02a5ab7f9313f3cf · report fixture: signature labels:list/1/str, heights:list/1/int, categories:list/1/tuple, palette:list/1/tuple, barheight:int/0/int, barwidth:int/0/int, show_labels:bool/0/bool, file_header:bool/0/bool · returned str · level recorded 2026-09-26
resfile dissect_resnet152.py ran fingerprinted MIT (permissive) · c9a9646005727be8 · report fixture: signature filename:str/0/str · returned str · level recorded 2026-09-26
activation_surface netdissect/actviz.py unverified MIT (permissive) · ea83571322bf19bd · report

Tasks

Image GenerationObject

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

ConvolutionNetwork Dissection

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