Papers › To grok or not to grok: Disentangling generalization and memorization on corrupted...

To grok or not to grok: Disentangling generalization and memorization on corrupted algorithmic datasets

19 Oct 2023arXiv:2310.13061archive 2025-07-28

Darshil Doshi, Aritra Das, Tianyu He, Andrey Gromov

Robust generalization is a major challenge in deep learning, particularly when the number of trainable parameters is very large. In general, it is very difficult to know if the network has memorized a particular set of examples or understood the underlying rule (or both). Motivated by this challenge, we study an interpretable model where generalizing representations are understood analytically, and are easily distinguishable from the memorizing ones. Namely, we consider multi-layer perceptron (MLP) and Transformer architectures trained on modular arithmetic tasks, where (ξ·100%) of labels are corrupted (i.e. some results of the modular operations in the training set are incorrect). We show that (i) it is possible for the network to memorize the corrupted labels and achieve 100% generalization at the same time; (ii) the memorizing neurons can be identified and pruned, lowering the accuracy on corrupted data and improving the accuracy on uncorrupted data; (iii) regularization methods such as weight decay, dropout and BatchNorm force the network to ignore the corrupted data during optimization, and achieve 100% accuracy on the uncorrupted dataset; and (iv) the effect of these regularization methods is (``mechanistically'') interpretable: weight decay and dropout force all the neurons to learn generalizing representations, while BatchNorm de-amplifies the output of memorizing neurons and amplifies the output of the generalizing ones. Finally, we show that in the presence of regularization, the training dynamics involves two consecutive stages: first, the network undergoes grokking dynamics reaching high train and test accuracy; second, it unlearns the memorizing representations, where the train accuracy suddenly jumps from 100% to 100 (1-ξ)%.

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 10 code samples harvested from 1 repository linked to this paper: 10 ran (of which 0 constructed an object rather than computing a result; 10 with no instrument failure: 0 honoured, 0 violated, 10 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code; here there are none.

By repository: official repository: 10 samples from 1 repository, 10 ran (10 with no instrument failure; 0 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 on 2026-09-24 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

d-doshi/Grokking officialmentioned in papermentioned on GitHubpytorchSyntology: 10 samples, 10 ran (0 instrument-failed) · 0 unverified 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

10 samples harvested: 10 ran (of which 0 constructed an object rather than computing a result; 10 with no instrument failure: 0 honoured, 0 violated, 10 with no contract checked; 0 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 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.

10ran

Licence: 10 of the 10 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 d-doshi/Grokking. “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.

d-doshi/Grokking official repository 10 samples, 10 ran (0 instrument-failed) · 0 unverified
bwd_hook_ utils/hooks.py ran no licence file found · pointer only · b9e7372bec816408 · report fixture: signature model:callable/0/?, module:callable/0/?, n:int/0/int · returned Hook · level recorded 2026-09-24
calculate_gini utils/data_processing.py ran fingerprinted no licence file found · pointer only · cc8f4098386e32f6 · report fixture: signature array:ndarray/1/float · returned float64 · level recorded 2026-09-24
calculate_ipr utils/data_processing.py ran fingerprinted no licence file found · pointer only · c9ea61e56776cdc8 · report fixture: signature array:ndarray/1/float, r:int/0/int · returned float64 · level recorded 2026-09-24
check_accuracy_grokking utils/optimization.py ran no licence file found · pointer only · f7c0fbb9c7ee31fc · report fixture: signature X:Tensor/2/float, Y:Tensor/1/int, model:callable/0/?, dtype:dtype/0/?, device:str/0/str, scaler:NoneType/0/? · returned tuple · level recorded 2026-09-24
fwd_hook_ utils/hooks.py ran no licence file found · pointer only · 1957ab794eb163a9 · report fixture: signature model:callable/0/?, module:callable/0/?, n:int/0/int · returned Hook · level recorded 2026-09-24
ipr_test utils/data_processing.py ran no licence file found · pointer only · 903c226243d14f33 · report fixture: signature array:ndarray/1/float, r:int/0/int, chi_ipr:float/0/float · returned bool · level recorded 2026-09-24
noisy_dataset utils/datasets.py ran no licence file found · pointer only · 02ddb24b51a41c52 · report fixture: signature p:int/0/int, pair_seed:int/0/int, frac:float/0/float, noise_level:float/0/float, device:str/0/str, dtype:dtype/0/?, operation:str/0/str, fixed_seed:bool/0/bool · returned dict · level recorded 2026-09-24
noisy_dataset_int utils/datasets.py ran no licence file found · pointer only · 95c18c93cb4a4be6 · report fixture: signature p:int/0/int, pair_seed:int/0/int, frac:float/0/float, noise_level:float/0/float, device:str/0/str, dtype:dtype/0/?, operation:str/0/str, fixed_seed:bool/0/bool · returned dict · level recorded 2026-09-24
test_loss_grokking utils/optimization.py ran no licence file found · pointer only · ba1581a612dca015 · report fixture: signature model:callable/0/?, X:Tensor/2/float, Y:Tensor/1/int, losstype:str/0/str, dtype:dtype/0/?, device:str/0/str, scaler:NoneType/0/? · returned Tensor · level recorded 2026-09-24
train_one_epoch_grokking utils/optimization.py ran no licence file found · pointer only · e3f8da4b30b63eb0 · report fixture: signature model:callable/0/?, optimizer:SGD/0/?, time:int/0/int, X_train:Tensor/2/float, Y_train:Tensor/1/int, X_test:Tensor/2/float, Y_test:Tensor/1/int, dtype:dtype/0/?, device:str/0/str, stopwatch:int/0/int, losstype:str/0/str, scheduler:StepLR/0/?, if_data:bool/0/bool, verbose:bool/0/bool, scaler:NoneType/0/?, if_freeze:bool/0/bool · returned dict · level recorded 2026-09-24

Tasks

Memorization

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

Absolute Position EncodingsAdamAttentionBPEDense ConnectionsDropoutLabel SmoothingLayer NormalizationLinear LayerMulti-Head AttentionPosition-Wise Feed-Forward LayerResidual ConnectionSETSoftmaxTransformerWeight Decay

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