SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2405.16002 · ICLR · 2025

Does SGD really happen in tiny subspaces?

Chulhee Yun, Kwangjun Ahn, Minhak Song

arXiv · PDF · Open in the Atlas

Code that ran

14 of 15 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 6 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 8 failures of our instrument; 1 no recorded run), measured 2026-11-22; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2405.16002.json · badge: /api/ran/2405.16002.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
locuslab/edge-of-stability canonical (classifier label)
classifier basis: census
14 ran · 1 no recorded run
FunctionStatusWhere it lives
center Ran
ran, honoured the contract we drafted
locuslab/edge-of-stability/src/cifar.py
pointer only (licence: NONE) · get_code("b37b8602123cd711")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, locuslab/edge-of-stability when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

conv3x3 Ran
ran; our contract draft was wrong, not the code
locuslab/edge-of-stability/src/resnet_cifar.py
pointer only (licence: NONE); the stored copy get_code serves, from lixiaotong97/DSU, is pointer only (licence: NONE) · get_code("fac5364e2f53c6db")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, locuslab/edge-of-stability when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

flatten Ran
ran, honoured the contract we drafted
locuslab/edge-of-stability/src/cifar.py
pointer only (licence: NONE) · get_code("5ae58d45938cf313")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, locuslab/edge-of-stability when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

get_activation Ran
ran; our contract draft was wrong, not the code
locuslab/edge-of-stability/src/archs.py
pointer only (licence: NONE) · get_code("9d866a28174c7773")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, locuslab/edge-of-stability when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

get_adam_nu Ran
ran, honoured the contract we drafted
locuslab/edge-of-stability/src/adam.py
pointer only (licence: NONE) · get_code("c9020a109e8d40e6")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, locuslab/edge-of-stability when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

get_flow_directory Ran
ran; our contract draft was wrong, not the code
locuslab/edge-of-stability/src/utilities.py
pointer only (licence: NONE) · get_code("6fc397cdc1cb7bfd")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, locuslab/edge-of-stability when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

get_gd_directory Ran
ran; our contract draft was wrong, not the code
locuslab/edge-of-stability/src/utilities.py
pointer only (licence: NONE) · get_code("668f4bd271368cb3")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, locuslab/edge-of-stability when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

get_modified_flow_directory Ran
ran; our contract draft was wrong, not the code
locuslab/edge-of-stability/src/utilities.py
pointer only (licence: NONE) · get_code("469c24f9be20c5be")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, locuslab/edge-of-stability when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

get_pooling Ran
ran; our contract draft was wrong, not the code
locuslab/edge-of-stability/src/archs.py
pointer only (licence: NONE) · get_code("295abaa4814b4e95")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, locuslab/edge-of-stability when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

make_chebyshev_dataset Ran
ran, honoured the contract we drafted
locuslab/edge-of-stability/src/synthetic.py
pointer only (licence: NONE) · get_code("5107ac691376fc43")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, locuslab/edge-of-stability when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

make_layers Ran
ran; our contract draft was wrong, not the code
locuslab/edge-of-stability/src/vgg.py
pointer only (licence: NONE) · get_code("ac62432dc5134b0d")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, locuslab/edge-of-stability when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

num_input_channels Ran
ran, honoured the contract we drafted
locuslab/edge-of-stability/src/data.py
pointer only (licence: NONE) · get_code("3f99563e280b4081")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, locuslab/edge-of-stability when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

standardize Ran
ran, honoured the contract we drafted
locuslab/edge-of-stability/src/cifar.py
pointer only (licence: NONE) · get_code("c931ef45cfe86923")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, locuslab/edge-of-stability when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

unflatten Ran
ran; our contract draft was wrong, not the code
locuslab/edge-of-stability/src/data.py
pointer only (licence: NONE) · get_code("6e1ce477ef41fc78")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, locuslab/edge-of-stability when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

make_linear_dataset Not yet run
unverified (harvested, no recorded run)
locuslab/edge-of-stability/src/synthetic.py
pointer only (licence: NONE) · get_code("f04e278f3a337f1a")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, locuslab/edge-of-stability when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

Also present in other papers' code

Of 15 samples lifted from this paper's linked repositories, 15 carry a function identity and 0 do not. Of the 15 distinct functions with an identity, 2 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 1 the same month, 2 later) and 13 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

FunctionAlso present inEarliest-datedOthers named
conv3x3 686 other papers
588 dated earlier, 7 the same month, 74 later, 17 same year, order not determinable · 31 only as an assembly brick
1406.2661 (2014-06)
Generative Adversarial Networks
1409.0575 (2014-09)
ImageNet Large Scale Visual Recognition Challenge
1412.6572 (2014-12)
Explaining and Harnessing Adversarial Examples
and 683 more not named here
make_layers 59 other papers
50 dated earlier, 7 later, 2 same year, order not determinable
1411.4038 (2014-11)
Fully Convolutional Networks for Semantic Segmentation
1511.00561 (2015-11)
SegNet: A Deep Convolutional Encoder-Decoder Architecture for Image Segmentation
1511.05950 (2015-11)
Staleness-aware Async-SGD for Distributed Deep Learning
and 56 more not named here

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Understanding the training dynamics of deep neural networks is challenging due to their high-dimensional nature and intricate loss landscapes. Recent studies have revealed that, along the training trajectory, the gradient approximately aligns with a low-rank top eigenspace of the training loss Hessian, referred to as the dominant subspace. Given this alignment, this paper explores whether neural networks can be trained within the dominant subspace, which, if feasible, could lead to more efficient training methods. Our primary observation is that when the SGD update is projected onto the dominant subspace, the training loss does not decrease further. This suggests that the observed alignment between the gradient and the dominant subspace is spurious. Surprisingly, projecting out the dominant subspace proves to be just as effective as the original update, despite removing the majority of the original update component. We observe similar behavior across practical setups, including the large learning rate regime (also known as Edge of Stability), Sharpness-Aware Minimization, momentum, and adaptive optimizers. We discuss the main causes and implications of this spurious alignment, shedding light on the dynamics of neural network training.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2405.16002")
get_code_for_paper("2405.16002")
have("2405.16002")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2405.16002.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2405.16002.svg)](https://syntology.ai/paper/2405.16002)

Connect an agent — have() is free.