SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2006.04884 · ICLR · 2021

On the Stability of Fine-tuning BERT: Misconceptions, Explanations, and Strong Baselines

Dietrich Klakow, Marius Mosbach, Maksym Andriushchenko

arXiv · PDF · Open in the Atlas

Code that ran

13 of 20 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; 4 honoured the contract we drafted, 0 violated it and 5 have no contract verdict; 4 failures of our instrument; 7 no recorded run); 2 of them were also run on the host, not in the jail, by the peer-agreement lane that set their level, measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; 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/2006.04884.json · badge: /api/ran/2006.04884.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
uds-lsv/bert-stable-fine-tuning canonical (classifier label)
classifier basis: census
8 ran · 2 no recorded run
ruanchaves/eplm community (classifier label)
classifier basis not recorded on the link
5 ran · 5 no recorded run
FunctionStatusWhere it lives
swish Ran
ran; our contract draft was wrong, not the code
uds-lsv/bert-stable-fine-tuning/src/transformers/activations.py
code served (permissive licence) · get_code("0f786c407fb1ee4c")
How it ran

Ran in Syntology's jail on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned Tensor, ndim 2, dtype family float, shape (2, 4) as a later run recorded it. The image it ran in was not recorded. The comparison behind this sample's level is a different run: the peer-agreement lane ran it again on the host, not in the jail (report dated 2026-09-04T22:33:52Z). If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# The function swish takes a single parameter 'x' which is used with:
# - torch.sigmoid(x) - requires a tensor
# - x * ... - requires a tensor that supports multiplication
# So x should be a tensor

x = torch.randn(2, 4)

ARGS = {
    "x": x
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
whitespace_tokenize Ran
ran; our contract draft was wrong, not the code
uds-lsv/bert-stable-fine-tuning/src/transformers/tokenization_bert.py
code served (permissive licence) · get_code("cf9ffa02a42184af")
How it ran

Ran in Syntology's jail on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. The comparison behind this sample's level is a different run: the peer-agreement lane ran it again on the host, not in the jail (report dated 2026-09-05T02:55:57Z). If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

def whitespace_tokenize(text):
    """Runs basic whitespace cleaning and splitting on a piece of text."""
    text = text.strip()
    if not text:
        return []
    tokens = text.split()
    return tokens

# Setup code
torch.manual_seed(0)

# The function takes a single parameter: text (a string)
# It performs basic whitespace cleaning and splitting
# We need to provide a valid string argument

ARGS = {
    "text": "  hello world  this is a test  "
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
add_dataset Ran
ran on a synthesized input
ruanchaves/eplm/leaderboard/manage_data.py
code served (permissive licence) · get_code("a6f1e11c54df3de5")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned dict. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np
from typing import Dict, Any

torch.manual_seed(0)

# Initialize data structure with datasets dict
data = {'datasets': {}}

dataset_name = 'cifar10'
name = 'CIFAR-10'
description = 'A dataset of 60000 32x32 color images in 10 classes'
tasks = ['image_classification', 'transfer_learning']
url = 'https://www.cs.toronto.edu/~kriz/cifar.html'

ARGS = {
    'data': data,
    'dataset_name': dataset_name,
    'name': name,
    'description': description,
    'tasks': tasks,
    'url': url
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
check_dependency Ran
ran on a synthesized input
ruanchaves/eplm/leaderboard/run_app.py
code served (permissive licence) · get_code("d87861334682866a")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned bool, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import importlib.util

def check_dependency(package_name):
    """Check if a package is installed."""
    spec = importlib.util.find_spec(package_name)
    return spec is not None

# The function takes a single parameter: package_name (a string)
# It checks if a package with that name is installed using importlib.util.find_spec
# We'll test with a package that definitely exists (like 'sys' or 'os')

ARGS = {
    "package_name": "sys"
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type bool, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
find_json_files Ran
ran on a synthesized input
ruanchaves/eplm/leaderboard/extract_portuguese_leaderboard.py
code served (permissive licence) · get_code("6e1f4e02403886e3")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import json
import tempfile
from pathlib import Path

# Create a temporary directory structure with valid JSON files
temp_dir = tempfile.mkdtemp()
temp_path = Path(temp_dir)

# Create a valid JSON file with dict content
valid_json_1 = temp_path / "config.json"
with open(valid_json_1, 'w') as f:
    json.dump({"key": "value", "number": 42}, f)

# Create another valid JSON file with dict content
valid_json_2 = temp_path / "settings.json"
with open(valid_json_2, 'w') as f:
    json.dump({"nested": {"inner": "data"}}, f)

# Create a subdirectory with a JSON file
subdir = temp_path / "subdir"
subdir.mkdir()
valid_json_3 = subdir / "data.json"
with open(valid_json_3, 'w') as f:
    json.dump({"array": [1, 2, 3]}, f)

# Create an invalid JSON file (will be skipped)
invalid_json = temp_path / "invalid.json"
with open(invalid_json, 'w') as f:
    f.write("{invalid json content")

# Create a JSON file with list instead of dict (will be skipped)
list_json = temp_path / "list.json"
with open(list_json, 'w') as f:
    json.dump([1, 2, 3], f)

ARGS = {
    "repo_path": str(temp_path)
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
gelu Ran
ran, honoured the contract we drafted
uds-lsv/bert-stable-fine-tuning/src/transformers/modeling_tf_bert.py
code served (permissive licence) · get_code("9bc31d06256e1691")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned EagerTensor, ndim 2, dtype family float, shape (2, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import tensorflow as tf
import numpy as np

# Create a small TensorFlow tensor as input
# The function expects a tensor and applies element-wise operations
x = tf.constant([[-2.0, -1.0, 0.0, 1.0], [2.0, 3.0, -0.5, 0.5]], dtype=tf.float32)

ARGS = {
    "x": x
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
gelu_new Ran
ran, honoured the contract we drafted
uds-lsv/bert-stable-fine-tuning/src/transformers/activations.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from lonePatient/BERT-SDA, is pointer only (licence: NONE) · get_code("77601724cad03f95")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from lonePatient/BERT-SDA, and GitHub detected no licence in that repository 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.

gelu_new Ran
ran, honoured the contract we drafted
uds-lsv/bert-stable-fine-tuning/src/transformers/modeling_tf_bert.py
code served (permissive licence) · get_code("83257a2b015273fa")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned EagerTensor, ndim 2, dtype family float, shape (2, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np
import tensorflow as tf

# Create a small TensorFlow tensor for testing
x = tf.constant([[1.0, 2.0, 0.5, -1.0],
                 [0.1, -0.5, 1.5, 0.0]], dtype=tf.float32)

ARGS = {
    "x": x
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
is_valid_json_file Ran
ran on a synthesized input
ruanchaves/eplm/leaderboard/extract_portuguese_leaderboard.py
code served (permissive licence) · get_code("e0c0f0759208e777")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned bool, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import json
import tempfile
import os

# Create a temporary valid JSON file with a dict
temp_dir = tempfile.mkdtemp()
valid_json_file = os.path.join(temp_dir, "valid.json")
with open(valid_json_file, 'w', encoding='utf-8') as f:
    json.dump({"key": "value", "nested": {"data": 123}}, f)

ARGS = {
    "file_path": valid_json_file
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type bool, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
load_vocab Ran
ran; our contract draft was wrong, not the code
uds-lsv/bert-stable-fine-tuning/src/transformers/tokenization_bert.py
code served (permissive licence) · get_code("e7fbc7a74a3457c7")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned OrderedDict. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import collections
import tempfile
import os

# Create a temporary vocabulary file
temp_dir = tempfile.mkdtemp()
vocab_file = os.path.join(temp_dir, 'vocab.txt')

# Write some sample tokens to the vocab file
sample_tokens = ['[PAD]', '[UNK]', '[CLS]', '[SEP]', 'the', 'quick', 'brown', 'fox']
with open(vocab_file, 'w', encoding='utf-8') as f:
    for token in sample_tokens:
        f.write(token + '\n')

ARGS = {
    'vocab_file': vocab_file
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type OrderedDict — a OrderedDict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a OrderedDict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
mish Ran
ran; our contract draft was wrong, not the code
uds-lsv/bert-stable-fine-tuning/src/transformers/modeling_bert.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from llyx97/sparse-and-robust-PLM, is pointer only (licence: NONE) · get_code("060527bb34794b59")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from llyx97/sparse-and-robust-PLM, and GitHub detected no licence in that repository 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.

swish Ran
ran, honoured the contract we drafted
uds-lsv/bert-stable-fine-tuning/src/transformers/modeling_tf_bert.py
code served (permissive licence) · get_code("86d13b9f67274f19")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned EagerTensor, ndim 2, dtype family float, shape (2, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import tensorflow as tf
import numpy as np

# Create a small TensorFlow tensor for testing
x = tf.constant([[1.0, 2.0, 3.0, 4.0],
                 [0.5, 1.5, 2.5, 3.5]], dtype=tf.float32)

ARGS = {
    'x': x
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
validate_yaml_structure Ran
ran on a synthesized input
ruanchaves/eplm/leaderboard/manage_data.py
code served (permissive licence) · get_code("03c4a8263ac6e7a1")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned bool, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

ARGS = {
    'data': {
        'datasets': {
            'imagenet': {
                'name': 'ImageNet',
                'description': 'Large-scale visual recognition challenge dataset',
                'tasks': ['image_classification'],
                'url': 'https://www.image-net.org/'
            },
            'cifar10': {
                'name': 'CIFAR-10',
                'description': 'Labeled subset of the 80 Million Tiny Images dataset',
                'tasks': ['image_classification'],
                'url': 'https://www.cs.toronto.edu/~kriz/cifar.html'
            }
        },
        'benchmark_results': {
            'imagenet': {
                'resnet50': {
                    'accuracy': 0.76,
                    'top5_accuracy': 0.93,
                    'inference_time': 25.3
                },
                'vgg16': {
                    'accuracy': 0.71,
                    'top5_accuracy': 0.90,
                    'inference_time': 45.2
                }
            },
            'cifar10': {
                'resnet18': {
                    'accuracy': 0.95,
                    'loss': 0.18
                }
            }
        },
        'model_metadata': {
            'resnet50': {
                'parameters': 25557032,
                'architecture': 'ResNet',
                'base_model': 'ResNet-50',
                'task': 'image_classification'
            },
            'vgg16': {
                'parameters': 138357544,
                'architecture': 'VGG',
                'base_model': 'VGG-16',
                'task': 'image_classification'
            },
            'resnet18': {
                'parameters': 11689512,
                'architecture': 'ResNet',
                'base_model': 'ResNet-18',
                'task': 'image_classification'
            }
        }
    }
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type bool, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
add_benchmark_result Not yet run
unverified (harvested, no recorded run)
ruanchaves/eplm/leaderboard/manage_data.py
code served (permissive licence) · get_code("33aeaf51efd2501f")
How it ran

no invocation recorded.

convert_pytorch_checkpoint_to_tf Not yet run
unverified (harvested, no recorded run)
uds-lsv/bert-stable-fine-tuning/src/transformers/convert_bert_pytorch_checkpoint_to_original_tf.py
code served (permissive licence) · get_code("d772ba25f4de22b0")
How it ran

no invocation recorded.

convert_to_completion_format Not yet run
unverified (harvested, no recorded run)
ruanchaves/eplm/src/napolab/napolab.py
code served (permissive licence) · get_code("ef9ba0bf81d816aa")
How it ran

no invocation recorded.

export_napolab_benchmark Not yet run
unverified (harvested, no recorded run)
ruanchaves/eplm/src/napolab/napolab.py
code served (permissive licence) · get_code("02907c137bc813d1")
How it ran

no invocation recorded.

extract_data_from_json Not yet run
unverified (harvested, no recorded run)
ruanchaves/eplm/leaderboard/extract_portuguese_leaderboard.py
code served (permissive licence) · get_code("653c7bc811a7f863")
How it ran

no invocation recorded.

load_napolab_benchmark Not yet run
unverified (harvested, no recorded run)
ruanchaves/eplm/src/napolab/napolab.py
code served (permissive licence) · get_code("8db96d079a008116")
How it ran

no invocation recorded.

load_tf_weights_in_bert Not yet run
unverified (harvested, no recorded run)
uds-lsv/bert-stable-fine-tuning/src/transformers/modeling_bert.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from llyx97/sparse-and-robust-PLM, is pointer only (licence: NONE) · get_code("0766f48ad6e01b56")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from llyx97/sparse-and-robust-PLM, and GitHub detected no licence in that repository 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 20 samples lifted from this paper's linked repositories, 20 carry a function identity and 0 do not. Of the 20 distinct functions with an identity, 10 are also present in other papers whose code we hold (7 with at least one paper dated earlier, 4 the same month, 10 later) and 10 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
swish 206 other papers
39 dated earlier, 2 the same month, 164 later, 1 same year, order not determinable · 44 only as an assembly brick
1408.5882 (2014-08)
Convolutional Neural Networks for Sentence Classification
1805.12114 (2018-05)
Deep Reinforcement Learning in a Handful of Trials using Probabilistic Dynamics Models
1809.01357 (2018-09)
Zero Shot Learning for Code Education: Rubric Sampling with Deep Learning Inference
and 203 more not named here
whitespace_tokenize 43 other papers
7 dated earlier, 36 later
1408.5882 (2014-08)
Convolutional Neural Networks for Sentence Classification
1810.04805 (2018-10)
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
1906.01698 (2019-06)
Open Sesame: Getting Inside BERT's Linguistic Knowledge
and 40 more not named here
load_vocab 29 other papers
3 dated earlier, 26 later
1810.04805 (2018-10)
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
1909.11942 (2019-09)
ALBERT: A Lite BERT for Self-supervised Learning of Language Representations
2004.09297 (2020-04)
MPNet: Masked and Permuted Pre-training for Language Understanding
and 26 more not named here
gelu_new 23 other papers
7 dated earlier, 1 the same month, 15 later
1810.04805 (2018-10)
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
1909.11942 (2019-09)
ALBERT: A Lite BERT for Self-supervised Learning of Language Representations
1911.03631 (2019-11)
Hierarchical Graph Network for Multi-hop Question Answering
and 20 more not named here
mish 19 other papers
19 later
2007.12223 (2020-07)
The Lottery Ticket Hypothesis for Pre-trained BERT Networks
2012.07463 (2020-12)
Parameter-Efficient Transfer Learning with Diff Pruning
2021.acl-long.390 (2021 (year only))
Few-Shot Text Ranking with Meta Adapted Synthetic Weak Supervision
and 16 more not named here
load_tf_weights_in_bert 16 other papers
16 later
2007.12223 (2020-07)
The Lottery Ticket Hypothesis for Pre-trained BERT Networks
2012.07463 (2020-12)
Parameter-Efficient Transfer Learning with Diff Pruning
2021.acl-long.390 (2021 (year only))
Few-Shot Text Ranking with Meta Adapted Synthetic Weak Supervision
and 13 more not named here
swish 8 other papers
4 dated earlier, 4 later · 2 only as an assembly brick
1810.04805 (2018-10)
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
1902.08408 (2019-02)
PhysNet: A Neural Network for Predicting Energies, Forces, Dipole Moments and Partial Charges
1910.01108 (2019-10)
DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter
and 5 more not named here
gelu 6 other papers
2 dated earlier, 1 the same month, 3 later · 1 only as an assembly brick
1810.04805 (2018-10)
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
1910.01108 (2019-10)
DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter
2006.04558 (2020-06)
FastSpeech 2: Fast and High-Quality End-to-End Text to Speech
and 3 more not named here
gelu_new 6 other papers
2 dated earlier, 1 the same month, 3 later · 1 only as an assembly brick
1810.04805 (2018-10)
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
1910.01108 (2019-10)
DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter
2006.04558 (2020-06)
FastSpeech 2: Fast and High-Quality End-to-End Text to Speech
and 3 more not named here
convert_pytorch_checkpoint_to_tf 1 other paper
1 later
2007.12223 (2020-07)
The Lottery Ticket Hypothesis for Pre-trained BERT Networks

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

Fine-tuning pre-trained transformer-based language models such as BERT has become a common practice dominating leaderboards across various NLP benchmarks. Despite the strong empirical performance of fine-tuned models, fine-tuning is an unstable process: training the same model with multiple random seeds can result in a large variance of the task performance. Previous literature (Devlin et al., 2019;Lee et al., 2020;Dodge et al., 2020) identified two potential reasons for the observed instability: catastrophic forgetting and small size of the fine-tuning datasets. In this paper, we show that both hypotheses fail to explain the fine-tuning instability. We analyze BERT, RoBERTa, and ALBERT, fine-tuned on commonly used datasets from the GLUE benchmark, and show that the observed instability is caused by optimization difficulties that lead to vanishing gradients. Additionally, we show that the remaining variance of the downstream task performance can be attributed to differences in generalization where fine-tuned models with the same training loss exhibit noticeably different test performance. Based on our analysis, we present a simple but strong baseline that makes fine-tuning BERT-based models significantly more stable than the previously proposed approaches. Code to reproduce our results is available online: https://github.com/uds-lsv/bert-stable-fine-tuning.

For agents

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

get_harvested_code_for_paper("2006.04884")
get_code_for_paper("2006.04884")
have("2006.04884")

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

curl https://syntology.ai/api/ran/2006.04884.json

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

[![Syntology run record](https://syntology.ai/api/ran/2006.04884.svg)](https://syntology.ai/paper/2006.04884)

Connect an agent — have() is free.