SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2006.00555 · 2020

Transferring Inductive Biases through Knowledge Distillation

arXiv · PDF · Open in the Atlas

Code that ran

7 of 7 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; 1 honoured the contract we drafted, 0 violated it and 4 have no contract verdict; 2 failures of our instrument; 0 no recorded run), 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/2006.00555.json · badge: /api/ran/2006.00555.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
samiraabnar/Reflect canonical (classifier label)
classifier basis: census
7 ran · 0 no recorded run
FunctionStatusWhere it lives
distill_loss Ran
ran on a synthesized input
samiraabnar/Reflect/tf2_models/metrics.py
code served (permissive licence) · get_code("02c6d20363dfb076")
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 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 tensorflow as tf
import numpy as np

# Set seed for reproducibility
np.random.seed(0)
tf.random.set_seed(0)

# Create y_true: labels for softmax cross entropy
# Shape: (batch_size, num_classes) with one-hot encoding
# Using batch_size=2, num_classes=4
y_true = tf.constant([
    [1.0, 0.0, 0.0, 0.0],
    [0.0, 1.0, 0.0, 0.0]
], dtype=tf.float32)

# Create y_pred: logits from model
# Shape: (batch_size, num_classes) = (2, 4)
y_pred = tf.constant([
    [2.1, 0.5, 0.3, 0.1],
    [0.2, 1.8, 0.4, 0.6]
], dtype=tf.float32)

# Create tmp: temperature parameter (scalar)
tmp = tf.constant(0.5, dtype=tf.float32)

ARGS = {
    "y_true": y_true,
    "y_pred": y_pred,
    "tmp": tmp
}
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 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
gelu Ran
ran, honoured the contract we drafted
samiraabnar/Reflect/tf2_models/common_layers.py
code served (permissive licence) · get_code("47451fbef579f4fc")
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
x = tf.constant([[-0.5, 0.0, 0.5, 1.0], [0.2, -0.3, 0.8, -0.1]], 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
get_initializer Ran
ran; our contract draft was wrong, not the code
samiraabnar/Reflect/tf2_models/common_layers.py
code served (permissive licence) · get_code("488a1d723fa95280")
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 TruncatedNormal. 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 tensorflow as tf

ARGS = {
    "initializer_range": 0.02
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type TruncatedNormal — a TruncatedNormal output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a TruncatedNormal 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
masked_sequence_loss Ran
ran on a synthesized input
samiraabnar/Reflect/tf2_models/metrics.py
code served (permissive licence) · get_code("0800e43230ecb35d")
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 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 tensorflow as tf
import numpy as np

# Set seed for reproducibility
np.random.seed(0)
tf.random.set_seed(0)

# Create small test tensors
batch_size = 2
sequence_length = 4
num_classes = 3

# y_true: shape (batch_size, sequence_length) with class indices
# Values should be in range [0, num_classes) or padding_symbol
y_true = tf.constant([
    [1, 2, 1, 0],
    [2, 1, 0, 0]
], dtype=tf.int32)

# y_pred: shape (batch_size, sequence_length, num_classes) - logits
# sparse_softmax_cross_entropy_with_logits expects logits of shape (batch, seq_len, num_classes)
y_pred = tf.random.normal((batch_size, sequence_length, num_classes), dtype=tf.float32)

# padding_symbol: the value that represents padding in y_true
padding_symbol = 0

ARGS = {
    "y_true": y_true,
    "y_pred": y_pred,
    "padding_symbol": padding_symbol
}
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 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
max_out Ran
ran on a synthesized input
samiraabnar/Reflect/tf2_models/cnn.py
code served (permissive licence) · get_code("3686362b18aeccb8")
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 with shape (batch_size, features)
# The function expects num_channels (features dimension) to be divisible by num_units
# Let's use batch_size=2, num_channels=8 (divisible by num_units=4)
inputs = tf.constant(np.random.randn(2, 8).astype(np.float32))

# num_units should divide evenly into the number of channels
num_units = 4

# axis can be None or an integer specifying which axis to apply max-out on
axis = -1

ARGS = {
    'inputs': inputs,
    'num_units': num_units,
    'axis': axis,
}
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
sequence_distill_loss Ran
ran on a synthesized input
samiraabnar/Reflect/tf2_models/metrics.py
code served (permissive licence) · get_code("7629a5fc4fbe87b4")
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 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 tensorflow as tf
import numpy as np

# Set random seeds for reproducibility
tf.random.set_seed(0)
np.random.seed(0)

# Create small test tensors
batch_size = 2
sequence_length = 4
num_classes = 3
padding_symbol = 2

# y_true: one-hot encoded labels with shape (batch_size, sequence_length, num_classes)
# The function expects y_true to have a dimension at index `padding_symbol`
y_true = tf.constant([
    [[1.0, 0.0, 0.0], [0.0, 1.0, 0.0], [0.0, 0.0, 1.0], [1.0, 0.0, 0.0]],
    [[0.0, 1.0, 0.0], [1.0, 0.0, 0.0], [0.0, 1.0, 0.0], [0.0, 0.0, 1.0]]
], dtype=tf.float32)

# y_pred: logits with shape (batch_size, sequence_length, num_classes)
y_pred = tf.constant([
    [[0.5, -0.2, 1.0], [-0.3, 0.8, 0.2], [0.1, 0.3, -0.5], [0.9, -0.1, 0.3]],
    [[0.2, 0.7, -0.1], [0.4, -0.2, 0.5], [-0.1, 0.6, 0.2], [0.3, 0.2, 0.8]]
], dtype=tf.float32)

# padding_symbol: index used to identify padding in y_true
padding_symbol = 2

# tmp: temperature parameter for distillation
tmp = 2.0

ARGS = {
    "y_true": y_true,
    "y_pred": y_pred,
    "padding_symbol": padding_symbol,
    "tmp": tmp
}
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 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
shape_list Ran
ran; our contract draft was wrong, not the code
samiraabnar/Reflect/tf2_models/common_layers.py
code served (permissive licence) · get_code("6220c13c591c68a7")
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 int; 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 tensorflow as tf
import numpy as np

# Create a small TensorFlow tensor with some dynamic dimensions
# The function uses x.shape.as_list() and tf.shape(x)
x = tf.constant(np.random.randn(2, 4, 3), 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 list, ndim 1, dtype family int — 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

Also present in other papers' code

Of 7 samples lifted from this paper's linked repositories, 7 carry a function identity and 0 do not. Of the 7 distinct functions with an identity, 3 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 1 the same month, 2 later) and 4 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
gelu 26 other papers
16 dated earlier, 10 later · 4 only as an assembly brick
1706.03762 (2017-06)
Attention Is All You Need
1810.04805 (2018-10)
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
1901.02860 (2019-01)
Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context
and 23 more not named here
get_initializer 6 other papers
5 dated earlier, 1 the same month · 1 only as an assembly brick
1810.04805 (2018-10)
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
1909.05017 (2019-09)
Question Generation by Transformers
1909.11942 (2019-09)
ALBERT: A Lite BERT for Self-supervised Learning of Language Representations
and 3 more not named here
shape_list 3 other papers
2 dated earlier, 1 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
2601.10926 (2026-01)
Selecting Language Models for Social Science: Start Small, Start Open, and Validate Preprint XX(X):1-22 ©The Author(s) 2026

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

Having the right inductive biases can be crucial in many tasks or scenarios where data or computing resources are a limiting factor, or where training data is not perfectly representative of the conditions at test time. However, defining, designing and efficiently adapting inductive biases is not necessarily straightforward. In this paper, we explore the power of knowledge distillation for transferring the effect of inductive biases from one model to another. We consider families of models with different inductive biases, LSTMs vs. Transformers and CNNs vs. MLPs, in the context of tasks and scenarios where having the right inductive biases is critical. We study the effect of inductive biases on the solutions the models converge to and investigate how and to what extent the effect of inductive biases is transferred through knowledge distillation, in terms of not only performance but also different aspects of converged solutions.

For agents

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

get_harvested_code_for_paper("2006.00555")
get_code_for_paper("2006.00555")
have("2006.00555")

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

curl https://syntology.ai/api/ran/2006.00555.json

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

[![Syntology run record](https://syntology.ai/api/ran/2006.00555.svg)](https://syntology.ai/paper/2006.00555)

Connect an agent — have() is free.