SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2106.01613 · ICLR · 2022

NODE-GAM: Neural Generalized Additive Model for Interpretable Deep Learning

Rich Caruana, Anna Goldenberg, Chun-Hao Chang

arXiv · PDF · Open in the Atlas

Code that ran

9 of 15 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (5 of them constructed an object rather than computing a result; 1 honoured the contract we drafted, 0 violated it and 7 have no contract verdict; 1 failure of our instrument; 6 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/2106.01613.json · badge: /api/ran/2106.01613.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
zzzace2000/nodegam canonical (classifier label)
classifier basis: classifier_2025
7 ran · 5 no recorded run
enouenj/sparse-interaction-additive-networks community
Papers with Code archive
2 ran · 1 no recorded run
FunctionStatusWhere it lives
Entmax15Function Ran
ran on a synthesized input · constructed an object; no result was computed
zzzace2000/nodegam/nodegam/arch.py
code served (permissive licence) · get_code("b6ad3ad95af90645")
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 an instance of its own class, Entmax15Function: the run constructed the object and did not call it, so no result was computed. 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
from torch.autograd import Function

torch.manual_seed(0)

# The function takes input tensor and dim parameter
# input: a multi-dimensional tensor
# dim: dimension along which to apply entmax

# Create a small input tensor - batch size 2, feature dim 4
input_tensor = torch.randn(2, 4, requires_grad=True)

# dim parameter - use -1 to operate on last dimension
dim = -1

ARGS = {
    "input": input_tensor,
    "dim": dim,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Entmax15Function — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a Entmax15Function 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
FeedForwardDNN Ran
ran on a synthesized input
enouenj/sparse-interaction-additive-networks/src/sian/models/models.py
pointer only (licence: NONE) · get_code("c484ea17a487fd2a")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, enouenj/sparse-interaction-additive-networks 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.

GAM_ODST Ran
ran on a synthesized input · constructed an object; no result was computed
zzzace2000/nodegam/nodegam/arch.py
code served (permissive licence) · get_code("98280132e17d7ecb")
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 an instance of its own class, GAM_ODST: the run constructed the object and did not call it, so no result was computed. 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 torch.nn as nn
import numpy as np

torch.manual_seed(0)

ARGS = {
    'in_features': 4,
    'num_trees': 2,
    'depth': 3,
    'tree_dim': 1,
    'choice_function': lambda input, dim=-1: input,
    'bin_function': lambda input: input,
    'initialize_response_': nn.init.normal_,
    'initialize_selection_logits_': nn.init.uniform_,
    'threshold_init_beta': 1.0,
    'threshold_init_cutoff': 1.0,
    'colsample_bytree': 1.0,
}
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 GAM_ODST — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a GAM_ODST 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
ModuleWithInit Ran
ran on a synthesized input · constructed an object; no result was computed
zzzace2000/nodegam/nodegam/arch.py
code served (permissive licence) · get_code("eda16069762b4af9")
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 an instance of its own class, ModuleWithInit: the run constructed the object and did not call it, so no result was computed. 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 torch.nn as nn

torch.manual_seed(0)

class ConcreteModule(nn.Module):
    """Concrete implementation of ModuleWithInit for testing."""
    
    def __init__(self):
        super().__init__()
        self.weight = None
    
    def initialize(self, *args, **kwargs):
        """Initialize weight on first batch."""
        if self.weight is None:
            self.weight = nn.Parameter(torch.randn(4, 3))
    
    def forward(self, x):
        """Forward pass using initialized weight."""
        if self.weight is None:
            return x
        return x @ self.weight

# Create an instance of the concrete module
module = ConcreteModule()

ARGS = {}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ModuleWithInit — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a ModuleWithInit 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
ODST Ran
ran on a synthesized input · constructed an object; no result was computed
zzzace2000/nodegam/nodegam/arch.py
code served (permissive licence) · get_code("c9033c4ad6b84b0c")
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 an instance of its own class, ODST: the run constructed the object and did not call it, so no result was computed. 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 torch.nn as nn
import torch.nn.functional as F

torch.manual_seed(0)

ARGS = {
    'in_features': 4,
    'num_trees': 3,
    'depth': 2,
    'tree_dim': 2,
    'choice_function': lambda input, dim=-1: input,
    'bin_function': lambda input: input,
    'initialize_response_': nn.init.normal_,
    'initialize_selection_logits_': nn.init.uniform_,
    'threshold_init_beta': 1.0,
    'threshold_init_cutoff': 1.0,
    'colsample_bytree': 1.0,
}
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 ODST — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a ODST 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
SmallModel Ran
ran on a synthesized input
enouenj/sparse-interaction-additive-networks/src/sian/models/models.py
pointer only (licence: NONE) · get_code("f84da3894ffcc730")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, enouenj/sparse-interaction-additive-networks 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.

Timer Ran
ran on a synthesized input · constructed an object; no result was computed
zzzace2000/nodegam/nodegam/arch.py
code served (permissive licence) · get_code("328250adf4274596")
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 an instance of its own class, Timer: the run constructed the object and did not call it, so no result was computed. 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 time

ARGS = {
    'name': 'model training',
    'remove_start_msg': True,
}
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 Timer — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a Timer 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
_make_ix_like Ran
ran; our contract draft was wrong, not the code
zzzace2000/nodegam/nodegam/arch.py
code served (permissive licence) · get_code("d5065b01c734ab2a")
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 Tensor, ndim 4, dtype family float, shape (1, 4, 1, 1) 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

torch.manual_seed(0)

# Construct a small input tensor with distinct dimensions
input_tensor = torch.randn(2, 4, 3, 5)

# The function uses:
# - input.size(dim) to get the dimension size
# - input.device to get the device
# - input.dtype to get the dtype
# - input.dim() to get the number of dimensions
# - All of these are standard tensor attributes/methods

ARGS = {
    "input": input_tensor,
    "dim": 1
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
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 4, dtype family float, shape (1, 4, 1, 1) — shape (1, 4, 1, 1) 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
check_numpy Ran
ran, honoured the contract we drafted
zzzace2000/nodegam/nodegam/arch.py
code served (permissive licence) · get_code("f43369bac5c86b96")
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 ndarray, 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 torch

torch.manual_seed(0)

# The function check_numpy takes a single parameter x
# It handles: torch.Tensor, lists/array-like objects, and numpy arrays
# We'll test with a torch.Tensor as the primary case since that's what the function explicitly handles

x_tensor = torch.randn(2, 4)

ARGS = {
    "x": x_tensor
}
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 ndarray, 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
Blocksparse_Deep_Relu_GAM Not yet run
unverified (harvested, no recorded run)
enouenj/sparse-interaction-additive-networks/src/sian/models/models.py
pointer only (licence: NONE) · get_code("b507292035444fc2")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, enouenj/sparse-interaction-additive-networks 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.

Entmoid15 Not yet run
unverified (harvested, no recorded run)
zzzace2000/nodegam/nodegam/arch.py
code served (permissive licence) · get_code("36b224091f68a938")
How it ran

no invocation recorded.

GAMAdditiveMixin Not yet run
unverified (harvested, no recorded run)
zzzace2000/nodegam/nodegam/arch.py
code served (permissive licence) · get_code("dc7e97327b2af44f")
How it ran

no invocation recorded.

GAMBlock Not yet run
unverified (harvested, no recorded run)
zzzace2000/nodegam/nodegam/arch.py
code served (permissive licence) · get_code("4a3d4c115bd3fa46")
How it ran

no invocation recorded.

ODSTBlock Not yet run
unverified (harvested, no recorded run)
zzzace2000/nodegam/nodegam/arch.py
code served (permissive licence) · get_code("c1c7c834a428ca47")
How it ran

no invocation recorded.

process_in_chunks Not yet run
unverified (harvested, no recorded run)
zzzace2000/nodegam/nodegam/arch.py
code served (permissive licence) · get_code("bc0337a39644d30e")
How it ran

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, 4 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 0 the same month, 2 later) and 11 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
_make_ix_like
brick of this paper's assembly: helper
2 other papers
2 later · 1 only as an assembly brick
2406.00426 (2024-06)
InterpreTabNet: Distilling Predictive Signals from Tabular Data by Salient Feature Interpretation
2604.19547 (2026-04)
Emotion-Cause Pair Extraction in Conversations via Semantic Decoupling and Graph Alignment
check_numpy
brick of this paper's assembly: helper
2 other papers
2 dated earlier
1909.06312 (2019-09)
Neural Oblivious Decision Ensembles for Deep Learning on Tabular Data
2004.00345 (2020-04)
Editable Neural Networks
process_in_chunks
brick of this paper's assembly: helper
2 other papers
2 dated earlier
1909.06312 (2019-09)
Neural Oblivious Decision Ensembles for Deep Learning on Tabular Data
2004.00345 (2020-04)
Editable Neural Networks
Entmax15Function
brick of this paper's assembly: helper
1 other paper
1 later
2406.00426 (2024-06)
InterpreTabNet: Distilling Predictive Signals from Tabular Data by Salient Feature Interpretation

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

Deployment of machine learning models in real high-risk settings (e.g. healthcare) often depends not only on the model's accuracy but also on its fairness, robustness, and interpretability. Generalized Additive Models (GAMs) are a class of interpretable models with a long history of use in these high-risk domains, but they lack desirable features of deep learning such as differentiability and scalability. In this work, we propose a neural GAM (NODE-GAM) and neural GA 2 M (NODE-GA 2 M) that scale well and perform better than other GAMs on large datasets, while remaining interpretable compared to other ensemble and deep learning models. We demonstrate that our models find interesting patterns in the data. Lastly, we show that we improve model accuracy via self-supervised pre-training, an improvement that is not possible for non-differentiable GAMs.

For agents

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

get_harvested_code_for_paper("2106.01613")
get_code_for_paper("2106.01613")
have("2106.01613")

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

curl https://syntology.ai/api/ran/2106.01613.json

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

[![Syntology run record](https://syntology.ai/api/ran/2106.01613.svg)](https://syntology.ai/paper/2106.01613)

Connect an agent — have() is free.