SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2510.12503 · ICLR · 2025

The robustness of differentiable Causal Discovery in misspecified Scenarios

He Wang, Mingyu Kang, Huiyang Yi, Yanyan He, Duxin Chen, Wenwu Yu

arXiv · PDF · Open in the Atlas

Code that ran

12 of 13 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 0 have no contract verdict; 11 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/2510.12503.json · badge: /api/ran/2510.12503.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
fishmoon1234/DAG-NoCurl canonical (classifier label)
classifier basis: census
6 ran · 0 no recorded run
CausalDisco/CausalDisco canonical (classifier label)
classifier basis: census
6 ran · 0 no recorded run
copy not recorded community
not in the paper's declared repositories
0 ran · 1 no recorded run
FunctionStatusWhere it lives
elu Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
fishmoon1234/DAG-NoCurl/BPR.py
code served (permissive licence) · get_code("2d1e7d81125cf16e")
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

# Construct arguments for the elu function
x = np.array([[-2.0, -1.0, 0.0, 1.0],
              [2.0, -0.5, 0.5, -1.5]], dtype=np.float32)

derivative = False
alpha = 0.1

ARGS = {
    'x': x,
    'derivative': derivative,
    'alpha': alpha
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — 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
order_alignment_adjacent_pairs Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
CausalDisco/CausalDisco/CausalDisco/analytics.py
code served (permissive licence) · get_code("3970791be77e4b8e")
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 float64, ndim 0, 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 numpy as np

# Create a small binary adjacency matrix (4x4)
# This represents a DAG with a few edges
B = np.array([
    [0, 1, 0, 0],
    [0, 0, 1, 0],
    [0, 0, 0, 1],
    [0, 0, 0, 0]
], dtype=int)

# Create scores array matching the dimension of B
scores = np.array([0.1, 0.5, 0.8, 0.9], dtype=float)

# Tolerance parameter
tol = 0.0

ARGS = {
    "B": B,
    "scores": scores,
    "tol": tol
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
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
order_alignment_paths Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
CausalDisco/CausalDisco/CausalDisco/analytics.py
code served (permissive licence) · get_code("b349ed375942004d")
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 float64, ndim 0, 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 numpy as np

# Construct valid arguments for order_alignment_paths
np.random.seed(0)

# Create a small DAG adjacency matrix (4x4)
# A valid DAG should have no cycles
W = np.array([
    [0, 1, 1, 0],
    [0, 0, 1, 1],
    [0, 0, 0, 1],
    [0, 0, 0, 0]
], dtype=float)

# Create a scores vector (4 entries, one per node)
scores = np.array([0.1, 0.3, 0.5, 0.9], dtype=float)

# Tolerance parameter
tol = 0.1

ARGS = {
    "W": W,
    "scores": scores,
    "tol": tol
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
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
r2coeff Ran
ran, honoured the contract we drafted
CausalDisco/CausalDisco/CausalDisco/analytics.py
code served (permissive licence) · get_code("83a544e51ab9e249")
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 1, dtype family float, shape (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
from scipy import linalg
from sklearn.linear_model import LinearRegression

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

# Create a small valid dataset
# X should be d × n (note: different from typical sklearn convention)
# d = number of variables, n = number of samples
d = 4  # number of variables
n = 3  # number of samples

# Create a random data matrix with shape (d, n)
X = np.random.randn(d, n)

ARGS = {
    "X": X
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy, sklearn — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family float, shape (4,) — shape (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
random_sort_regress Ran
ran; our contract draft was wrong, not the code
CausalDisco/CausalDisco/CausalDisco/baselines.py
code served (permissive licence) · get_code("9d66114e8b2c8090")
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 (4, 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

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

# Create small data matrix X with shape (n_samples, n_features)
# Using n=5 samples and d=4 features for a tiny but valid setup
n_samples = 5
n_features = 4
X = np.random.randn(n_samples, n_features)

# Define ARGS dictionary with all parameters for random_sort_regress
ARGS = {
    'X': X,
    'seed': 42
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn — 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 (4, 4) — shape (4, 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
setup_logger Ran
ran; our contract draft was wrong, not the code
fishmoon1234/DAG-NoCurl/utils.py
code served (permissive licence) · get_code("f33787df9c8f67b4")
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 Logger. 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 logging

ARGS = {
    'mode': 'debug'
}
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 Logger — a Logger output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Logger 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
sigmoid Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
fishmoon1234/DAG-NoCurl/BPR.py
code served (permissive licence) · get_code("d39a275f8a95ded5")
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 1, dtype family float, shape (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

# The sigmoid function takes two parameters: x and derivative
# x should be a numeric input (numpy array, scalar, etc.)
# derivative is a boolean flag

# Create a small numpy array for x
x_input = np.array([0.0, 1.0, -1.0, 0.5])

# Create ARGS dict with all parameters
ARGS = {
    "x": x_input,
    "derivative": False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family float, shape (4,) — shape (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
simulate_random_dag Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
fishmoon1234/DAG-NoCurl/utils.py
code served (permissive licence) · get_code("5a8fa66922d5bfb9")
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 DiGraph. 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 numpy as np
import networkx as nx

np.random.seed(0)

ARGS = {
    'd': 5,
    'degree': 2.5,
    'graph_type': 'erdos-renyi',
    'w_range': (0.5, 2.0)
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
networkx, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DiGraph — a DiGraph output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DiGraph output the recorded ndim 1 and dtype family int 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
simulate_sem_nonlinear Ran
ran; our contract draft was wrong, not the code
fishmoon1234/DAG-NoCurl/utils.py
code served (permissive licence) · get_code("e9a02ce9615c9da6")
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 3, dtype family float, shape (2, 3, 3) 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 networkx as nx

# Create a small weighted DAG
G = nx.DiGraph()
G.add_weighted_edges_from([
    (0, 1, 0.5),
    (0, 2, 0.3),
    (1, 2, 0.7),
])

# Define arguments
ARGS = {
    "G": G,
    "n": 2,
    "x_dims": 3,
    "sem_type": "linear-gauss",
    "linear_type": "linear",
    "noise_scale": 1.0,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
networkx, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 3, dtype family float, shape (2, 3, 3) — shape (2, 3, 3) 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
softrelu Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
fishmoon1234/DAG-NoCurl/BPR.py
code served (permissive licence) · get_code("caea6fcf470e6e1e")
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

# Examine the function: softrelu takes x, derivative (bool), and alpha (float)
# The function uses x in numpy operations (log, exp)
# derivative parameter controls which branch is taken
# alpha parameter is defined but not used in the function body

# Create a small numpy array for x
x_val = np.array([[-0.5, 0.0, 0.5, 1.0], 
                   [0.2, -0.1, 0.3, -0.2]], dtype=np.float32)

ARGS = {
    "x": x_val,
    "derivative": False,
    "alpha": 0.1
}
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
sort_regress Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
CausalDisco/CausalDisco/CausalDisco/baselines.py
code served (permissive licence) · get_code("6e174813119d541f")
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 (4, 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

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

# Create small sample data
n_samples = 10  # number of observations
n_features = 4  # number of variables (d)

# X: n x d matrix of data
X = np.random.randn(n_samples, n_features)

# scores: vector of d scores used to determine ordering
scores = np.random.randn(n_features)

ARGS = {
    "X": X,
    "scores": scores,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn — 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 (4, 4) — shape (4, 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
var_sort_regress Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
CausalDisco/CausalDisco/CausalDisco/baselines.py
code served (permissive licence) · get_code("7b9b0e44e8ab6838")
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 (4, 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

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

# Create small synthetic data for var_sort_regress
# X should be n x d (n samples, d variables)
# Using n=10 samples, d=4 variables
n_samples = 10
n_variables = 4

X = np.random.randn(n_samples, n_variables).astype(np.float64)

ARGS = {
    "X": X,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn — 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 (4, 4) — shape (4, 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
notears_linear Not yet run
unverified (harvested, no recorded run)
this paper's copy was not recorded; identical code first harvested from xunzheng/notears
pointer only · get_code("8121d0e3bc3faff7")
How it ran

The body of this sample is withheld: this paper's copy of the code was not recorded (its link names no repository or source path), so no licence was read for it. 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 13 samples lifted from this paper's linked repositories, 13 carry a function identity and 0 do not. Of the 13 distinct functions with an identity, 1 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 0 later) and 12 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
notears_linear 8 other papers
8 dated earlier
1803.01422 (2018-03)
DAGs with NO TEARS: Continuous Optimization for Structure Learning
1909.13189 (2019-09)
Learning Sparse Nonparametric DAGs
2006.10201 (2020-06)
On the Role of Sparsity and DAG Constraints for Learning Linear DAGs
and 5 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

Causal discovery aims to learn causal relationships between variables from targeted data, making it a fundamental task in machine learning. However, causal discovery algorithms often rely on unverifiable causal assumptions, which are usually difficult to satisfy in real-world data, thereby limiting the broad application of causal discovery in practical scenarios. Inspired by these considerations, this work extensively benchmarks the empirical performance of various mainstream causal discovery algorithms, which assume i.i.d. data, under eight model assumption violations. Our experimental results show that differentiable causal discovery methods exhibit robustness under the metrics of Structural Hamming Distance and Structural Intervention Distance of the inferred graphs in commonly used challenging scenarios, except for scale variation. We also provide the theoretical explanations for the performance of differentiable causal discovery methods. Finally, our work aims to comprehensively benchmark the performance of recent differentiable causal discovery methods under model assumption violations, and provide the standard for reasonable evaluation of causal discovery, as well as to further promote its application in real-world scenarios.

For agents

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

get_harvested_code_for_paper("2510.12503")
get_code_for_paper("2510.12503")
have("2510.12503")

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

curl https://syntology.ai/api/ran/2510.12503.json

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

[![Syntology run record](https://syntology.ai/api/ran/2510.12503.svg)](https://syntology.ai/paper/2510.12503)

Connect an agent — have() is free.