SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2301.09633 · 2023

Prediction-Powered Inference

arXiv · PDF · Open in the Atlas

Code that ran

12 of 16 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; 5 honoured the contract we drafted, 2 violated it and 0 have no contract verdict; 5 failures of our instrument; 4 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/2301.09633.json · badge: /api/ran/2301.09633.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
aangelopoulos/prediction-powered-inference canonical (classifier label)
classifier basis: classifier_2025
8 ran · 0 no recorded run
aangelopoulos/ppi_py canonical (classifier label)
classifier basis: classifier_2025
1 ran · 4 no recorded run
kclip/r_autoeval_plus extension (classifier label)
classifier basis: classifier_2025
2 ran · 0 no recorded run
copy not recorded community
not in the paper's declared repositories
1 ran · 0 no recorded run
FunctionStatusWhere it lives
WSR_PPI_p_value Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
this paper's copy was not recorded; identical code first harvested from kclip/r_autoeval_plus
pointer only; the stored copy get_code serves, from kclip/r_autoeval_plus, is pointer only (licence: NONE) · get_code("3fcb0d4ac9f94b5b")
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. A run is recorded.

acs_filter Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
aangelopoulos/prediction-powered-inference/census/logistic_utils.py
code served (permissive licence) · get_code("85b650f112d3ffcd")
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 DataFrame. 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 pandas as pd

# Create a small DataFrame with the required columns
np.random.seed(0)

# Create outcome column with values that when transformed will be 0 or 1
# Original values: 1 and 2, after transformation (1 - (x - 1)): 1 and 0
outcome_data = np.array([1.0, 2.0, 1.0, 2.0])

# Create regression feature column with numeric values
reg_feat_data = np.array([0.5, 1.5, 2.5, 3.5])

# Create DataFrame
df = pd.DataFrame({
    'outcome': outcome_data,
    'reg_feature': reg_feat_data
})

ARGS = {
    'df': df,
    'outcome_name': 'outcome',
    'reg_feat_name': 'reg_feature'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, pandas — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DataFrame — a DataFrame output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DataFrame output the recorded ndim 2 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
bentkus_iid Ran
ran, honoured the contract we drafted
aangelopoulos/prediction-powered-inference/ppi.py
code served (permissive licence) · get_code("c11f36381d80b1b9")
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 (2,) 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.stats import binom, norm
from scipy.optimize import brentq

def binomial_iid(N, alpha, muhat):
    def invert_upper_tail(mu):
        return binom.cdf(N * muhat, N, mu) - alpha / 2

    def invert_lower_tail(mu):
        return binom.cdf(N * muhat, N, mu) - (1 - alpha / 2)
    u = brentq(invert_upper_tail, 0, 1)
    l = brentq(invert_lower_tail, 0, 1)
    return np.array([l, u])

def bentkus_iid(N, alpha, muhat):
    return binomial_iid(N, alpha / np.e, muhat)

ARGS = {
    "N": 100,
    "alpha": 0.05,
    "muhat": 0.3
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy — 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 (2,) — shape (2,) 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
binomial_iid Ran
ran, honoured the contract we drafted
aangelopoulos/prediction-powered-inference/ppi.py
code served (permissive licence) · get_code("4180e3b78f99bc9a")
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 (2,) 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.stats import binom, norm
from scipy.optimize import brentq

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

# Define arguments for binomial_iid function
# N: number of trials (integer)
# alpha: significance level (float between 0 and 1)
# muhat: estimated probability (float between 0 and 1)

ARGS = {
    "N": 100,
    "alpha": 0.05,
    "muhat": 0.35
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy — 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 (2,) — shape (2,) 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
clt_iid Ran
ran, violated the contract we drafted
aangelopoulos/prediction-powered-inference/ppi.py
code served (permissive licence) · get_code("b232f82b39915429")
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 (2,) 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.stats import binom, norm

# Construct valid arguments for clt_iid function
# The function takes:
# - x: array-like with shape[0] (sample size), std() method, mean() method
# - alpha: float between 0 and 1 (significance level)

# Create a small sample array
np.random.seed(0)
x = np.random.normal(loc=5.0, scale=2.0, size=30)

# Set alpha to a typical significance level
alpha = 0.05

ARGS = {
    'x': x,
    'alpha': alpha
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy — 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 (2,) — shape (2,) 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_confusion_matrix Ran
ran; our contract draft was wrong, not the code
aangelopoulos/prediction-powered-inference/alphafold/utils.py
code served (permissive licence) · get_code("786b270c6fb127b3")
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, 2) 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 pandas as pd
import numpy as np

# Create a sample dataframe with predictions and true labels
np.random.seed(0)
n_samples = 20

df = pd.DataFrame({
    'predictions': np.random.rand(n_samples),
    'true_labels': np.random.randint(0, 2, n_samples)
})

ARGS = {
    'df': df,
    'pred_col_name': 'predictions',
    'true_column_name': 'true_labels'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, pandas — 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, 2) — shape (2, 2) 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_logical_ci Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
aangelopoulos/prediction-powered-inference/alphafold/utils.py
code served (permissive licence) · get_code("e2c42412c53cbfa6")
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 tuple, 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 numpy as np

# Analyze the function:
# - x_n: used with .size attribute, np.cumsum(), and indexing with [-1]
#        appears to be a 1D numpy array of numeric values
# - N: used in division operations, appears to be a scalar (int or float)

# Create test data
x_n = np.array([0.2, 0.3, 0.1, 0.4, 0.5])  # 1D array of values between 0 and 1
N = 10  # scalar representing total population size

ARGS = {
    "x_n": x_n,
    "N": N,
}
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 tuple, ndim 1, dtype family float — a tuple 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
get_odds_ratio_ci_from_mu_ci Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
aangelopoulos/prediction-powered-inference/alphafold/utils.py
code served (permissive licence) · get_code("683c1e8c7e14a6e5")
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 tuple, 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 numpy as np

# The function takes two confidence intervals for means (mu1 and mu0)
# Each CI is a tuple of (lower_bound, upper_bound)
# The values should be probabilities (between 0 and 1) to make sense in odds ratio context
# We need mu1_ci[0] <= mu1_ci[1] and mu0_ci[0] <= mu0_ci[1]
# We also need values that won't cause division by zero when computing ratio = mu / (1 - mu)
# So we need mu values != 0 and != 1

mu1_ci = (0.2, 0.4)
mu0_ci = (0.1, 0.3)

ARGS = {
    'mu1_ci': mu1_ci,
    'mu0_ci': mu0_ci,
}
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 tuple, ndim 1, dtype family float — a tuple 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
pp_mean_iid_asymptotic Ran
ran, violated the contract we drafted
aangelopoulos/prediction-powered-inference/ppi.py
code served (permissive licence) · get_code("06abf03d289229ef")
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 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 numpy as np
from scipy.stats import binom, norm

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

# Create small sample data
n_labeled = 10
n_unlabeled = 20

# Y_labeled: true labels for labeled data
Y_labeled = np.random.binomial(1, 0.5, n_labeled).astype(float)

# Yhat_labeled: predictions on labeled data
Yhat_labeled = np.random.uniform(0, 1, n_labeled)

# Yhat_unlabeled: predictions on unlabeled data
Yhat_unlabeled = np.random.uniform(0, 1, n_unlabeled)

# alpha: significance level
alpha = 0.05

ARGS = {
    "Y_labeled": Y_labeled,
    "Yhat_labeled": Yhat_labeled,
    "Yhat_unlabeled": Yhat_unlabeled,
    "alpha": alpha
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, scipy — 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
ppi_money Ran
ran, honoured the contract we drafted
kclip/r_autoeval_plus/testing_by_betting.py
pointer only (licence: NONE) · get_code("773cd343dca4fc75")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, kclip/r_autoeval_plus 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.

rectified_p_value Ran
ran, honoured the contract we drafted
aangelopoulos/ppi_py/ppi_py/ppi.py
code served (permissive licence) · get_code("2d2336bd457be717")
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, 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 numpy as np

ARGS = {
    'rectifier': 0.5,
    'rectifier_std': 0.1,
    'imputed_mean': 1.2,
    'imputed_std': 0.3,
    'null': 0,
    'alternative': 'two-sided'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, statsmodels — 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, 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
val_p_value_naive Ran
ran, honoured the contract we drafted
kclip/r_autoeval_plus/testing_by_betting.py
pointer only (licence: NONE) · get_code("cf13375c4da83210")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, kclip/r_autoeval_plus 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.

_calc_lam_glm Not yet run
unverified (harvested, no recorded run)
aangelopoulos/ppi_py/ppi_py/ppi.py
code served (permissive licence) · get_code("e83a98bde0af8f43")
How it ran

no invocation recorded.

ppi_mean_ci Not yet run
unverified (harvested, no recorded run)
aangelopoulos/ppi_py/ppi_py/ppi.py
code served (permissive licence) · get_code("dd0387892f74fb4f")
How it ran

no invocation recorded.

ppi_mean_ci Not yet run
unverified (harvested, no recorded run)
aangelopoulos/ppi_py/ppi_py/ppi.py
code served (permissive licence) · get_code("b9bafa1a1f6c6fe8")
How it ran

no invocation recorded.

ppi_mean_pointestimate Not yet run
unverified (harvested, no recorded run)
aangelopoulos/ppi_py/ppi_py/ppi.py
code served (permissive licence) · get_code("cbf78f3128be74b1")
How it ran

no invocation recorded.

Also present in other papers' code

Of 16 samples lifted from this paper's linked repositories, 16 carry a function identity and 0 do not. Of the 15 distinct functions with an identity, 6 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 6 later) and 9 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
ppi_mean_ci
brick of this paper's assembly: entry
3 other papers
3 later
2309.16598 (2023-09)
Cross-Prediction-Powered Inference
2311.01453 (2023-11)
PPI++: Efficient Prediction-Powered Inference
2405.18379 (2024-05)
A Note on the Prediction-Powered Bootstrap
ppi_mean_pointestimate
brick of this paper's assembly: model
3 other papers
3 later
2309.16598 (2023-09)
Cross-Prediction-Powered Inference
2311.01453 (2023-11)
PPI++: Efficient Prediction-Powered Inference
2405.18379 (2024-05)
A Note on the Prediction-Powered Bootstrap
rectified_p_value 3 other papers
3 later
2309.16598 (2023-09)
Cross-Prediction-Powered Inference
2311.01453 (2023-11)
PPI++: Efficient Prediction-Powered Inference
2405.18379 (2024-05)
A Note on the Prediction-Powered Bootstrap
WSR_PPI_p_value
brick of this paper's assembly: entry
1 other paper
1 later
2505.18659 (2025-05)
Adaptive Prediction-Powered AutoEval with Reliability and Efficiency Guarantees
ppi_money
brick of this paper's assembly: loss
1 other paper
1 later
2505.18659 (2025-05)
Adaptive Prediction-Powered AutoEval with Reliability and Efficiency Guarantees
val_p_value_naive 1 other paper
1 later
2505.18659 (2025-05)
Adaptive Prediction-Powered AutoEval with Reliability and Efficiency Guarantees

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

Prediction-powered inference is a framework for performing valid statistical inference when an experimental dataset is supplemented with predictions from a machine-learning system. The framework yields simple algorithms for computing provably valid confidence intervals for quantities such as means, quantiles, and linear and logistic regression coefficients, without making any assumptions on the machine-learning algorithm that supplies the predictions. Furthermore, more accurate predictions translate to smaller confidence intervals. Prediction-powered inference could enable researchers to draw valid and more data-efficient conclusions using machine learning. The benefits of prediction-powered inference are demonstrated with datasets from proteomics, astronomy, genomics, remote sensing, census analysis, and ecology.

For agents

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

get_harvested_code_for_paper("2301.09633")
get_code_for_paper("2301.09633")
have("2301.09633")

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

curl https://syntology.ai/api/ran/2301.09633.json

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

[![Syntology run record](https://syntology.ai/api/ran/2301.09633.svg)](https://syntology.ai/paper/2301.09633)

Connect an agent — have() is free.