SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2106.01703 · ACL Findings · 2021

Fingerprinting Fine-tuned Language Models in the Wild *

Nirav Diwan, Tanmoy Chakravorty, Zubair Shafiq

arXiv · PDF · Open in the Atlas

Code that ran

13 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; 0 honoured the contract we drafted, 0 violated it and 13 have no contract verdict; 0 failures of our instrument; 3 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/2106.01703.json · badge: /api/ran/2106.01703.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
LCS2-IIITD/ACL-FFLM canonical (classifier label)
classifier basis: census
13 ran · 3 no recorded run
FunctionStatusWhere it lives
convert Ran
ran on a synthesized input
LCS2-IIITD/ACL-FFLM/models/generate-embed/pretrained/pretrained-generate-embed.py
code served (permissive licence) · get_code("a34a68ec1f87d1c1")
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. 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 itertools import islice

torch.manual_seed(0)

# The function takes two parameters:
# - lst: an iterable that will be converted to an iterator
# - var_lst: a list of integers specifying slice sizes

# Create a simple list to iterate over
lst = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]

# Create a list of slice sizes
var_lst = [2, 3, 5]

ARGS = {
    'lst': lst,
    'var_lst': var_lst
}
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 list — a list output has no array shape of its own; none is 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
convert_timestamp Ran
ran on a synthesized input
LCS2-IIITD/ACL-FFLM/preprocess/create_full_dataset.py
code served (permissive licence) · get_code("2e298fac74427689")
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 datetime

def convert_timestamp(timestamp):
    """
    Convert  timestamp from reddit to day, month and year
    """
    date = datetime.datetime.fromtimestamp(timestamp).strftime('%Y-%m-%d %H:%M:%S')
    date_split = date.split('-')
    month = int(date_split[1])
    year = int(date_split[0])
    return (month, year, date)

# Create a valid Unix timestamp (seconds since epoch)
# Using a timestamp from January 15, 2023
timestamp_value = 1673740800

ARGS = {
    'timestamp': timestamp_value
}
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 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
generate_embeddings Ran
ran on a synthesized input
LCS2-IIITD/ACL-FFLM/models/generate-embed/pretrained/pretrained-generate-embed.py
code served (permissive licence) · get_code("a3aad5d3c0c10547")
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 (1, 768) 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

class MockTokenizer:
    def __init__(self):
        self.cls_token_id = 101
    
    def encode(self, sentence, add_special_tokens=True):
        # Return a list of token ids
        return [101, 2054, 2003, 1045, 102]

class MockRobertaModel(tf.Module):
    def __init__(self):
        super().__init__()
    
    def __call__(self, input_ids):
        # Simulate model output
        batch_size = tf.shape(input_ids)[0]
        seq_length = tf.shape(input_ids)[1]
        # Return tuple with embeddings of shape [batch_size, seq_length, hidden_dim]
        embeddings = tf.random.normal([batch_size, seq_length, 768])
        return (embeddings,)

class MockDoc2VecModel:
    def infer_vector(self, sentence):
        # Return a 1D vector
        return np.random.randn(768)

class MockXLNetModel(tf.Module):
    def __init__(self):
        super().__init__()
    
    def __call__(self, input_ids):
        batch_size = tf.shape(input_ids)[0]
        seq_length = tf.shape(input_ids)[1]
        embeddings = tf.random.normal([batch_size, seq_length, 768])
        return (embeddings,)

class MockGPT2Tokenizer:
    def __init__(self):
        self.cls_token_id = 50256
    
    def encode(self, sentence, add_special_tokens=True):
        return [50256, 2054, 2003, 1045, 50256]

class MockGPT2Model(tf.Module):
    def __init__(self):
        super().__init__()
    
    def __call__(self, input_ids):
        batch_size = tf.shape(input_ids)[0]
        seq_length = tf.shape(input_ids)[1]
        embeddings = tf.random.normal([batch_size, seq_length, 768])
        return (embeddings,)

# Create instances for testing with roberta-base
model_name = 'roberta-base'
model = MockRobertaModel()
sentence = "This is a test sentence"
tokenizer = MockTokenizer()

ARGS = {
    'model_name': model_name,
    'model': model,
    'sentence': sentence,
    'tokenizer': tokenizer,
    'batch_size': 32
}
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 ndarray, ndim 2, dtype family float, shape (1, 768) — shape (1, 768) 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_dist Ran
ran on a synthesized input
LCS2-IIITD/ACL-FFLM/models/generate-embed/ft/seq_classification.py
code served (permissive licence) · get_code("0d7d5e73ca198d40")
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 torch
import numpy as np

torch.manual_seed(0)

# Analyze the function:
# - labels: iterable of integers (class indices)
# - nfclass: integer representing number of classes
# The function iterates over labels and indexes into per_class_count with each label

# Create small test data
labels = [0, 1, 0, 2, 1, 2, 2]  # 7 samples across 3 classes
nfclass = 3

ARGS = {
    "labels": labels,
    "nfclass": nfclass
}
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 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
initialize_classifiers Ran
ran on a synthesized input
LCS2-IIITD/ACL-FFLM/models/sklearn_classifiers.py
code served (permissive licence) · get_code("217dfb86b70c79c5")
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 torch
import numpy as np
from sklearn.neural_network import MLPClassifier
from sklearn.ensemble import RandomForestClassifier, AdaBoostClassifier
from sklearn.naive_bayes import GaussianNB
from sklearn.neighbors import KNeighborsClassifier
from sklearn.svm import SVC
from sklearn.tree import DecisionTreeClassifier
from sklearn.discriminant_analysis import QuadraticDiscriminantAnalysis
from xgboost import XGBClassifier
from sklearn.discriminant_analysis import LinearDiscriminantAnalysis

torch.manual_seed(0)

ARGS = {
    'num_classes': 3,
    'classifier_type': 'fast'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn, torch, xgboost — 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
loadJsonFile Ran
ran on a synthesized input
LCS2-IIITD/ACL-FFLM/models/read_write_functions.py
code served (permissive licence) · get_code("65d81b283d6d14fe")
How it ran

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

import json
import tempfile
import os

# Create a temporary JSON file with test data
test_data = {
    "key1": "value1",
    "key2": 42,
    "key3": [1, 2, 3],
    "nested": {
        "inner_key": "inner_value"
    }
}

# Write the test data to a temporary file
temp_dir = tempfile.gettempdir()
test_filepath = os.path.join(temp_dir, "test_data.json")
with open(test_filepath, 'w') as f:
    json.dump(test_data, f)

ARGS = {
    "filepath": test_filepath
}
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 dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
loadPickleFile Ran
ran on a synthesized input
LCS2-IIITD/ACL-FFLM/models/read_write_functions.py
code served (permissive licence) · get_code("d069cf3bd43db648")
How it ran

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

import pickle
import tempfile
import os

# Create a temporary pickle file with valid content
temp_dir = tempfile.mkdtemp()
temp_filepath = os.path.join(temp_dir, "test_file.pkl")

# Create a sample dictionary and pickle it
sample_dict = {
    "key1": "value1",
    "key2": [1, 2, 3],
    "key3": {"nested": "dict"}
}

with open(temp_filepath, 'wb') as f:
    pickle.dump(sample_dict, f)

ARGS = {
    "filepath": temp_filepath
}
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 dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
loadPickleFile Ran
ran on a synthesized input
LCS2-IIITD/ACL-FFLM/create-splits/create_splits.py
code served (permissive licence) · get_code("f940e406bb65ae47")
How it ran

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

import pickle
import tempfile
import os

# Create a temporary pickle file with sample data
temp_dir = tempfile.gettempdir()
temp_filepath = os.path.join(temp_dir, "test_pickle_file.pkl")

sample_dict = {
    "data": [1, 2, 3, 4, 5],
    "metadata": {"name": "test", "version": 1},
    "nested": {"key": "value"}
}

with open(temp_filepath, 'wb') as f:
    pickle.dump(sample_dict, f)

# Define arguments for the function
ARGS = {
    "filepath": temp_filepath,
    "verbose": True,
    "print_obj": False
}
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 dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
load_dataset Ran
ran on a synthesized input
LCS2-IIITD/ACL-FFLM/models/generate-embed/ft/seq_classification.py
code served (permissive licence) · get_code("ddffc0e8663801cd")
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. 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 pandas as pd
import json

# Mock the loadJsonFile function since we can't do file I/O
def loadJsonFile(path):
    """Mock function that returns test data"""
    pass

# Set the global min_nf_tokens
min_nf_tokens = 6

# Create test data structure
# The complete_dataset structure based on how it's used:
# complete_dataset[label1]['data'][commentID] should have:
# - 'classno': the label
# - 'comment': the text
# - 'nftokens': number of tokens (must be >= min_nf_tokens)

complete_dataset = {
    '1': {
        'data': {
            'c1-comment1': {
                'classno': 1,
                'comment': 'This is a test comment with enough tokens.',
                'nftokens': 8
            },
            'c1-comment2': {
                'classno': 1,
                'comment': 'Another test comment for validation.',
                'nftokens': 7
            }
        }
    },
    '2': {
        'data': {
            'c2-comment1': {
                'classno': 2,
                'comment': 'A different class comment here.',
                'nftokens': 6
            }
        }
    },
    '3': {
        'data': {
            'c3-comment1': {
                'classno': 3,
                'comment': 'Third class test comment with tokens.',
                'nftokens': 7
            },
            'c3-comment2': {
                'classno': 3,
                'comment': 'More test data for testing purposes.',
                'nftokens': 6
            }
        }
    }
}

# Create ids structure with train/val/test splits
ids = {
    'train': ['c1-comment1', 'c1-comment2'],
    'val': ['c2-comment1'],
    'test': ['c3-comment1', 'c3-comment2']
}

# Mock the loadJsonFile to return our test data
original_loadJsonFile = loadJsonFile
def loadJsonFile_mock(path):
    if 'dataset' in path:
        return complete_dataset
    elif 'ids' in path:
        return ids
    return {}

# Patch it in the module where make_dataset and load_dataset are defined
import sys
current_module = sys.modules[__name__]
current_module.loadJsonFile = loadJsonFile_mock

ARGS = {
    'dataset_path': 'test_dataset.json',
    'ids_path': 'test_ids.json'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
pandas — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is 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_dataset Ran
ran on a synthesized input
LCS2-IIITD/ACL-FFLM/models/generate-embed/ft/seq_classification.py
code served (permissive licence) · get_code("728e40674f6bf35c")
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 pandas as pd

# Construct the complete_dataset structure
complete_dataset = {
    '1': {
        'data': {
            'c1-comment1': {
                'classno': 1,
                'comment': 'This is a sample comment about the topic.',
                'nftokens': 10
            },
            'c1-comment2': {
                'classno': 1,
                'comment': 'Another comment with sufficient tokens.',
                'nftokens': 8
            }
        }
    },
    '2': {
        'data': {
            'c2-comment1': {
                'classno': 2,
                'comment': 'A different class comment here.',
                'nftokens': 7
            }
        }
    }
}

# Construct the ids list
ids = ['c1-comment1', 'c1-comment2', 'c2-comment1']

ARGS = {
    'ids': ids,
    'complete_dataset': complete_dataset
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
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
make_sets Ran
ran on a synthesized input
LCS2-IIITD/ACL-FFLM/create-splits/create_splits.py
code served (permissive licence) · get_code("14a777d170409d8d")
How it ran

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

import torch
import numpy as np

torch.manual_seed(0)

# The function takes:
# - indexes: a list of indices
# - comment_ids: a list-like object that can be indexed

indexes = [0, 2, 1]
comment_ids = ["comment_001", "comment_002", "comment_003", "comment_004"]

ARGS = {
    "indexes": indexes,
    "comment_ids": comment_ids,
}
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 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
openCSVfile Ran
ran on a synthesized input
LCS2-IIITD/ACL-FFLM/models/read_write_functions.py
code served (permissive licence) · get_code("656cc72eb9b46d98")
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 2, 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 csv
import tempfile
import os

# Create a temporary CSV file for testing
temp_dir = tempfile.mkdtemp()
temp_filepath = os.path.join(temp_dir, "test.csv")

# Write some sample CSV data
with open(temp_filepath, 'w', newline='') as f:
    writer = csv.writer(f, delimiter=',')
    writer.writerow(['col1', 'col2', 'col3'])
    writer.writerow(['val1', 'val2', 'val3'])
    writer.writerow(['val4', 'val5', 'val6'])

ARGS = {
    'filepath': temp_filepath,
    'delimiter': ','
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 2, 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
openCSVfile Ran
ran on a synthesized input
LCS2-IIITD/ACL-FFLM/preprocess/create_full_dataset.py
code served (permissive licence) · get_code("51db9a13268db992")
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 2, 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 tempfile
import os

# Create a temporary CSV file with some sample data
temp_dir = tempfile.mkdtemp()
csv_filepath = os.path.join(temp_dir, "test.csv")

with open(csv_filepath, 'w') as f:
    f.write("name,age,city\n")
    f.write("Alice,30,New York\n")
    f.write("Bob,25,Los Angeles\n")
    f.write("Charlie,35,Chicago\n")

ARGS = {
    'filepath': csv_filepath,
    'delimiter': ','
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 2, 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
make_paths Not yet run
unverified (harvested, no recorded run)
LCS2-IIITD/ACL-FFLM/models/classifiers/pretrained_classify_main_final.py
code served (permissive licence) · get_code("0cc7fd039d01bf50")
How it ran

no invocation recorded.

new_classifier Not yet run
unverified (harvested, no recorded run)
LCS2-IIITD/ACL-FFLM/models/sklearn_classifiers.py
code served (permissive licence) · get_code("246180ab9a03e813")
How it ran

no invocation recorded.

tokenize_words Not yet run
unverified (harvested, no recorded run)
LCS2-IIITD/ACL-FFLM/preprocess/create_full_dataset.py
code served (permissive licence) · get_code("61dfa1c9e178c8b1")
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, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 15 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.

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

There are concerns that the ability of language models (LMs) to generate high quality synthetic text can be misused to launch spam, disinformation, or propaganda. Therefore, the research community is actively working on developing approaches to detect whether a given text is organic or synthetic. While this is a useful first step, it is important to be able to further fingerprint the author LM to attribute its origin. Prior work on fingerprinting LMs is limited to attributing synthetic text generated by a handful (usually < 10) of pre-trained LMs. However, LMs such as GPT2 are commonly fine-tuned in a myriad of ways (e.g., on a domain-specific text corpus) before being used to generate synthetic text. It is challenging to fingerprinting fine-tuned LMs because the universe of fine-tuned LMs is much larger in realistic scenarios. To address this challenge, we study the problem of large-scale fingerprinting of fine-tuned LMs in the wild. Using a real-world dataset of synthetic text generated by 108 different fine-tuned LMs, we conduct comprehensive experiments to demonstrate the limitations of existing fingerprinting approaches. Our results show that fine-tuning itself is the most effective in attributing the synthetic text generated by fine-tuned LMs.

For agents

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

get_harvested_code_for_paper("2106.01703")
get_code_for_paper("2106.01703")
have("2106.01703")

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

curl https://syntology.ai/api/ran/2106.01703.json

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

[![Syntology run record](https://syntology.ai/api/ran/2106.01703.svg)](https://syntology.ai/paper/2106.01703)

Connect an agent — have() is free.