SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2206.07796 · 2022

FixEval: Execution-based Evaluation of Program Fixes for Programming Problems

arXiv · PDF · Open in the Atlas

Code that ran

11 of 14 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, 1 violated it and 4 have no contract verdict; 6 failures of our instrument; 3 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/2206.07796.json · badge: /api/ran/2206.07796.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
mahimanzum/fixeval canonical (classifier label)
classifier basis not recorded on the link
11 ran · 3 no recorded run
FunctionStatusWhere it lives
add_lang_by_task Ran
ran; our contract draft was wrong, not the code
mahimanzum/fixeval/codet5/_utils.py
code served (permissive licence) · get_code("2e3e72e1e860afb2")
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 str. 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

ARGS = {
    'target_str': 'This is a test string',
    'task': 'summarize',
    'sub_task': 'java-cs'
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
calculate_similarity Ran
ran on a synthesized input
mahimanzum/fixeval/src/split.py
code served (permissive licence) · get_code("2bd99fb1f49c213d")
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 float, 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 torch
import numpy as np

# Construct valid arguments for calculate_similarity
code1_tokens = ['def', 'hello', '(', 'x', ')', ':', 'return', 'x', '+', '1']
code2_tokens = ['def', 'hello', '(', 'y', ')', ':', 'return', 'y', '+', '1']

ARGS = {
    'code1_tokens': code1_tokens,
    'code2_tokens': code2_tokens
}
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 float, 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
check_floating Ran
ran on a synthesized input
mahimanzum/fixeval/src/run_codes_on_tc.py
code served (permissive licence) · get_code("934eab567f180187")
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 bool, 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 torch
import numpy as np

# Analyze the function:
# check_floating(n1, n2) takes two parameters
# Both are used as string arguments to only_digits()
# only_digits() calls .replace() and .isdigit() on them, treating them as strings
# The function converts them to float and compares absolute difference
# So n1 and n2 should be strings representing numbers

# Test with valid numeric strings that are close enough to pass the check
n1 = "3.14159"
n2 = "3.14160"

ARGS = {
    "n1": n1,
    "n2": n2
}
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 bool, 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
convert_clone_examples_to_features Ran
ran; our contract draft was wrong, not the code
mahimanzum/fixeval/codet5/_utils.py
pointer only (licence: MIT); the stored copy get_code serves, from veronicium/cmtrans, is pointer only (licence: NONE) · get_code("cc181d5e6e9504de")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from veronicium/cmtrans, and GitHub detected no licence in that repository 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.

convert_examples_to_features Ran
ran; our contract draft was wrong, not the code
mahimanzum/fixeval/codet5/_utils.py
code served (permissive licence) · get_code("70674cae33ae9c61")
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 InputFeatures. 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 transformers import T5Tokenizer

# Create a minimal example class
class Example:
    def __init__(self, source, target, url=None):
        self.source = source
        self.target = target
        self.url = url

# Create a minimal args class
class Args:
    def __init__(self):
        self.model_type = 't5'
        self.add_task_prefix = True
        self.task = 'summarize'
        self.sub_task = 'none'
        self.max_source_length = 128
        self.max_target_length = 64
        self.add_lang_ids = True

# Create a minimal tokenizer stand-in
class MinimalTokenizer:
    def __init__(self):
        self.eos_token_id = 1
        self.pad_token_id = 0
    
    def encode(self, text, max_length=128, padding='max_length', truncation=True):
        # Return a list of token ids with exactly one eos_token_id at the end
        tokens = [2] * (max_length - 1) + [self.eos_token_id]
        return tokens

# Create example instance
example = Example(
    source="def hello(): pass",
    target="This function says hello",
    url="http://example.com"
)

# Create args instance
args = Args()

# Create tokenizer
tokenizer = MinimalTokenizer()

ARGS = {
    'item': (
        example,
        0,
        tokenizer,
        args,
        'train'
    )
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch, transformers — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type InputFeatures — a InputFeatures output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a InputFeatures 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
create_folder Ran
ran on a synthesized input
mahimanzum/fixeval/src/upload_to_drive.py
code served (permissive licence) · get_code("6eb9a48d1e06e7cf")
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 str. 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.

# Mock the Google Drive API
class MockFile:
    def __init__(self, metadata):
        self.data = metadata.copy()
        self.data['id'] = 'folder_123456789'
    
    def CreateFile(self, metadata):
        return MockFileInstance(metadata)
    
    def __getitem__(self, key):
        return self.data[key]

class MockFileInstance:
    def __init__(self, metadata):
        self.metadata = metadata
        self.data = metadata.copy()
        self.data['id'] = 'folder_id_abc123'
    
    def Upload(self):
        pass
    
    def __getitem__(self, key):
        return self.data[key]

# Create mock drive object
drive = MockFile({})

ARGS = {
    'drive': drive,
    'folder_name': 'test_folder',
    'parent_folder_id': 'parent_id_xyz789'
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
getJsonData Ran
ran; our contract draft was wrong, not the code
mahimanzum/fixeval/src/merge.py
code served (permissive licence) · get_code("304f3eb32f111d55")
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 valid data
temp_dir = tempfile.gettempdir()
json_file_path = os.path.join(temp_dir, "test_data.json")

test_data = {
    "key1": "value1",
    "key2": [1, 2, 3],
    "key3": {"nested": "object"}
}

with open(json_file_path, 'w', encoding='utf8') as f:
    json.dump(test_data, f)

ARGS = {
    "JsonFile": json_file_path
}
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
get_model_size Ran
ran; our contract draft was wrong, not the code
mahimanzum/fixeval/codet5/models.py
code served (permissive licence) · get_code("8cfe38019e9d6a49")
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 str. 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)

# Create a minimal model with some trainable parameters
class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear1 = nn.Linear(4, 3)
        self.linear2 = nn.Linear(3, 2)
    
    def forward(self, x):
        x = self.linear1(x)
        x = self.linear2(x)
        return x

model = SimpleModel()

ARGS = {
    'model': model
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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_codet5 Ran
ran; our contract draft was wrong, not the code
mahimanzum/fixeval/codet5/models.py
pointer only (licence: MIT); the stored copy get_code serves, from veronicium/cmtrans, is pointer only (licence: NONE) · get_code("b57e5f4b7c4d2011")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from veronicium/cmtrans, and GitHub detected no licence in that repository 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.

only_digits Ran
ran, violated the contract we drafted
mahimanzum/fixeval/src/run_codes_on_tc.py
code served (permissive licence) · get_code("2d959fcbc7b2caed")
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 bool, 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 torch

ARGS = {
    'num': '123'
}
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 bool, 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
read_dic Ran
ran on a synthesized input
mahimanzum/fixeval/src/make_submission_list_json.py
code served (permissive licence) · get_code("c5b926cebaadf75d")
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

def read_dic(path):
    return json.loads(path)

# The function takes a 'path' parameter which should be a JSON string
# that can be parsed by json.loads()
path_value = '{"key": "value", "number": 42, "list": [1, 2, 3]}'

ARGS = {
    "path": path_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 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
add_args Not yet run
unverified (harvested, no recorded run)
mahimanzum/fixeval/codet5/configs.py
code served (permissive licence) · get_code("414ce77640f6ccc5")
How it ran

no invocation recorded.

build_or_load_gen_model Not yet run
unverified (harvested, no recorded run)
mahimanzum/fixeval/codet5/models.py
code served (permissive licence) · get_code("75da417c97a30704")
How it ran

no invocation recorded.

load_collected_test_suit Not yet run
unverified (harvested, no recorded run)
mahimanzum/fixeval/src/split.py
code served (permissive licence) · get_code("81617394dadc5689")
How it ran

no invocation recorded.

Also present in other papers' code

Of 14 samples lifted from this paper's linked repositories, 14 carry a function identity and 0 do not. Of the 14 distinct functions with an identity, 7 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 0 the same month, 7 later) and 7 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
get_model_size 6 other papers
1 dated earlier, 5 later
1909.09436 (2019-09)
CodeSearchNet Challenge: Evaluating the State of Semantic Code Search
2207.11680 (2022-07)
No More Fine-Tuning? An Experimental Evaluation of Prompt Tuning in Code Intelligence
2305.17506 (2023-05)
Backdooring Neural Code Search
and 3 more not named here
add_lang_by_task 4 other papers
1 dated earlier, 3 later
1909.09436 (2019-09)
CodeSearchNet Challenge: Evaluating the State of Semantic Code Search
2207.11680 (2022-07)
No More Fine-Tuning? An Experimental Evaluation of Prompt Tuning in Code Intelligence
2311.00317 (2023-11)
Data Augmentation for Code Translation with Comparable Corpora and Multiple References
and 1 more not named here
convert_examples_to_features 3 other papers
3 later
2207.11680 (2022-07)
No More Fine-Tuning? An Experimental Evaluation of Prompt Tuning in Code Intelligence
2311.00317 (2023-11)
Data Augmentation for Code Translation with Comparable Corpora and Multiple References
2312.04234 (2023-12)
Graph Convolutions Enrich the Self-Attention in Transformers!
convert_clone_examples_to_features 1 other paper
1 later
2311.00317 (2023-11)
Data Augmentation for Code Translation with Comparable Corpora and Multiple References
getJsonData 1 other paper
1 later
2310.10628 (2023-10)
Data Contamination Through the Lens of Time
load_codet5 1 other paper
1 later
2311.00317 (2023-11)
Data Augmentation for Code Translation with Comparable Corpora and Multiple References
only_digits 1 other paper
1 later
2310.10628 (2023-10)
Data Contamination Through the Lens of Time

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

The complexity of modern software has led to a drastic increase in the time and cost associated with detecting and rectifying software bugs. In response, researchers have explored various methods to automatically generate fixes for buggy code. However, due to the large combinatorial space of possible fixes for any given bug, few tools and datasets are available to evaluate model-generated fixes effectively. To address this issue, we introduce FixEval, a benchmark comprising of buggy code submissions to competitive programming problems and their corresponding fixes. FixEval offers an extensive collection of unit tests to evaluate the correctness of model-generated program fixes and assess further information regarding time, memory constraints, and acceptance based on a verdict. We consider two Transformer language models pretrained on programming languages as our baseline and compare them using match-based and execution-based evaluation metrics. Our experiments show that match-based metrics do not reflect model-generated program fixes accurately. At the same time, execution-based methods evaluate programs through all cases and scenarios designed explicitly for that solution. Therefore, we believe FixEval provides a step towards real-world automatic bug fixing and model-generated code evaluation. The dataset and models are open-sourced at https://github.com/mahimanzum/FixEval.

For agents

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

get_harvested_code_for_paper("2206.07796")
get_code_for_paper("2206.07796")
have("2206.07796")

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

curl https://syntology.ai/api/ran/2206.07796.json

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

[![Syntology run record](https://syntology.ai/api/ran/2206.07796.svg)](https://syntology.ai/paper/2206.07796)

Connect an agent — have() is free.