SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2603.22018 · 2026

Do Papers Tell the Whole Story? A Benchmark and Framework for Uncovering Hidden Implementation Gaps in Bioinformatics

Xin Lai, Xiaoyan Zhu, Jiayin Wang, Tianxiang Xu, Xin Lian, Sizhe Dang, Hangyu Cheng, Codebases Codebases, Codebases, Paper Pdfs

arXiv · PDF · Open in the Atlas

Code that ran

2 of 2 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (1 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 2 have no contract verdict; 0 failures of our instrument; 0 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/2603.22018.json · badge: /api/ran/2603.22018.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
Ricardo1998-Xu/BioCon community
Papers with Code archive
2 ran · 0 no recorded run
FunctionStatusWhere it lives
FocalLoss_Weight Ran
ran on a synthesized input
Ricardo1998-Xu/BioCon/model.py
code served (permissive licence) · get_code("9b5a0deecd913898")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned Tensor, 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 torch
import torch.nn as nn

# INIT_ARGS: parameters for FocalLoss_Weight.__init__
INIT_ARGS = {
    'alpha': torch.tensor([0.25, 0.75, 1.0]),  # weights for 3 classes
    'gamma': 2,
    'reduction': 'mean'
}

# ARGS: parameters for FocalLoss_Weight.forward
# input: (batch_size=2, num_classes=3)
# target: (batch_size=2) with class indices
ARGS = {
    'input': torch.randn(2, 3),  # logits for 2 samples, 3 classes
    'target': torch.tensor([0, 2])  # class labels
}
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 0, dtype family float — no output shape is stated: observe_signature.py observed this class with its constructor sizes canonicalised where it could, which the node does not record, so its shape is not stated as this setup block's
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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
Model Ran
ran on a synthesized input · constructed an object; no result was computed
Ricardo1998-Xu/BioCon/model.py
code served (permissive licence) · get_code("bff3fb7e46fce7df")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned an instance of its own class, Model: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

# Create a minimal encoder that satisfies the usage in Model.forward()
class MinimalEncoder(nn.Module):
    def __init__(self, hidden_size=4):
        super().__init__()
        self.config = type('Config', (), {'hidden_size': hidden_size})()
        self.hidden_size = hidden_size
    
    def forward(self, input_ids, attention_mask=None, output_hidden_states=True):
        batch_size = input_ids.shape[0]
        seq_length = input_ids.shape[1]
        
        # Return outputs with hidden_states attribute
        hidden_state = torch.randn(batch_size, seq_length, self.hidden_size)
        
        outputs = type('Outputs', (), {
            'hidden_states': [hidden_state, hidden_state]
        })()
        return outputs

# Create a minimal tokenizer that satisfies the usage in Model.forward()
class MinimalTokenizer:
    def __init__(self):
        self.pad_token_id = 0

# Create a minimal config
class MinimalConfig:
    def __init__(self):
        self.hidden_size = 4

# Create minimal args
class MinimalArgs:
    def __init__(self):
        self.num_labels = 2
        self.dropout_probability = 0.1

encoder = MinimalEncoder(hidden_size=4)
config = MinimalConfig()
tokenizer = MinimalTokenizer()
args = MinimalArgs()

ARGS = {
    'encoder': encoder,
    'config': config,
    'tokenizer': tokenizer,
    'args': args
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Model — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a Model 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

Also present in other papers' code

Of 2 samples lifted from this paper's linked repositories, 2 carry a function identity and 0 do not. Of the 2 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 2 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

Ensuring consistency between research papers and their corresponding software code implementations is a fundamental prerequisite for guaranteeing the reproducibility of scientific findings and the reliability of software systems. However, this issue has received limited attention to date, particularly in the field of bioinformatics, where inconsistencies between methodological descriptions in papers and their actual code implementations are prevalent. To address this gap, we introduce a novel research task, namely paper-code consistency detection, which aims to characterize the cross-modal semantic alignment between methodological descriptions in papers and their corresponding code implementations. At the data level, we construct the first benchmark dataset for this task in the bioinformatics domain, termed BioCon, comprising 48 bioinformatics software projects and their associated publications. BioCon is built by fine-grained alignment between sentence-level methodological descriptions in papers and function-level code snippets, combined with expert annotation and hard negative sampling strategies, resulting in a high-quality sentence-code paired dataset. At the methodological level, we propose a unified cross-modal consistency detection framework that leverages pre-trained models to jointly encode paper sentences and code functions. To address class imbalance, a weighted focal loss is introduced to enhance discriminative capability. At the evaluation level, we conduct a systematic analysis from three perspectives: sentence-level classification, cross-modal retrieval, and project-level consistency assessment. Experimental results demonstrate that the proposed approach achieves strong performance in both consistency discrimination and semantic alignment. Furthermore, a case study on 23 real-world bioinformatics projects from Bioinformatics journal highlights the practical effectiveness of our framework in identifying potential inconsistencies. Overall, this work establishes the first systematic benchmark and framework for paper-code consistency analysis, opening a new research direction and providing a foundation for improving reproducibility and reliability in bioinformatics software. The data and code are available at https://github.com/Ricardo1998-Xu/BioCon.

For agents

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

get_harvested_code_for_paper("2603.22018")
get_code_for_paper("2603.22018")
have("2603.22018")

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

curl https://syntology.ai/api/ran/2603.22018.json

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

[![Syntology run record](https://syntology.ai/api/ran/2603.22018.svg)](https://syntology.ai/paper/2603.22018)

Connect an agent — have() is free.