@inproceedings{sahlen-etal-2026-miun,
title = "{MIUN} {B}ias{P}atrol at {S}em{E}val-2026 Task 13: Why {TF}-{IDF} can Beat Transformers for {OOD} Code Detection",
author = "Sahlen, Loviza and
Springfeldt, Thomas and
Fatima, Mehwish and
Shahzad, Raja Khurram",
editor = "Kochmar, Ekaterina and
Ghosh, Debanjan and
North, Kai and
Komachi, Mamoru",
booktitle = "Proceedings of the 20th {I}nternational {W}orkshop on {S}emantic {E}valuation (2026)",
month = jul,
year = "2026",
address = "San Diego, California, USA",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.semeval-1.312/",
pages = "2469--2474",
ISBN = "979-8-89176-414-9",
abstract = "The increasing use of AI-generated code underscores the need for effective detection systems. However, their performance often deteriorates when faced with distribution shifts. This paper presents our system for SemEval-2026 Task 13: A, which focuses on binary classification of human-written versus machine-generated code across various programming languages and domains. We systematically compare traditional classifiers, such as Random Forest and XGBoost, which utilize statistical and TF-IDF features, against pipelines that leverage frozen embeddings from advanced code transformers like UniXcoder and GraphCodeBERT. Our results reveal a notable trade-off, i.e., while transformer-based pipelines excel in in-distribution validation (reaching up to 0.89 Macro F1), they experience severe performance drops up to 77{\%}; when applied to out-of-distribution languages and domains. In contrast, models employing TF-IDF with tree-based classifiers demonstrate significantly greater stability. We identify this fragility as a result of a bias toward superficial formatting, particularly whitespace, which is accentuated by transformers. By implementing simple space normalization, we enhance the out-of-distribution robustness of traditional models; however, this also highlights the ongoing dependence of embeddings on these non-semantic features. Our findings suggest that for creating generalizable code detection systems, straightforward, well-normalized lexical features may be more reliable than complex, unrefined embeddings."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="sahlen-etal-2026-miun">
<titleInfo>
<title>MIUN BiasPatrol at SemEval-2026 Task 13: Why TF-IDF can Beat Transformers for OOD Code Detection</title>
</titleInfo>
<name type="personal">
<namePart type="given">Loviza</namePart>
<namePart type="family">Sahlen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Thomas</namePart>
<namePart type="family">Springfeldt</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mehwish</namePart>
<namePart type="family">Fatima</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Raja</namePart>
<namePart type="given">Khurram</namePart>
<namePart type="family">Shahzad</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 20th International Workshop on Semantic Evaluation (2026)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Ekaterina</namePart>
<namePart type="family">Kochmar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Debanjan</namePart>
<namePart type="family">Ghosh</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kai</namePart>
<namePart type="family">North</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mamoru</namePart>
<namePart type="family">Komachi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">San Diego, California, USA</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-414-9</identifier>
</relatedItem>
<abstract>The increasing use of AI-generated code underscores the need for effective detection systems. However, their performance often deteriorates when faced with distribution shifts. This paper presents our system for SemEval-2026 Task 13: A, which focuses on binary classification of human-written versus machine-generated code across various programming languages and domains. We systematically compare traditional classifiers, such as Random Forest and XGBoost, which utilize statistical and TF-IDF features, against pipelines that leverage frozen embeddings from advanced code transformers like UniXcoder and GraphCodeBERT. Our results reveal a notable trade-off, i.e., while transformer-based pipelines excel in in-distribution validation (reaching up to 0.89 Macro F1), they experience severe performance drops up to 77%; when applied to out-of-distribution languages and domains. In contrast, models employing TF-IDF with tree-based classifiers demonstrate significantly greater stability. We identify this fragility as a result of a bias toward superficial formatting, particularly whitespace, which is accentuated by transformers. By implementing simple space normalization, we enhance the out-of-distribution robustness of traditional models; however, this also highlights the ongoing dependence of embeddings on these non-semantic features. Our findings suggest that for creating generalizable code detection systems, straightforward, well-normalized lexical features may be more reliable than complex, unrefined embeddings.</abstract>
<identifier type="citekey">sahlen-etal-2026-miun</identifier>
<location>
<url>https://aclanthology.org/2026.semeval-1.312/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>2469</start>
<end>2474</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T MIUN BiasPatrol at SemEval-2026 Task 13: Why TF-IDF can Beat Transformers for OOD Code Detection
%A Sahlen, Loviza
%A Springfeldt, Thomas
%A Fatima, Mehwish
%A Shahzad, Raja Khurram
%Y Kochmar, Ekaterina
%Y Ghosh, Debanjan
%Y North, Kai
%Y Komachi, Mamoru
%S Proceedings of the 20th International Workshop on Semantic Evaluation (2026)
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, California, USA
%@ 979-8-89176-414-9
%F sahlen-etal-2026-miun
%X The increasing use of AI-generated code underscores the need for effective detection systems. However, their performance often deteriorates when faced with distribution shifts. This paper presents our system for SemEval-2026 Task 13: A, which focuses on binary classification of human-written versus machine-generated code across various programming languages and domains. We systematically compare traditional classifiers, such as Random Forest and XGBoost, which utilize statistical and TF-IDF features, against pipelines that leverage frozen embeddings from advanced code transformers like UniXcoder and GraphCodeBERT. Our results reveal a notable trade-off, i.e., while transformer-based pipelines excel in in-distribution validation (reaching up to 0.89 Macro F1), they experience severe performance drops up to 77%; when applied to out-of-distribution languages and domains. In contrast, models employing TF-IDF with tree-based classifiers demonstrate significantly greater stability. We identify this fragility as a result of a bias toward superficial formatting, particularly whitespace, which is accentuated by transformers. By implementing simple space normalization, we enhance the out-of-distribution robustness of traditional models; however, this also highlights the ongoing dependence of embeddings on these non-semantic features. Our findings suggest that for creating generalizable code detection systems, straightforward, well-normalized lexical features may be more reliable than complex, unrefined embeddings.
%U https://aclanthology.org/2026.semeval-1.312/
%P 2469-2474
Markdown (Informal)
[MIUN BiasPatrol at SemEval-2026 Task 13: Why TF-IDF can Beat Transformers for OOD Code Detection](https://aclanthology.org/2026.semeval-1.312/) (Sahlen et al., SemEval 2026)
ACL