# FMI_SU_Yotkova_Kastreva at SemEval-2026 Task 13: Lightweight Detection of LLM-Generated Code via Stylometric Signals

arXiv: https://arxiv.org/abs/2605.04157 · HTML: https://mlsanity.com/paper/2605.04157/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 10 of its 21 references that are in the map.
- First posted on arXiv 5 May 2026.

## Builds on (most cited first, 10 of 10)

- [GraphCodeBERT: Pre-training Code Representations with Data Flow](https://mlsanity.com/paper/2009.08366/) (2020)
- [Asleep at the Keyboard? Assessing the Security of GitHub Copilot's Code Contributions](https://mlsanity.com/paper/2108.09293/) (2021)
- [Efficient Training of Language Models to Fill in the Middle](https://mlsanity.com/paper/2207.14255/) (2022)
- [TestGenEval: A Real World Unit Test Generation and Test Completion Benchmark](https://mlsanity.com/paper/2410.00752/) (2024)
- [CoRNStack: High-Quality Contrastive Data for Better Code Retrieval and Reranking](https://mlsanity.com/paper/2412.01007/) (2024)
- [A Comparative Study of Programming Languages in Rosetta Code](https://mlsanity.com/paper/1409.0252/) (2014)
- [CoDet-M4: Detecting Machine-Generated Code in Multi-Lingual, Multi-Generator and Multi-Domain Settings](https://mlsanity.com/paper/2503.13733/) (2025)
- [textttDroid: A Resource Suite for AI-Generated Code Detection](https://mlsanity.com/paper/2507.10583/) (2025)
- [CodeVision: Detecting LLM-Generated Code Using 2D Token Probability Maps and Vision Models](https://mlsanity.com/paper/2501.03288/) (2025)
- [Automatic Detection of LLM-Generated Code: A Comparative Case Study of Contemporary Models Across Function and Class Granularities](https://mlsanity.com/paper/2409.01382/) (2024)

## Abstract (arXiv)

SemEval-2026 Task 13 investigates machine-generated code detection across multiple programming languages and application scenarios, asking participating systems to generalize to unseen languages and domains. This paper describes our participation in Subtask A (binary classification) and explores both pretrained code encoders and lightweight feature-based methods. We design ratio-based features that are less sensitive to snippet length. To support the extraction of descriptiveness-related signals, we use parsing engines and a programming-language classifier. Additionally, we train a separate code-vs-text line classifier to identify raw natural language segments embedded within samples. We combine a shallow decision tree with heuristic rules derived from data analysis to produce the final predictions. Our approach is computationally efficient, requires only CPU resources for training, and achieves near-instant inference time, offering a lightweight alternative to large pretrained models.