LLMSniffer: Detecting LLM-Generated Code via GraphCodeBERT and Supervised Contrastive Learning

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Dihan, Mahir Labib, Muhtasim, Abir
Format: Preprint
Publié: 2026
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866914483269533696
author Dihan, Mahir Labib
Muhtasim, Abir
author_facet Dihan, Mahir Labib
Muhtasim, Abir
contents The rapid proliferation of Large Language Models (LLMs) in software development has made distinguishing AI-generated code from human-written code a critical challenge with implications for academic integrity, code quality assurance, and software security. We present LLMSniffer, a detection framework that fine-tunes GraphCodeBERT using a two-stage supervised contrastive learning pipeline augmented with comment removal preprocessing and an MLP classifier. Evaluated on two benchmark datasets - GPTSniffer and Whodunit - LLMSniffer achieves substantial improvements over prior baselines: accuracy increases from 70% to 78% on GPTSniffer (F1: 68% to 78%) and from 91% to 94.65% on Whodunit (F1: 91% to 94.64%). t-SNE visualizations confirm that contrastive fine-tuning yields well-separated, compact embeddings. We release our model checkpoints, datasets, codes and a live interactive demo to facilitate further research.
format Preprint
id arxiv_https___arxiv_org_abs_2604_16058
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle LLMSniffer: Detecting LLM-Generated Code via GraphCodeBERT and Supervised Contrastive Learning
Dihan, Mahir Labib
Muhtasim, Abir
Software Engineering
Computation and Language
The rapid proliferation of Large Language Models (LLMs) in software development has made distinguishing AI-generated code from human-written code a critical challenge with implications for academic integrity, code quality assurance, and software security. We present LLMSniffer, a detection framework that fine-tunes GraphCodeBERT using a two-stage supervised contrastive learning pipeline augmented with comment removal preprocessing and an MLP classifier. Evaluated on two benchmark datasets - GPTSniffer and Whodunit - LLMSniffer achieves substantial improvements over prior baselines: accuracy increases from 70% to 78% on GPTSniffer (F1: 68% to 78%) and from 91% to 94.65% on Whodunit (F1: 91% to 94.64%). t-SNE visualizations confirm that contrastive fine-tuning yields well-separated, compact embeddings. We release our model checkpoints, datasets, codes and a live interactive demo to facilitate further research.
title LLMSniffer: Detecting LLM-Generated Code via GraphCodeBERT and Supervised Contrastive Learning
topic Software Engineering
Computation and Language
url https://arxiv.org/abs/2604.16058