EVIL (Exploiting software VIa natural Language) is an approach to automatically generate software exploits in assembly/Python language from descriptions in natural language. The approach leverages Neural Machine Translation (NMT) techniques and a dataset that we developed for this work.
Neural search engine for discovering semantically similar Python repositories on GitHub
Augmenting the Interpretability of GraphCodeBERT for Code Similarity Tasks
Repository about small code models
This repository contains the code, the dataset and the experimental results related to the paper "Vulnerabilities in AI Code Generators: Exploring Targeted Data Poisoning Attacks" accepted for publication at The 32nd IEEE/ACM International Conference on Program Comprehension (ICPC 2024).
Improving Source Code Similarity Detection with GraphCodeBERT and Additional Feature Integration
Fine-tuning CodeBERT for Vulnerability Detection
Implementation and dataset for A Zero-Shot Framework for Cross-Project Vulnerability Detection in Source Code (Empirical Software Engineering, 2026).
CodeOpt: A framework for optimizing code performance using Two-Stage Sampling, Few-Shot Learning, and Iterative Self-Reflection with support for Genetic Algorithm Inspired Chain-of-Thought (GA-COT).
The modern web development landscape is plagued by a peculiar paradox: despite the abundance of UI components and design systems, developers still spend countless hours reimplementing similar interfaces. S0 addresses this challenge by introducing a novel approach that combines advanced vector search capabilities.
BERT-based MLM pretrained on LilyPond datasets
AI-based code vulnerability analysis system using Transformer & AST
Django implementation of CodeBERT for detecting vulnerable code.
A deterministic and neuro-symbolic framework for evaluating LLM-generated code using Abstract Syntax Trees, Semantic Embeddings, and Integrated Gradients. Think of it as a 'Digital Polygraph' for AI. It uses a three-step verification process to ensure the AI didn't 'misunderstand' your instructions
An advanced AI-driven SAST engine that utilizes Tree-sitter for AST parsing, Microsoft CodeBERT for semantic featurization, and PyTorch Geometric GCNs for static code vulnerability detection, paired with a local DeepSeek LLM triage assistant loop.
MSc Software Engineering Dissertation — AI-Assisted Semantic Feature Location using CodeBERT and CodeSearchNet
Benchmarking parameter-efficient fine-tuning for code models: 179x smaller checkpoints for 1-5 points of accuracy. 24 runs, 3 seeds, reproducible on a free GPU.
Official implementation of OmniPhish: a hybrid CodeBERT + 1D-CNN + XGBoost stacking ensemble for robust zero-day phishing detection.
To associate your repository with the codebert topic, visit your repo's landing page and select "manage topics."