R2Vul - Data and Models
R2Vul - Data and Models
This Zenodo repository contains the datasets and models for our paper "R2Vul: Learning to Reason about Software Vulnerabilities with Reinforcement Learning and Structured Reasoning Distillation". Datasets raw_dataset.json - the raw dataset mined from NVD. r2vul_dataset.zip - the dataset used for fine-tuning and testing. external_java_test.zip - the external, manually annotated Java dataset (RQ2). Models and Checkpoints cls.zip - includes all models checkpoints fine-tuned using CLS. sft.zip - includes all models checkpoints fine-tuned using SFT. orpo.zip - includes all models checkpoints fine-tuned using ORPO (R2Vul). MSIVD.zip - downloaded from https://zenodo.org/records/11403208 (codellama-13b - bigvul_expl). VulLLM.zip - downloaded from https://zenodo.org/records/10677069 (codellama-13b-multi-r16-2048).
selected citations These citations are derived from selected sources.This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).0 popularity This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.Average influence This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).Average impulse This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.Average
Citations provided by BIP!Popularity provided by BIP!