starsofchance/PrimeVul
Dataset Card: PrimeVul Dataset Splits Overview PrimeVul is a dataset crafted for vulnerability detection in C/C++ code, aimed at training and evaluating code language models under realistic conditions. This dataset card describes the pre-split version (train, validation, and test) uploaded to Hugging Face, based on the PrimeVul-v0.1 release. It includes approximately 7,000 vulnerable functions and 229,000 benign functions from real-world projects, covering over… See the full description on the dataset page: https://huggingface.co/datasets/starsofchance/PrimeVul.
Removing processed_dataset folder and its contents
Delete test_split_labels.json with huggingface_hub
Delete test_split_functions.json with huggingface_hub
Delete cleaned_test_split.jsonl with huggingface_hub
Add processed_dataset folder structure
Add cleaned test splits/clened_split/functions/labels (streaming-friendly)
Add dataset card
Upload primevul_train.jsonl with huggingface_hub
Upload primevul_train_paired.jsonl with huggingface_hub
Upload file_info.json with huggingface_hub
Upload primevul_valid_paired.jsonl with huggingface_hub
Upload primevul_test_paired.jsonl with huggingface_hub
Upload primevul_valid.jsonl with huggingface_hub
Upload primevul_test.jsonl with huggingface_hub
initial commit