datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
shenava-phaseb-processedRakshakAI-phase-b
RakshakAI Phase B Dataset (v2 — cleaned)
Bug-fixed version of the RakshakAI Phase B security dataset:
Bug 1: NVD/OSV advisory text masquerading as code — source-based filter removed ~22K rows
Bug 2: OWASP false positives labeled as vulnerable — 2,740 rows corrected to is_vulnerable=false
Bug 3: BigVul identical before/after patches — 57,203 vuln rows nulled (whitespace normalization revealed 98% identity)
Bug 4: Partial snippet patches (<50% of vuln lines) — 1,187 rows nulled… See the full description on the dataset page: https://huggingface.co/datasets/Muneerali199/RakshakAI-phase-b.