Datasetpublic

highflame/palisade-scan-results

Palisade Model Security Scans This dataset contains security scan results for popular ML models, generated by Palisade. Dataset Structure data/train.parquet: Leaderboard table with security metrics sarif/: SARIF 2.1.0 scan reports organized by org/model-name Metrics Security Score: Weighted sum of findings (lower is better) Risk Level: safe, low, medium, high, critical Severity Counts: critical, high, medium, low breakdowns Usage… See the full description on the dataset page: https://huggingface.co/datasets/highflame/palisade-scan-results.

sourceHugging Facemitupdated 5mo agoView on Hugging Face

0likes25downloads

Dataset Card

Palisade Model Security Scans

This dataset contains security scan results for popular ML models, generated by Palisade.

Dataset Structure

—data/train.parquet: Leaderboard table with security metrics
—sarif/: SARIF 2.1.0 scan reports organized by org/model-name

Metrics

—Security Score: Weighted sum of findings (lower is better)
—Risk Level: safe, low, medium, high, critical
—Severity Counts: critical, high, medium, low breakdowns

Usage

python

from datasets import load_dataset

ds = load_dataset("palisade-security/model-scans", split="train")
df = ds.to_pandas()

# Get safest models
safe_models = df[df['risk_level'] == 'safe'].sort_values('security_score')

License

Public Domain (CC0)