repo_id stringlengths 15 56 | security_score int64 0 184 | risk_level stringclasses 3
values | total_warnings int64 0 8 | critical_count int64 0 0 | high_count int64 0 3 | medium_count int64 0 3 | low_count int64 0 3 | scan_date stringdate 2025-12-15 02:45:52 2026-04-12 04:17:47 | palisade_version stringclasses 1
value | model_url stringlengths 38 79 | threat_categories stringclasses 3
values | policy_compliance stringclasses 1
value | top_validators stringclasses 3
values | mitre_techniques stringclasses 2
values |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
google/gemma-3-270m-it | 0 | safe | 0 | 0 | 0 | 0 | 0 | 2025-12-15T02:45:52.194529 | 0.4.0 | https://huggingface.co/google/gemma-3-270m-it | {} | {} | [] | [] |
matrixportalx/Llama-2-7b-chat-hf-Q4_K_M-GGUF | 184 | high | 8 | 0 | 3 | 3 | 2 | 2025-12-15T03:58:52.761612 | 0.4.0 | https://huggingface.co/matrixportalx/Llama-2-7b-chat-hf-Q4_K_M-GGUF | {"buffer_overflow": 3, "model_integrity": 1, "behavior_analysis": 1, "gguf_safety": 2, "tool_call_security": 1} | {} | [{"name": "BufferOverflowValidator", "count": 3}, {"name": "GGUFSafetyValidator", "count": 2}, {"name": "ModelIntegrityValidator", "count": 1}, {"name": "BehaviorAnalysisValidator", "count": 1}, {"name": "ToolCallSecurityValidator", "count": 1}] | ["AML.T0051.002"] |
Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled | 0 | safe | 0 | 0 | 0 | 0 | 0 | 2026-04-07T05:21:59.907447 | 0.4.0 | https://huggingface.co/Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled | {} | {} | [] | [] |
sentence-transformers/all-MiniLM-L6-v2 | 0 | safe | 0 | 0 | 0 | 0 | 0 | 2026-04-12T03:15:45.488073 | 0.4.0 | https://huggingface.co/sentence-transformers/all-MiniLM-L6-v2 | {} | {} | [] | [] |
sentence-transformers/all-mpnet-base-v2 | 0 | safe | 0 | 0 | 0 | 0 | 0 | 2026-04-12T03:15:59.224960 | 0.4.0 | https://huggingface.co/sentence-transformers/all-mpnet-base-v2 | {} | {} | [] | [] |
BAAI/bge-small-en-v1.5 | 0 | safe | 0 | 0 | 0 | 0 | 0 | 2026-04-12T03:16:04.943139 | 0.4.0 | https://huggingface.co/BAAI/bge-small-en-v1.5 | {} | {} | [] | [] |
BAAI/bge-base-en-v1.5 | 0 | safe | 0 | 0 | 0 | 0 | 0 | 2026-04-12T03:16:25.218686 | 0.4.0 | https://huggingface.co/BAAI/bge-base-en-v1.5 | {} | {} | [] | [] |
microsoft/phi-2 | 6 | safe | 3 | 0 | 0 | 0 | 3 | 2026-04-12T03:20:28.478308 | 0.4.0 | https://huggingface.co/microsoft/phi-2 | {"backdoor": 3} | {} | [{"name": "BackdoorDetectionValidator", "count": 3}] | [] |
google/gemma-2b | 0 | safe | 0 | 0 | 0 | 0 | 0 | 2026-04-12T03:23:35.355605 | 0.4.0 | https://huggingface.co/google/gemma-2b | {} | {} | [] | [] |
TinyLlama/TinyLlama-1.1B-Chat-v1.0 | 0 | safe | 0 | 0 | 0 | 0 | 0 | 2026-04-12T03:24:48.960818 | 0.4.0 | https://huggingface.co/TinyLlama/TinyLlama-1.1B-Chat-v1.0 | {} | {} | [] | [] |
google/gemma-2b-it | 0 | safe | 0 | 0 | 0 | 0 | 0 | 2026-04-12T03:27:55.580948 | 0.4.0 | https://huggingface.co/google/gemma-2b-it | {} | {} | [] | [] |
Qwen/Qwen3-1.7B | 0 | safe | 0 | 0 | 0 | 0 | 0 | 2026-04-12T03:30:12.109696 | 0.4.0 | https://huggingface.co/Qwen/Qwen3-1.7B | {} | {} | [] | [] |
meta-llama/Llama-2-7b-hf | 50 | medium | 1 | 0 | 1 | 0 | 0 | 2026-04-12T03:40:32.027813 | 0.4.0 | https://huggingface.co/meta-llama/Llama-2-7b-hf | {} | {} | [] | [] |
mistralai/Mistral-7B-v0.1 | 50 | medium | 1 | 0 | 1 | 0 | 0 | 2026-04-12T03:50:28.238539 | 0.4.0 | https://huggingface.co/mistralai/Mistral-7B-v0.1 | {} | {} | [] | [] |
tiiuae/falcon-7b | 50 | medium | 1 | 0 | 1 | 0 | 0 | 2026-04-12T04:00:22.738340 | 0.4.0 | https://huggingface.co/tiiuae/falcon-7b | {} | {} | [] | [] |
EleutherAI/gpt-neo-2.7B | 50 | medium | 1 | 0 | 1 | 0 | 0 | 2026-04-12T04:08:24.888148 | 0.4.0 | https://huggingface.co/EleutherAI/gpt-neo-2.7B | {} | {} | [] | [] |
codellama/CodeLlama-7b-hf | 50 | medium | 1 | 0 | 1 | 0 | 0 | 2026-04-12T04:17:47.710648 | 0.4.0 | https://huggingface.co/codellama/CodeLlama-7b-hf | {} | {} | [] | [] |
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 metricssarif/: 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
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)
- Downloads last month
- 111