Hugging Face's logo

Datasets:
intertwine-ai
/
security-verifiers-e1-metadata

Tasks:
Text Classification
Modalities:
Text
Formats:
parquet
Languages:
English
Size:
< 1K
Tags:
security
rl
network-security
anomaly-detection
verifiers
metadata-only
Libraries:
Datasets
pandas
License:
Dataset card Data Studio Files
xet
Community
1
Dataset Viewer
Auto-converted to Parquet
section
stringclasses
4 values
name
stringclasses
6 values
description
stringclasses
6 values
payload_json
stringclasses
6 values
version
stringclasses
1 value
created_at
stringdate
2025-10-23 02:16:29
2025-10-23 02:16:29
sampling
iot23-train-dev-test
IoT-23 primary dataset sampling metadata (train/dev/test splits)
{"mode":"full","seed":42,"total":1800,"by_split":{"train":1253,"dev":269,"test":278}}
v1
2025-10-23T02:16:29.295365Z
sampling
e1-ood-datasets
Out-of-distribution datasets sampling metadata (CIC-IDS-2017, UNSW-NB15)
{"mode":"full","seed":42,"n_requested":600,"deduplication":"content-based (5-tuple + bytes + duration + state + service)","datasets":{"cic-ids-2017":{"hf_id":"bvk/CICIDS-2017","total":600,"labels":{"Malicious":300,"Benign":300},"attack_families":{"NA":600},"unique_hashes":600},"unsw-nb15":{"hf_id":"Mireu-Lab/UNSW-NB15","total":600,"labels":{"Malicious":330,"Benign":270},"attack_families":{"Normal":270,"Exploits":178,"Fuzzers":81,"DoS":41,"Reconnaissance":9,"Backdoor":5,"Generic":8,"Analysis":4,"Shellcode":3,"Worms":1},"unique_hashes":600}}}
v1
2025-10-23T02:16:29.295365Z
ood
cic-ids-2017-ood
CIC-IDS-2017 out-of-distribution test set (600 samples)
{"hf_id":"bvk/CICIDS-2017","total":600,"labels":{"Malicious":300,"Benign":300},"attack_families":{"NA":600},"unique_hashes":600}
v1
2025-10-23T02:16:29.295365Z
ood
unsw-nb15-ood
UNSW-NB15 out-of-distribution test set (600 samples)
{"hf_id":"Mireu-Lab/UNSW-NB15","total":600,"labels":{"Malicious":330,"Benign":270},"attack_families":{"Normal":270,"Exploits":178,"Fuzzers":81,"DoS":41,"Reconnaissance":9,"Backdoor":5,"Generic":8,"Analysis":4,"Shellcode":3,"Worms":1},"unique_hashes":600}
v1
2025-10-23T02:16:29.295365Z
provenance
dataset-sources
Original dataset sources and references
{"iot23":{"source":"Avast AIC IoT-23 dataset","url":"https://www.stratosphereips.org/datasets-iot23","hf_id":"stratosphereips/iot-23"},"cic-ids-2017":{"source":"Canadian Institute for Cybersecurity IDS 2017","url":"https://www.unb.ca/cic/datasets/ids-2017.html","hf_id":"bvk/CICIDS-2017"},"unsw-nb15":{"source":"UNSW-NB15 Network Intrusion Dataset","url":"https://research.unsw.edu.au/projects/unsw-nb15-dataset","hf_id":"Mireu-Lab/UNSW-NB15"}}
v1
2025-10-23T02:16:29.295365Z
notes
privacy-rationale
Explanation of private dataset hosting to prevent training contamination
{"reason":"Training contamination prevention","details":"Datasets are private to preserve benchmark validity and enable fair model comparisons over time.","access":"Request access via GitHub Issues: https://github.com/intertwine/security-verifiers/issues"}
v1
2025-10-23T02:16:29.295365Z

🔒 Security Verifiers E1: Network Log Anomaly Detection (Public Metadata)

⚠️ This is a PUBLIC metadata-only repository. The full datasets are hosted privately to prevent training contamination. See below for access instructions.

Overview

E1 is a network log anomaly detection environment with calibrated classification and abstention. This repository contains only the sampling metadata that describes how the private datasets were constructed.

Why Private Datasets?

Training contamination is a critical concern for benchmark integrity. If datasets leak into public training corpora:

  • Models can memorize answers instead of learning to reason
  • Evaluation metrics become unreliable
  • Research reproducibility suffers
  • True capabilities become obscured

By keeping evaluation datasets private with gated access, we:

  • ✅ Preserve benchmark validity over time
  • ✅ Enable fair model comparisons
  • ✅ Maintain research integrity
  • ✅ Allow controlled access for legitimate research

Dataset Composition

The private E1 datasets include:

Primary Dataset: IoT-23

  • Samples: 1,800 network flows (train/dev/test splits)
  • Source: IoT-23 botnet dataset
  • Features: Network flow statistics, timestamps, protocols
  • Labels: Benign vs Malicious with confidence scores
  • Sampling: Stratified by label and split

Out-of-Distribution Datasets

  • CIC-IDS-2017: 600 samples (different attack patterns)
  • UNSW-NB15: 600 samples (different network environment)
  • Purpose: Test generalization and OOD detection

What's in This Repository?

This public repository contains:

  1. Sampling Metadata (sampling-*.json):

    • Dataset versions and sources
    • Sampling strategies and random seeds
    • Label distributions
    • Split ratios
    • Reproducibility parameters

  2. Tools Versions (referenced in metadata):

    • Exact versions of all preprocessing tools
    • Dataset library versions
    • Python environment specifications

  3. This README: Instructions for requesting access

Reward Components

E1 uses composable reward functions:

  • Accuracy: Correctness of malicious/benign classification
  • Calibration: Alignment between confidence and actual accuracy
  • Abstention: Reward for declining on uncertain examples
  • Asymmetric Costs: Higher penalty for false negatives (security context)

Requesting Access

🔑 To access the full private datasets:

  1. Open an access request issue: Security Verifiers Issues
  2. Use the title: "Dataset Access Request: E1"
  3. Include:
    • Your name and affiliation
    • Research purpose / use case
    • HuggingFace username
    • Commitment to not redistribute or publish the raw data

Approval criteria:

  • Legitimate research or educational use
  • Understanding of contamination concerns
  • Agreement to usage terms

We typically respond within 2-3 business days.

Citation

If you use this environment or metadata in your research:

@misc{security-verifiers-2025,
  title={Open Security Verifiers: Composable RL Environments for AI Safety},
  author={intertwine},
  year={2025},
  url={https://github.com/intertwine/security-verifiers},
  note={E1: Network Log Anomaly Detection}
}

Related Resources

License

MIT License - See repository for full terms.

Contact

Built with ❤️ for the AI safety research community

Downloads last month
38
Size of downloaded dataset files:
5.61 kB
Size of the auto-converted Parquet files:
5.61 kB
Number of rows:
6