README.md

---
dataset_info:
- config_name: cti_attack_techniques
  features:
  - name: id
    dtype: int64
  - name: technique_id
    dtype: string
  - name: name
    dtype: string
  - name: tactic
    dtype: string
  - name: revoked_by
    dtype: string
  - name: version
    dtype: 'null'
  splits:
  - name: train
    num_bytes: 55844
    num_examples: 858
  download_size: 24471
  dataset_size: 55844
- config_name: cti_items
  features:
  - name: id
    dtype: string
  - name: task
    dtype: string
  - name: external_id
    dtype: string
  - name: source
    dtype: string
  - name: input_text
    dtype: string
  - name: input_ref
    dtype: string
  - name: label
    dtype: string
  - name: label_provenance
    dtype: string
  - name: source_revision
    dtype: string
  - name: input_date
    dtype: string
  - name: label_date
    dtype: string
  - name: first_available_date
    dtype: string
  - name: authority_agreement
    dtype: string
  - name: withhold
    dtype: bool
  - name: status
    dtype: string
  - name: created_at
    dtype: string
  splits:
  - name: train
    num_bytes: 7752024
    num_examples: 4032
  download_size: 2197511
  dataset_size: 7752024
- config_name: cti_kev
  features:
  - name: id
    dtype: int64
  - name: cve_id
    dtype: string
  - name: date_added
    dtype: string
  - name: vendor
    dtype: string
  - name: product
    dtype: string
  splits:
  - name: train
    num_bytes: 114295
    num_examples: 1622
  download_size: 43490
  dataset_size: 114295
- config_name: cti_results
  features:
  - name: id
    dtype: int64
  - name: run_id
    dtype: string
  - name: item_id
    dtype: string
  - name: model_id
    dtype: string
  - name: prompt_hash
    dtype: string
  - name: score
    dtype: float64
  - name: score_breakdown
    dtype: string
  - name: correct
    dtype: bool
  - name: pre_cutoff
    dtype: bool
  - name: created_at
    dtype: string
  splits:
  - name: train
    num_bytes: 752318
    num_examples: 2369
  download_size: 118407
  dataset_size: 752318
- config_name: cti_runs
  features:
  - name: id
    dtype: string
  - name: model_id
    dtype: string
  - name: task
    dtype: string
  - name: triggered_by
    dtype: string
  - name: status
    dtype: string
  - name: model_cutoff_start
    dtype: string
  - name: model_cutoff_end
    dtype: string
  - name: item_count
    dtype: int64
  - name: scored_count
    dtype: int64
  - name: prequential_score
    dtype: float64
  - name: config
    dtype: string
  - name: started_at
    dtype: string
  - name: completed_at
    dtype: string
  - name: created_at
    dtype: string
  splits:
  - name: train
    num_bytes: 17680
    num_examples: 72
  download_size: 13951
  dataset_size: 17680
- config_name: cti_threat_actors
  features:
  - name: id
    dtype: int64
  - name: canonical_name
    dtype: string
  - name: aliases
    dtype: string
  - name: techniques
    dtype: 'null'
  - name: related_groups
    dtype: string
  splits:
  - name: train
    num_bytes: 54013
    num_examples: 1010
  download_size: 34523
  dataset_size: 54013
- config_name: models
  features:
  - name: id
    dtype: string
  - name: name
    dtype: string
  - name: provider
    dtype: string
  - name: cutoff_start
    dtype: string
  - name: cutoff_end
    dtype: string
  - name: created_at
    dtype: string
  splits:
  - name: train
    num_bytes: 6715
    num_examples: 50
  download_size: 6937
  dataset_size: 6715
configs:
- config_name: cti_attack_techniques
  data_files:
  - split: train
    path: cti_attack_techniques/train-*
- config_name: cti_items
  data_files:
  - split: train
    path: cti_items/train-*
- config_name: cti_kev
  data_files:
  - split: train
    path: cti_kev/train-*
- config_name: cti_results
  data_files:
  - split: train
    path: cti_results/train-*
- config_name: cti_runs
  data_files:
  - split: train
    path: cti_runs/train-*
- config_name: cti_threat_actors
  data_files:
  - split: train
    path: cti_threat_actors/train-*
- config_name: models
  data_files:
  - split: train
    path: models/train-*
---