README.md

---
dataset_info:
  features:
  - name: artifact_name
    dtype: large_string
  - name: org
    dtype: large_string
  - name: artifact_type
    dtype: large_string
  - name: languages
    list: string
  - name: license
    dtype: large_string
  - name: size_category
    dtype: large_string
  - name: hf_task_categories
    list: string
  - name: arxiv_ids
    list: string
  - name: data_role
    dtype: large_string
  - name: modality
    dtype: large_string
  - name: topic
    list: string
  - name: format
    list: string
  - name: per_language
    list:
    - name: lang
      dtype: string
    - name: provenance
      dtype: string
    - name: in_language_verified
      dtype: string
  - name: has_human_baseline
    dtype: large_string
  - name: derived_from_public_source
    dtype: large_string
  - name: evidence
    struct:
    - name: data_role
      dtype: string
    - name: topic
      dtype: string
    - name: format
      dtype: string
    - name: per_language
      dtype: string
    - name: modality
      dtype: string
    - name: has_human_baseline
      dtype: string
    - name: derived_from_public_source
      dtype: string
    - name: in_language_verified
      dtype: string
    - name: provenance
      dtype: string
    - name: topic:Crime & Law
      dtype: string
    - name: topic:Food & Dining
      dtype: string
    - name: topic:Art & Design
      dtype: string
    - name: topic:Science & Technology
      dtype: string
    - name: topic:History
      dtype: string
    - name: topic:Health
      dtype: string
    - name: topic:Literature
      dtype: string
    - name: format:Structured Data
      dtype: string
    - name: format:Knowledge Article
      dtype: string
    - name: format:Creative Writing
      dtype: string
    - name: format:Tutorial
      dtype: string
    - name: format:Q&A Forum
      dtype: string
    - name: per_language_provenance
      dtype: string
  - name: low_confidence_fields
    list: string
  - name: needs_human_review
    dtype: bool
  - name: agent_trace
    struct:
    - name: artifact_name
      dtype: string
    - name: model
      dtype: string
    - name: messages
      list:
      - name: role
        dtype: string
      - name: content
        dtype: string
      - name: tool_calls
        list:
        - name: index
          dtype: int64
        - name: function
          struct:
          - name: arguments
            dtype: string
          - name: name
            dtype: string
        - name: id
          dtype: string
        - name: type
          dtype: string
      - name: function_call
        dtype: 'null'
      - name: reasoning_content
        dtype: string
      - name: provider_specific_fields
        struct:
        - name: refusal
          dtype: 'null'
        - name: reasoning_content
          dtype: string
      - name: tool_call_id
        dtype: string
  splits:
  - name: train
    num_bytes: 355792473
    num_examples: 6410
  download_size: 351553575
  dataset_size: 355792473
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---