---
task_categories:
  - text-classification
tags:
  - data-use
  - gate
  - domain
  - multi-label
license: cc-by-4.0
configs:
  - config_name: default
    data_files:
      - split: train
        path: encoder_train.jsonl
      - split: val
        path: encoder_val.jsonl
      - split: holdout
        path: encoder_holdout.jsonl
---
# datause-encoder-data

Training data for the data-use encoder: a page-level `has_data` gate plus a
document-level `teratopic` domain classifier, in one joint dataset.

Columns (same schema on every row):
- `task` — `gate` (page-level binary) or `domain` (document-level multi-label)
- `doc_id` — source document id
- `text` — page text (gate) or title+abstract (domain)
- `has_data` — 0/1 for gate rows (0 placeholder on domain rows)
- `labels` — `teratopic` label list for domain rows (empty on gate rows)

Splits are document-id disjoint across both tasks. The 30 teratopic labels
(in column order) are in `encoder_labels.json`.