Add dataset card

README.md

@@ -1,19 +1,38 @@
 ---
-dataset_info:
-  features:
-  - name: image_id
-    dtype: string
-  - name: image
-    dtype: image
-  splits:
-  - name: train
-    num_bytes: 3679389
-    num_examples: 30
-  download_size: 3681319
-  dataset_size: 3679389
-configs:
-- config_name: default
-  data_files:
-  - split: train
-    path: data/train-*
+license: cc0-1.0
+task_categories:
+- image-to-text
+language:
+- en
+tags:
+- ocr
+- structured-extraction
+- kie
+- index-cards
+- glam
+- manuscript-catalogue
+- held-out
+- unlabeled
+size_categories:
+- n<1K
+source_datasets:
+- biglam/rubenstein-manuscript-catalog
+pretty_name: Rubenstein Manuscript Cards — Held-out Schema Probe
 ---
+
+# Rubenstein manuscript cards — held-out schema probe
+
+**30 manuscript-catalogue card images** from the Duke Rubenstein collection
+([`biglam/rubenstein-manuscript-catalog`](https://huggingface.co/datasets/biglam/rubenstein-manuscript-catalog)),
+used as a **held-out generalization probe**: a collection and schema the
+[index-card-extractor-4b](#) model was **never trained on**. Tests whether the model still emits
+valid, schema-conforming JSON for an unseen schema (it does — better than the base model).
+
+No ground truth — this measures schema-following (valid-JSON rate, key-conformance, field coverage),
+not field accuracy. A candidate Rubenstein extraction schema ships with the project code.
+
+## Columns
+`image_id` · `image`.
+
+## Provenance & license
+Images from `biglam/rubenstein-manuscript-catalog` (CC0). Released **CC0-1.0**.