Datasetpublic

NewSonnet/uof-grapheme-crime-scenes

UOF Grapheme Segmentation Scenes v2 Synthetic extended-grapheme cases with escaped input, cluster targets, code-point counts, UTF-8 byte counts, and cursor/backspace step estimates. Segmentation uses the regex \X reference with version 2025.11.3. Cluster strings are stored in escaped form so that combining marks, joiners, variation selectors, and supplementary-plane characters remain inspectable in JSON. Release contract Generator: uof-frontier-v2.0.0 Regex… See the full description on the dataset page: https://huggingface.co/datasets/NewSonnet/uof-grapheme-crime-scenes.

sourceHugging Facecc-by-4.0updated 12d agoView on Hugging Face

0likes76downloads

Dataset Card

UOF Grapheme Segmentation Scenes v2

Synthetic extended-grapheme cases with escaped input, cluster targets, code-point counts, UTF-8 byte counts, and cursor/backspace step estimates.

Segmentation uses the regex \X reference with version 2025.11.3. Cluster strings are stored in escaped form so that combining marks, joiners, variation selectors, and supplementary-plane characters remain inspectable in JSON.

Release contract

—Generator: uof-frontier-v2.0.0
—Regex oracle version: 2025.11.3
—Seed: 20260915
—Rows: 1,440 (train 840; validation 120; test 240; ood 240)
—Split isolation: the publication gate found no duplicate IDs and no root overlap across splits.

Limitations

The \X result is a reference segmentation, not a guarantee that every editor, terminal, browser, or UI uses the same cursor model. Cursor and backspace values are estimates derived from the generated clusters. Confirm behavior in the target text stack before treating them as interaction measurements.

Load

python

from datasets import load_dataset
ds = load_dataset("NewSonnet/uof-grapheme-crime-scenes")

License

CC BY 4.0.