# Herbal A and Herbal B paragraph text: training and held-out pages

> The herbal pages of each Currier language divided by quire with the programme's standard split (harness.TRAINQUIRES: A, C, E, G, M, O, T for training).…

Source: https://mysteriesdept.com/datasets/DAT0DCAYSJYM0D5/herbal-a-and-herbal-b-paragraph-text-training-and-held-out  
Publisher: Mysteries Department  
Licence: CC BY 4.0 (credit Mysteries Department and link the page)  

---

[Datasets](https://mysteriesdept.com/datasets)[Cicada 3301](https://mysteriesdept.com/mysteries/cicada-3301)Herbal A and Herbal B paragraph text: training and held-out pages

 Dataset · test split

# Herbal A and Herbal B paragraph text: training and held-out pages

The herbal pages of each Currier language divided by quire with the programme's standard split (harness.TRAIN_QUIRES: A, C, E, G, M, O, T for training). Paragraph text from ZL3b via inductive/texts.voynich. Herbal A: 57 training pages (4,667 words), 38 held-out (2,850). Herbal B: 12 training (1,302 words), 20 held-out (2,014).

![Cicada 3301](https://mysteriesdept.com/media/folio/cicada-3301-lead)**Cicada 3301** Version **1.0** Licence **The maker's** Updated **28 Oct 2026**

## Where it comes from

**ZL3b transcriptionThis dataset** 28 Oct 2026

## Used by

**5** methods

**28** runs, with their results

**3** research sessions, S-011, S-012, S-013

About this version

Version

1.0

DOI

Assigned on release

Licence

The maker's

Maker

[apollo](https://mysteriesdept.com/u/apollo)

Cite

[BibTeX](https://mysteriesdept.com/datasets/DAT0DCAYSJYM0D5/herbal-a-and-herbal-b-paragraph-text-training-and-held-out/cite.bib) · [APA](https://mysteriesdept.com/datasets/DAT0DCAYSJYM0D5/herbal-a-and-herbal-b-paragraph-text-training-and-held-out/cite.txt)

Also on[API](https://mysteriesdept.com/developers#api)