Datasets:

mrapacz
/

greek-interlinear-translations

license: cc-by-nc-sa-4.0
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: validation
        path: data/validation-*
      - split: test
        path: data/test-*
dataset_info:
  features:
    - name: source_normalized
      sequence: string
    - name: source_diacritics
      sequence: string
    - name: tag_oblubienica
      sequence: string
    - name: tag_biblehub
      sequence: string
    - name: target_pl
      sequence: string
    - name: target_en
      sequence: string
    - name: book
      dtype: int64
    - name: chapter
      dtype: int64
    - name: verse
      dtype: int64
    - name: book_name_pl
      dtype: string
    - name: book_name_en
      dtype: string
  splits:
    - name: train
      num_bytes: 8491619
      num_examples: 6352
    - name: validation
      num_bytes: 1062996
      num_examples: 794
    - name: test
      num_bytes: 1066036
      num_examples: 794
  download_size: 2799395
  dataset_size: 10620651
tags:
  - interlinear-translation
task_categories:
  - translation
language:
  - en
  - pl
pretty_name: Interlinear Translations of the Greek New Testament

Dataset Card for Ancient Greek Interlinear Translations Dataset

This dataset provides word-level aligned interlinear translations of the New Testament from Ancient Greek to English and Polish with morphological tags sourced from Oblubienica (https://biblia.oblubienica.pl) and BibleHub (https://biblehub.com/interlinear).

See https://github.com/mrapacz/loreslm-interlinear-translation for more details.

Dataset Details

Dataset Description

The dataset contains interlinear translations where each Greek word is paired with its corresponding Polish and English translations, along with morphological tags from two different annotation systems. We applied a set of heuristics to align the corpora at the word level. The alignment process achieved over 99% word matching accuracy, with unmatched words being excluded. Subsequently, we trimmed all verses so that the least memory-efficient models tested in our research could encode them within the chosen limit of 512 tokens.

Curated by: Maciej Rapacz
Language(s): Ancient Greek, English, Polish
License: CC BY-NC-SA 4.0

Dataset Sources

Repository: https://huggingface.co/datasets/mrapacz/greek-interlinear-translations
Source Texts:
- English interlinear translation from BibleHub (NA27 critical edition) - https://biblehub.com/interlinear
- Polish interlinear translation from Oblubienica (NA28 critical edition) - https://biblia.oblubienica.pl

Dataset Structure

The dataset is divided into:

Training: 6,352 verses (80%)
Validation: 794 verses (10%)
Test: 794 verses (10%)

Each entry contains:

source_diacritics: Greek text with diacritics (BibleHub source)
source_normalized: Normalized Greek text (lowercase, no diacritics)
tag_biblehub: BibleHub morphological tags
tag_oblubienica: Oblubienica morphological tags
target_pl: Polish translation sourced from Oblubienica
target_en: English translation sourced from BibleHub
book: Book number
chapter: Chapter number
verse: Verse number
book_name_pl: Book name in Polish
book_name_en: Book name in English

Dataset Card Authors

Maciej Rapacz

Citation

@inproceedings{rapacz-smywinski-pohl-2025-low,
    title = "Low-Resource Interlinear Translation: Morphology-Enhanced Neural Models for {A}ncient {G}reek",
    author = "Rapacz, Maciej  and
      Smywi{\'n}ski-Pohl, Aleksander",
    editor = "Hettiarachchi, Hansi  and
      Ranasinghe, Tharindu  and
      Rayson, Paul  and
      Mitkov, Ruslan  and
      Gaber, Mohamed  and
      Premasiri, Damith  and
      Tan, Fiona Anting  and
      Uyangodage, Lasitha",
    booktitle = "Proceedings of the First Workshop on Language Models for Low-Resource Languages",
    month = jan,
    year = "2025",
    address = "Abu Dhabi, United Arab Emirates",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2025.loreslm-1.11/",
    pages = "145--165",
    abstract = "Contemporary machine translation systems prioritize fluent, natural-sounding output with flexible word ordering. In contrast, interlinear translation maintains the source text`s syntactic structure by aligning target language words directly beneath their source counterparts. Despite its importance in classical scholarship, automated approaches to interlinear translation remain understudied. We evaluated neural interlinear translation from Ancient Greek to English and Polish using four transformer-based models: two Ancient Greek-specialized (GreTa and PhilTa) and two general-purpose multilingual models (mT5-base and mT5-large). Our approach introduces novel morphological embedding layers and evaluates text preprocessing and tag set selection across 144 experimental configurations using a word-aligned parallel corpus of the Greek New Testament. Results show that morphological features through dedicated embedding layers significantly enhance translation quality, improving BLEU scores by 35{\%} (44.67 {\textrightarrow} 60.40) for English and 38{\%} (42.92 {\textrightarrow} 59.33) for Polish compared to baseline models. PhilTa achieves state-of-the-art performance for English, while mT5-large does so for Polish. Notably, PhilTa maintains stable performance using only 10{\%} of training data. Our findings challenge the assumption that modern neural architectures cannot benefit from explicit morphological annotations. While preprocessing strategies and tag set selection show minimal impact, the substantial gains from morphological embeddings demonstrate their value in low-resource scenarios."
}