MEDLINE-S1-LT

MEDLINE-S1-LT is a Lithuanian translation of pritamdeka/MEDLINE-S1. MEDLINE-S1 is a set of typed decisions over PubMed abstracts. Each abstract has several questions about it: the study design, the main disease area, and yes/no questions about humans, animals, adverse effects, drug therapy and two age or sex groups. Every label comes from NLM's MeSH indexing and publication types, so no language model was involved in labelling. It is the training and evaluation set introduced in From Retrieval to Typed Decisions: Calibrated System One Models from Biomedical Sentence Encoders. Code is available at https://github.com/pritamdeka/sbert2s1.

The original Hub release contains PMIDs, questions and labels, but no abstract text. This release keeps those records and adds the Lithuanian title and abstract in state. Question ids, option keys and gold labels are unchanged, so a prediction can be scored against the original dataset.

Translation

Titles and abstracts (state) were translated from English into Lithuanian with facebook/nllb-200-distilled-1.3B.

Hub id	facebook/nllb-200-distilled-1.3B
Revision	7be3e24664b38ce1cac29b8aeed6911aa0cf0576
Direction	eng_Latn → lit_Latn
Weights	float16
Decoding	greedy, num_beams=1
Length	tokenizer and generation max_length=512

NLLB did not translate the questions. Instructions and option descriptions come from a closed Lithuanian glossary. Option keys stay in English (systematic_review, Neoplasms, true, false, and the other original keys). Yes/no questions also carry Lithuanian criteria for true and false; the original release left those descriptions implicit.

1,239 abstracts (2.6%) were dropped because the NLLB output collapsed into repetition. Their decisions were dropped with them. PMIDs are not reused across splits. manifest.json records the written counts, the question counts, the SHA-256 of each split file, and the PubMed baseline files of the parent English split.

Splits

Split	Abstracts	Decisions	English source	Source baseline files
train	38,950	236,720	40,000 / 242,986	14 files of the 2026 PubMed baseline
validation	1,964	11,930	2,000 / 12,138	same 14 files
calibration	390	2,352	400 / 2,418	same 14 files (used only to fit temperatures)
test	4,857	29,654	5,000 / 30,522	3 other files, disjoint from training
Total	46,161	280,656	47,400 / 288,064

The validation file is data/dev.jsonl and the calibration file is data/calib.jsonl, as in the original release.

Questions

Question	Type	Label source	Asked for
design	choice of 6	publication types (+ observational-design MeSH); precedence systematic review/meta-analysis > RCT > other trial > case report > observational > narrative review	abstracts with an unambiguous design
disease_area	choice	the single MeSH C-tree category among the major-topic descriptors, against 3 to 7 random other categories	abstracts with exactly one such category
humans, animals	yes/no	MeSH check tags	all
adverse	yes/no	any heading with the subheading adverse effects	all
drug_therapy	yes/no	any heading with the subheading drug therapy	all
group_* (infants, children, adolescents, adults, older adults, female, male)	yes/no	MeSH check tags	2 sampled per abstract about humans

Only citations from 2005 to 2021 with manual or curated (not fully automated) MeSH indexing and an abstract of at least 400 characters are included. PMIDs in any dev or test set of the companion BioDecide suite are excluded. Those filters were applied by the original MEDLINE-S1 build. This translation does not add citations.

Format

Each line is one abstract:

{"id": "medline_s1:train:000000", "task": "medline_s1", "split": "train", "pmid": 29101975, "year": 2017,
 "state": "<Lithuanian title> <Lithuanian abstract>",
 "state_sha256": "ac1a8ae378d139a2...",
 "questions": "{\"design\": {\"type\": \"choice\", \"instructions\": \"Koks tai tyrimo ar straipsnio tipas?\", \"criteria\": {...}}, \"humans\": {\"type\": \"noul\", ...}}",
 "gold": "{\"design\": {\"probabilities\": {\"narrative_review\": 1.0}}, \"humans\": {\"probabilities\": {\"true\": 1.0}}, ...}",
 "language": "lt",
 "translation_model": "facebook/nllb-200-distilled-1.3B"}

questions and gold are JSON strings, because different abstracts have different questions. Decode them with json.loads. noul is a yes/no question with the options true and false. Gold labels are distributions over the option keys. state is the Lithuanian title, a space, and the Lithuanian abstract.

state_sha256 is the SHA-256 of the English title and abstract in the original MEDLINE-S1 record, not of state. It joins a row to the source release. It does not verify the Lithuanian text. The file hashes in manifest.json do.

Source dataset

Do not use this folder to reconstruct the English abstracts. The original release omits them. To restore English titles and abstracts from PubMed, use the original repository:

https://huggingface.co/datasets/pritamdeka/MEDLINE-S1

That repository ships hydrate.py. This translation does not. E-utilities returns the current citation, so a restored English abstract can differ from the text that was translated here. state_sha256 is the check for the English text used by the paper.

Licence and attribution

The questions, labels and split assignments are released under CC BY 4.0, the same licence as pritamdeka/MEDLINE-S1. The original Hub repository declares that licence in the dataset card and does not ship a separate licence file. LICENSE in this folder is the Creative Commons Attribution 4.0 International legal text.

The labels are derived from MEDLINE/PubMed citation data and MeSH.

This product uses publicly available data from the U.S. National Library of Medicine (NLM), National Institutes of Health, Department of Health and Human Services; NLM is not responsible for the product and does not endorse or recommend this or any other product. The data come from the 2026 MEDLINE/PubMed baseline and may not reflect the most current information available from NLM.

Please cite the original dataset and paper, and name the translation model:

• Pritam Deka, MEDLINE-S1, https://huggingface.co/datasets/pritamdeka/MEDLINE-S1
• Pritam Deka, From Retrieval to Typed Decisions: Calibrated System One Models from Biomedical Sentence Encoders, arXiv:2610.02486, https://arxiv.org/abs/2610.02486
• Titles and abstracts translated with facebook/nllb-200-distilled-1.3B (revision 7be3e24664b38ce1cac29b8aeed6911aa0cf0576), NLLB Team, arXiv:2207.04672

The original English release does not redistribute abstracts. Their copyright stays with the publishers or authors. This release contains machine translations of titles and abstracts that were retrieved from NLM. Copyright in the underlying English text remains with the publishers or authors. NLM's terms of use apply to that text and to these translations.

Citation

@misc{deka2026retrievaltypeddecisionscalibrated,
      title={From Retrieval to Typed Decisions: Calibrated System One Models from Biomedical Sentence Encoders},
      author={Pritam Deka},
      year={2026},
      eprint={2610.02486},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2610.02486},
}

@article{nllbteam2022language,
      title={No Language Left Behind: Scaling Human-Centered Machine Translation},
      author={{NLLB Team}},
      journal={arXiv preprint arXiv:2207.04672},
      year={2022},
      url={https://arxiv.org/abs/2207.04672}
}