datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
sklep
Dataset Card for skLEP
Dataset Description
skLEP (General Language Understanding Evaluation benchmark for Slovak) is the first comprehensive benchmark specifically designed for evaluating Slovak natural language understanding (NLU) models. The benchmark encompasses nine diverse tasks that span token-level, sentence-pair, and document-level challenges, thereby offering a thorough assessment of model capabilities.
To create this benchmark, we curated new, original datasets… See the full description on the dataset page: https://huggingface.co/datasets/slovak-nlp/sklep.hate_speech_slovak
Slovak Hate Speech and Offensive Language Database
The dataset contains posts from a social network with human annotations.
Annotations
The posts are marked 1 if the post contain hateful or offensive language, 0 otherwise.
Dataset Creation
The source data were scraped from a social network from a selection of public pages for sport, politics or general discussion. The gathered data were cleaned from span with a text clustering.
The posts were annotated by a… See the full description on the dataset page: https://huggingface.co/datasets/TUKE-KEMT/hate_speech_slovak.slovak-sft
Slovak SFT Dataset
A supervised fine-tuning (SFT) dataset for Slovak language instruction following, constructed from two publicly available Slovak resources:
saillab/alpaca-slovak-cleaned — Slovak instruction-response pairs
TUKE-DeutscheTelekom/skquad — Slovak question answering, rewritten into chat-style prompts
Format
Each example follows the standard messages format with three turns:
{
"messages": [
{"role": "system", "content": "Si užitočný slovenský… See the full description on the dataset page: https://huggingface.co/datasets/mbenco/slovak-sft.slovak-web-qa-pairsslovak-triplets
Slovak Triplets Dataset
This repository contains the Slovak Triplets Dataset, a collection of triplet sentences in the Slovak language designed for training and evaluating document embedding models. Each triplet consists of an anchor sentence, a positive sentence (similar to the anchor), and a negative sentence (dissimilar to the anchor).
Data Source
The dataset is extracted form the Slovak part of WebFAQ and MQA datasets, which are publicly available collections of… See the full description on the dataset page: https://huggingface.co/datasets/TUKE-KEMT/slovak-triplets.slovak-sft
Slovak SFT Dataset
A supervised fine-tuning (SFT) dataset for Slovak language instruction following, constructed from two publicly available Slovak resources:
saillab/alpaca-slovak-cleaned — Slovak instruction-response pairs
TUKE-DeutscheTelekom/skquad — Slovak question answering, rewritten into chat-style prompts
Format
Each example follows the standard messages format with three turns:
{
"messages": [
{"role": "system", "content": "Si užitočný slovenský… See the full description on the dataset page: https://huggingface.co/datasets/Adamos3000/slovak-sft.exam-slovak-mathbioPart of INCLUDE
see research: https://arxiv.org/abs/2411.19799
full dataset: https://huggingface.co/datasets/CohereForAI/include-base-44
Slovak_Relation_Extractionuner_llm_inst_slovak
Dataset Card for Universal NER v1 in the Aya format - Slovak subset
This dataset is a format conversion for the Slovak data in the original Universal NER v1 into the Aya instruction format and it's released here under the same CC-BY-SA 4.0 license and conditions.
The dataset contains different subsets and their dev/test/train splits, depending on language. For more details, please refer to:
Dataset Details
For the original Universal NER dataset v1 and more details, please… See the full description on the dataset page: https://huggingface.co/datasets/universalner/uner_llm_inst_slovak.SlovakBebrasChallenge
iBobor Slovak Questions Dataset
Dataset Description
This dataset contains questions from the Slovak iBobor competition, publicly available on the website http://demo.ibobor.sk/sutaz_demo/, which is part of the international Bebras Challenge focused on informatics and computational thinking.
The dataset was created for the purposes of a master's thesis at Comenius University in Bratislava, Faculty of Mathematics, Physics and Informatics. The thesis is concerned with… See the full description on the dataset page: https://huggingface.co/datasets/patriciavnencakova/SlovakBebrasChallenge.SlovakBabyLMSKJ1To appear in Marek Dobeš, Peter Bednár: SKJ1 - Benchmark for Slovak Language Capabilities of Large Language Models. TCL 2026
Tasks in these categories:
Gender agreement
Noun inflection
Use of reflective particles “sa” and “si”
Noun misspellings
Use of opening and closing quotation marks
Bohemisms
Anglicisms
Stylistics
Mix of errors
slovak-sts-synthetic
Dataset Card for slovak-sts-synthetic
Dataset Details
Dataset Description
This is a synthetic Slovak dataset intended for the semantic textual similarity (STS) task.
Each pair consists of a sentence from the Slovak Summarization dataset and a generated sentence produced by GPT-5.
GPT-5 was instructed to generate six sentences per original sentence (from SlovakSum) to each similarity score (0-5). Each original sentence had to contain at least 60… See the full description on the dataset page: https://huggingface.co/datasets/slovak-nlp/slovak-sts-synthetic.slovak-financial-exam
Dataset Card for Slovak Financial QA
Dataset Description
This dataset contains 1,334 multiple-choice questions from the financial domain in the Slovak language. It was created to address the limited availability of language resources for Slovak, providing a benchmark for evaluating language models' capabilities in a specialized, low-resource domain.
The questions are sourced from the official certification exams for financial advisors in Slovakia, covering a range of… See the full description on the dataset page: https://huggingface.co/datasets/TUKE-KEMT/slovak-financial-exam.slovak-tax-accounting-laws-jsonl-# Slovak Tax & Accounting Legislation (JSONL)
Clean, structured, and chunked dataset of core Slovak financial legislation. Prepared specifically for RAG (Retrieval-Augmented Generation), vector databases, and LLM fine-tuning.
What's in this sample?
A small preview (ukazkovy_dataset.jsonl) extracted from the main dataset.
Each entry is a JSON object with the following schema:
zakon: Name and number of the law
paragraf: Paragraph ID (§)
text: Raw text content of the paragraph… See the full description on the dataset page: https://huggingface.co/datasets/datasetrura1/slovak-tax-accounting-laws-jsonl.Slovak_name_entity_recognition