Korean Learning Text Fact and Event Extraction Dataset

#named entity recognition #event extraction #evidence span identification #structured JSON generation #Korean language learning #Korean reading comprehension #educational text information extraction

5000
0.3GB
JSON
CATL
MOBIUSI INC

Updated:2026-11-05

AI Analysis & Value Prop

This dataset pairs short passages for Korean reading practice with corresponding JSON annotations. The annotations capture people, places, and events explicitly mentioned in the text, including event times, participants, locations, supporting text, and explicitly stated relations. It supports training models to extract facts from Korean passages and produce structured outputs, as well as reading comprehension practice, information extraction evaluation, and educational tool development.

Dataset Insights

Sample Examples

{}

Technical Specifications

Field	Type	Description
file_name	string	File name
events	array	Events described in the passage, including their explicit time, participants, location, and supporting text. Use an empty array if none are described.
people	array	People explicitly mentioned in the passage. Use an empty array if none are mentioned.
places	array	Places explicitly mentioned in the passage. Use an empty array if none are mentioned.
relations	array	Relations between people, places, or events that are explicitly expressed in the passage. Use an empty array if none are stated.
source_text	string	The original Korean passage to be analyzed.

Compliance Statement

Authorization Type	Proprietary - Commercial AI Training License (No Redistribution)
Commercial Use	Requires exclusive subscription or authorization contract (monthly or per-invocation charging)
Privacy and Anonymization	No PII, no real company names, simulated scenarios follow industry standards
Compliance System	Compliant with China's Data Security Law / EU GDPR / supports enterprise data access logs

Frequently Asked Questions

What does the dataset contain?: It contains short Korean passages paired with JSON annotations for people, places, and events, including explicitly stated event times, participants, locations, and relations. Items not mentioned in a passage are represented with empty arrays.

Which training and evaluation tasks does it support?: The data can support entity extraction, event extraction, relation extraction, and structured information generation from Korean passages. It can also evaluate whether a model extracts facts grounded in the text without adding unsupported information.

What are the dataset format and quality specifications?: Samples are provided as Korean source passages with structured JSON annotations and clearly defined fields, including textual evidence. Check the product page and delivered package for the exact sample count, full schema details, and quality specifications.

How can I access the data and verify its license?: Check the dataset product page for available access methods, package formats, and license terms. Review the specific license before use to confirm that it covers your intended research, educational, or commercial application.

Can't find the data you need?

Post a request and let data providers reach out to you.

Get this Dataset

Verified for Enterprise Use

Cite this Work

@dataset{Mobiusi2026,
  title={Korean Learning Text Fact and Event Extraction Dataset},
  author={MOBIUSI INC},
  year={2026},
  url={https://www.mobiusi.com/datasets/194697c7c955fdf73717ce018da0cd41},
  urldate={2026-10-19},
  keywords={Korean learning dataset,Korean reading comprehension,Korean information extraction,Korean event extraction,Korean entity recognition,structured output training,Korean text annotation,Korean NLP dataset},
  version={1.0}
}

Using this in research? Please cite us.

placeholder

Scene

Domain

Modality

Task