OPENZEKA/openzeka_2025_turkish_dictionary_dataset

Turkish Dictionary Dataset

Dataset Description

The Turkish Dictionary Dataset is a semi-synthetic dataset created from the Turkish Language Association's (TDK) official dictionary. The actual dictionary entries are processed and transformed into instruction-following format, generating structured instruction-output pairs for NLP training purposes.

Dataset Structure

The dataset contains three subsets in JSONL format:

Full Definitions: Dictionary definitions converted to instruction-response format Synonyms: Synonym relationships presented as instruction-based queries Example Sentences: Usage examples structured as conversational prompts All entries follow the standard instruct format:

{"instruction": "...", "input": "", "output": "..."}

Use Cases

• Fine-tuning language models on Turkish linguistic knowledge
• Training instruction-following models
• Semantic analysis and vocabulary expansion research
• Turkish language learning applications

Data Source

The base data is sourced from the Turkish Language Association's (Türk Dil Kurumu) Güncel Türkçe Sözlük, transformed into semi-synthetic instruction data through automated processing.

Citation

If you use this Turkish Dictionary Dataset in your research or work, please consider citing it:

@misc{yourusername_2025_turkish_dictionary_hf,
  author       = {Openzeka},
  title        = {Turkish Dictionary Dataset},
  year         = {2025},
  publisher    = {Hugging Face},
  howpublished = {\url{https://huggingface.co/datasets/yourusername/turkish-dictionary-dataset}}
}