Turkish-English Parallel Words and Phrases with 3.3 Million AI-Generated Entries

NumanKaanKaratas's Turkish-English Words dataset contains 3,365,067 parallel translation entries. The data was generated by ChatGPT (OpenAI) and covers a range of Turkish vocabulary from root words to complex agglutinated forms and idiomatic expressions. The dataset was last updated on 2026-05-27 and is provided in JSONL format under an MIT license.

Use Cases

• Train machine translation models based on the 3.3 million parallel entries.
• Build bilingual dictionaries or lookup tools based on the word and phrase translations.
• Evaluate the quality of AI-generated translations for Turkish-English language pairs.
• Study Turkish morphology and agglutination based on the described coverage of complex forms.

Strengths

• Contains 3,365,067 parallel entries, providing a substantial volume of translation pairs.
• Covers a described range of Turkish vocabulary, including complex agglutinated forms and idiomatic expressions.

Limitations

• Translations are fully AI-generated with no human review, which may impact accuracy.
• Column-level documentation is absent; field semantics must be inferred after download.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

NumanKaanKaratas on Hugging Face

Collection Method

AI-generated by ChatGPT (OpenAI)

Freshness

Last updated 2026-05-27 16:12:23; freshness should be verified.