KOLD

When this dataset fits

KOLD is a Korean Toxicity Dataset from GitHub. It fits the korean toxicity dataset task. Do not mix it with a benchmark download. The catalog does not claim a license.

Task / category

Text · Text · hate-speech, safety, classification

Language

Korean

Format · Size

Text · Unknown (40,429 comments)

License

License unknown(No confirmed license in catalog)

Source & original link

GitHub · Original dataset link(Verified)

Verification status

Source: Verified · License: Unknown · Last checked: 2026-09-03

KOLD: Korean Toxicity Dataset from GitHub. License license on original source. Open the original link to download.

KOLD (2022). GitHub. license label: License unknown catalog commercial mark: conditional Original: https://github.com/boychaboy/KOLD. GearDel catalog: https://geardel.com/en/datasets/kold. Last checked: 2026-09-03. Author names are not stored in the GearDel catalog and are not invented here.

Differences below use catalog fields only: license, commercial mark, provider, size text, and year.

• APEACHKorean LLM Benchmark Dataset · Hugging FaceViewLicense labels differ (License unknown vs CC BY-SA 4.0). Commercial marks differ (Commercial use may be conditional vs Commercial use allowed (catalog)). Providers differ (GitHub vs Hugging Face).
• K-HATERSKorean Toxicity Dataset · Hugging FaceViewLicense labels differ (License unknown vs CC BY 4.0). Commercial marks differ (Commercial use may be conditional vs Commercial use allowed (catalog)). Providers differ (GitHub vs Hugging Face).
• BEEP! Korean Hate SpeechKorean Toxicity Dataset · GitHubViewLicense labels differ (License unknown vs CC BY-SA 4.0). Catalog years differ (2022 vs 2020).
• K-MHaSKorean Toxicity Dataset · Hugging FaceViewLicense labels differ (License unknown vs CC BY-SA 4.0). Commercial marks differ (Commercial use may be conditional vs Commercial use allowed (catalog)). Providers differ (GitHub vs Hugging Face).