Share your thoughts, 1 month free Claude Pro on usSee more

Home/Benchmarks

Toxic-neutral pair quality evaluation on KODOLI

3.293Overall O.

KODOLI

Updated 1mo ago

Evaluation Results

Method	Links
KODOLI 2025.06		3.293	2.554

SOTA Paper

KODOLI

K/DA: Automated Data Generation Pipeline for Detoxifying Implicitly Offensive Language in Korean

Dataset

KODOLI

Follow for update

@wizwand_team Discord

Related Benchmarks

Language Detoxification on Ours (test)Language Detoxification on KOLD (test)Language Detoxification on BEEP (test)Human Evaluation on K/DA and K-OMG (50 random samples)Detoxification Dataset Quality Evaluation on ParaDetox 500 neutral-toxic pairs

© 2026 wizwand

Blog Contact Changelog Swarm

Privacy Policy Terms of Service FAQs Swarm Docs