Datasets:
Tasks:
Text Classification
Modalities:
Text
Formats:
parquet
Languages:
Korean
Size:
10K - 100K
ArXiv:
License:
Add eval config
Browse files
eval.yaml
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
name: KorHateSpeechMLClassification
|
| 2 |
+
description: 'The Korean Multi-label Hate Speech Dataset, K-MHaS, consists of 109,692
|
| 3 |
+
utterances from Korean online news comments, labelled with 8 fine-grained hate speech
|
| 4 |
+
classes (labels: Politics, Origin, Physical, Age, Gender, Religion, Race, Profanity)
|
| 5 |
+
or Not Hate Speech class. Each utterance provides from a single to four labels that
|
| 6 |
+
can handles Korean language patterns effectively. For more details, please refer
|
| 7 |
+
to the paper about K-MHaS, published at COLING 2022. This dataset is based on the
|
| 8 |
+
Korean online news comments available on Kaggle and Github. The unlabeled raw data
|
| 9 |
+
was collected between January 2018 and June 2020. The language producers are users
|
| 10 |
+
who left the comments on the Korean online news platform between 2018 and 2020.'
|
| 11 |
+
evaluation_framework: mteb
|
| 12 |
+
tasks:
|
| 13 |
+
- id: KorHateSpeechMLClassification
|
| 14 |
+
- id: KorHateSpeechMLClassification_default_test
|
| 15 |
+
config: default
|
| 16 |
+
split: test
|