text
stringlengths
1
305
label
sequence
"자한당틀딱들.. 악플질 고만해라."
[ 2, 4 ]
정치적으로 편향된 평론한은 분은 별로...
[ 8 ]
적당히좀 쳐먹지.그랬냐??? 안그래도 문재인 때문에 나라 엉망진창인데...
[ 2 ]
"안서는 아재들 풀발기 ㅋㄲㅋ"
[ 4 ]
우와 ㅋ 능력자
[ 8 ]
맛녀석 콩트보다 약했음맛녀석 애청자로써 70%실력발휘
[ 8 ]
주영훈 솔직히 호감임 잉꼬부부로 소문났잖아
[ 8 ]
이게주간아이돌이랑머가달라...
[ 8 ]
아오 슈박 회사생활도 졑깥고 돈벌기 힘들어 죽겠구만 뭔 저딴것들 자꾸 tv나와서 사람 짜증나게하냐 외국서 편히살려면 아닥하고 살아라 대한민국서 취미로 돈벌어가지말고 좀 끄지라고!
[ 3 ]
"문재인 하는게 뭐 별거있냐?ㅂㅅㅅㅋ가 하는짓인데 어련하겠어.ㅋㅋㅋ"
[ 2, 3 ]
공통점은 한국남자?
[ 8 ]
이재진은 그게문제가아니자나 지금 ㅋㅋㅋㅋ
[ 8 ]
"야이 OOO 색햐. 주체사상 황장엽 국립묘지에 십어넣은 이명박은 어쩔래 드으응OOO아."
[ 3 ]
소지섭 걸어다니는 인간 수면제나피디가 편집을 잘하겠지만걱정된다.ㅠ.ㅠ
[ 8 ]
원중씨........... 잘가요........ 잘살아요..........
[ 8 ]
50어 애나서 언제 길러요?
[ 8 ]
정신나간년들 나라 돌아가는건 1도 모르지
[ 3 ]
"중국은 무슨 암덩어리 같음 어떻게 저렇게 세계각지에 골고루 민폐만 끼치냐"
[ 0 ]
알바풀었나 많이본뉴스 이프로그램기사 댓글하고 반응이 180도 다르네 ㅋㅋㅋ 의미없는 찬양질 왜케많냐 여긴
[ 8 ]
택이 아부지 카리스마 뿜뿜
[ 8 ]
기사에 낚이는 ㅂㅅ같은 댓글러들
[ 3 ]
김태리.. 나 3대째 돼지농장주 고석환의 신부후보 16위.. 하지만 오늘 9위로 올라갔다. 좀더 분발해라
[ 8 ]
지가 괜찮다는데 니들은 좀 닥₩/₩치고있어라 대리도 못단 새!.!3!;끼들이 말은 존/&나게 많네 ㅋㅋ
[ 3 ]
대체 모자란게 뭐지 완벽하네
[ 8 ]
다들 오지랖 좀...되게 할 짓 없어보여요
[ 8 ]
"찌질한 새끼들.... 왜? 동양 남자는 무섭냐?"
[ 3 ]
개나소나 트로트 가수래 노래 존나 못부르네
[ 3 ]
절세미녀라 좋아했었는데 통수친거보고 정뚝떨. 여우짓 ㅈㄴ싫음ㅋㅋㅋㅋㅋ 무튼 행쇼~
[ 3 ]
지인이란넘?지인 맞어 ㅋㅋ
[ 8 ]
몬생긴 여자인간들은 찌그러져있어라 확그냥 막그냥
[ 1, 5 ]
송중기는 사진에도 없구만..ㅋㅋ송혜교♡고양이.ㅋㅋㅋ
[ 8 ]
도대체 ㅋㅋㅋㅋ어디서 방송사 갑질이 튀어나온거?? 지금까지 밝혀진 사실은 전부 고현정 갑질이던데???ㅋㅋㅋㅋㅋㅋ촬영거부, 감독지시거부, 촬영이탈, 밀치고욕하는폭력행사 ㅋㅋㅋㅋㅋㅋㅋㅋㅋ 왜갑자기 고현정을 불의에 맞서싸누는 잔다르크행인지 ㅋㅋㅋㅋㅋ여자들 진짜 무논리
[ 5 ]
유부남이 처녀에게 너랑 결혼하고 싶다라고 했으면 개욕처먹었겠지 ㅋㅋㅋㅋ 미친헬조선에서 태어나 오늘도 감사합니다
[ 0 ]
연기못하는 방구쟁이년
[ 3 ]
"짱깨 왤케 꼴보기 싫냐. 중국내에서 반란 일어나서 뿔뿔이 뜯어졌음 좋겠다."
[ 0 ]
근데 더러운건 지디인데 왜 여자욕만함 ㅋㅋ
[ 8 ]
"쿵쾅이들 좌표찍고 화력 오지네 ㅋㅋ"
[ 1, 5 ]
"하나님은 변함이 없는데 개목사들이 하나님을 팔아서 지배만 불리지"
[ 0 ]
가좇영화를 킹스맨 따위를 비비냐 ㅋㅋㅋㅋ
[ 3 ]
"짱깨듣이여 짱깨는 대한민국에 전혀 도움이 안되는 인간들이야 추방이 답이다."
[ 0 ]
"목사 좋아하네~~ 종교팔이 하지마라 ! 개목사 주제에.."
[ 7 ]
"이여인 관상자체가.."
[ 8 ]
최종몸통이 문죄인이랑 관련있는거 같다살아있믄 권력이니 안잡고 쉬쉬하는거지
[ 2 ]
조용필 이선희씨 거기 가서 노래부르는 짓 하지 마세요 그냥 거절하시길
[ 8 ]
와..진짜 예쁘다..
[ 8 ]
가시나세상좋아했는데. . 난선미넘좋아~.
[ 8 ]
"교활한 개독의 전형. 독실한 개독 하나 더 쥐바기"
[ 2, 7 ]
고생 하셨습니다. 좀더 맘편한 삶을 행복한삶을 살수 있길 바랍니다.
[ 8 ]
남은 인생달달하게 행복하게 사세요축하합니다
[ 8 ]
지우라면 지워 말 존나게많네 싸OOO들
[ 3 ]
한심한새끼들ㅉ
[ 3 ]
나도 돈많은 아줌마랑 결혼하고싶다
[ 8 ]
뚱보끼린 뭔가 잘맞는게 있나보네
[ 1 ]
광대가 제일 아쉽 ㅠㅠㅠㅠ
[ 8 ]
"관상은 무시 못한다.. 못되게 생긴거 봐라.."
[ 1 ]
옛날 보리고개 시절에도 7남매는 흔할 정도로 다산했었고 가난했어도 애들 다 잘키웠었다, 그리고 그 시절엔 산모가 밭에서 일하다가 애낳는 경우도 흔했고 산후조리는 꿈도 못꾸고 애낳고 이틀정도만 쉬고 다시 일하러 갔었다. 이런 헝그리 정신이 오늘날의 대한민국을 만들었는데 요즘 젊은 것들은 그 때에 비하면 천국같이 살기 좋은 세상인데도 헬조선 운운하며 애들을 안 낳는데 한 마디로 배지때기가 처 불러서 정신을 못차리는거다
[ 4 ]
"보름이 원래 웃음상이라 오해받는건데 40대 새끼들 진짜 토나온다"
[ 3, 4 ]
외모만보믄한국토종시골아줌마죠좀못생긴줌마. 노랜잘하죠
[ 1, 5 ]
노래 몇번 망하더니 이제 헝그리 정신이 새로 생긴듯....아아 세월이 야속하네. 옛날에 전성기 시절엔 지금 수지 아이유 합쳐놓은 급이었는데.
[ 8 ]
여기 아줌마들 시샘이 장난아니네요 열폭할시간에 애기나보세요^^
[ 5 ]
어린 여자 만나야 하는건가..
[ 8 ]
와!! 못생겨따
[ 1 ]
문재앙 덮으려고 연예인들이 고생하네
[ 2 ]
남자도 성형이 많다니..성형공화국답다
[ 8 ]
누나 함부로 쓰지마라. 곧 환갑인데. 아줌마나 할머니다.
[ 8 ]
맏딸이해가안가고 가장이기적인뇬 아버지가황혼에서 첫사랑만나 행복하고싶다는데 훼방놓다니 무자식이상팔자다
[ 3 ]
합창 물타기 꼼수가 방송을 핵노잼 만듦
[ 8 ]
영자편만보고 잘려고 했더니 뒤로 밀어놨네 아놔~유병재꺼 보기싫었는데.
[ 8 ]
"어서들 피검사 받으시길~~"
[ 8 ]
아우~~ 진짜 볼깨물어주게 귀엽넹~~ 삼둥이 아프지말고 건강하게 자라길 바랄게~~
[ 8 ]
누구신지. . .존재감 없는
[ 8 ]
여자가 남자궁뎅이만지면 결론은 그럴수도있지 남자가 쪼잔하게 따지냐 여자가 만져주면 좋지뭐 어때?? 반면 여자궁뎅이 만지면 30년전일이라도 지금 기억나면 아 요즘 짜증나는데누구 미투할남자없나 찾아서 아싸 미투 니인생 조져 ^-----^
[ 5 ]
와 이거 완전 재앙이네. 깨끗한 나라 만든다면서 왜 이런 사건이 터지냐??
[ 2 ]
좀더 찍어봐요
[ 8 ]
암만 연예인이라도 남잔데 무슨 화장을 저리도 떡칠을 했냐? 무슨 게이샤냐?
[ 5 ]
어이구 돌문재앙
[ 2 ]
오늘 첫 끼여~
[ 8 ]
다들 존나 꼬였구나
[ 3 ]
"중국중국중국중국하고 전쟁이나 시원하게 하자"
[ 8 ]
운동이 아니라 혹사지....걍 밥잘먹고 잘자고 조깅만해도 된다.
[ 8 ]
토크몬 강호동 김희선땜시 본다^^
[ 8 ]
정말바르고 봐도괘안은것같은데 두고봐야지 지금은 반듯한거같은데 그래서이뽀
[ 8 ]
"엉덩이를 흔들어봐~"
[ 8 ]
승리고향 전라도광주ㅋㅋㅋㅋ더 이상 무슨말이필요?ㅋㅇㅋ
[ 8 ]
"생긴것도 개떡같은게 개떡치는 소리하네. 18뇬!"
[ 1, 3 ]
"니네 꼴보기 시러 나도 동참했다 정말 느그 조국 쪽바리한테 가라"
[ 0 ]
"무식한 할배들의 안보팔이 식상타. 이젠 젊은이들이 살 세상이다"
[ 4 ]
애초에 카지노를 왜 짓냐
[ 8 ]
"주사파, 사노맹, 대깨문, 달X, 호남인들이 발광하네"
[ 0, 2 ]
"섬나라로 가라..."
[ 8 ]
"이뇸이 똥오즘 못가리네"
[ 3 ]
화장안한얼굴이훨씬이쁜것같음.실제로보니 몸매도 날씬한것같드만.카메라지대안받는타입인듯ㅎㅎ기자들이안티 ㅎ요즘 눈에뛰는여자아나운서별로없던데 쫌기대됨
[ 8 ]
"수꼴틀딱시키들이 다 디져야 나라가 똑바로 될것같다..답이 없는 종자들ㅠ"
[ 2, 3, 4 ]
"이런기사는 항상 댓글통제 씹재앙 ㅋㅋㅋ 종북새끼"
[ 2, 3 ]
기집년들아 솔직히 얘기하자. 니들이 받고 싶은게 사과냐? 아님 돈이냐? 왜 돈달라는 얘기는 떳떳하지 못한거 같아서 못하는거야? 그렇게 사과를해도 기자찾아다니며 기사써달라고 하는건 결국은 니들 돈먹고 싶어요~ 하는 꼬라지아니냐~가식 쳐떠는 년들 극혐✔
[ 3, 5 ]
이년 이거 정신 못차렸네 이런년들은 연예계에서 퇴출안시키나?
[ 3, 5 ]
결국엔 애 미 창 년이란거잖아
[ 3 ]
"근데 짱깨국에서 원래 마약하면 사형 아니였음?? 시진핑 집권하고 바뀌었나?"
[ 0 ]
"그럼 짱개를 시켜처먹든가 그지새ㅡ키가징징대기는"
[ 0, 3 ]
송지효 남친있지앜ㅎ음? 그 말죽거리 남자 매니저엿던애
[ 8 ]

Dataset Card for K-MHaS

Sample Code

base

Dataset Summary

The Korean Multi-label Hate Speech Dataset, K-MHaS, consists of 109,692 utterances from Korean online news comments, labelled with 8 fine-grained hate speech classes (labels: Politics, Origin, Physical, Age, Gender, Religion, Race, Profanity) or Not Hate Speech class. Each utterance provides from a single to four labels that can handles Korean language patterns effectively. For more details, please refer to our paper about K-MHaS, published at COLING 2022.

Supported Tasks and Leaderboards

Hate Speech Detection

  • binary classification (labels: Hate Speech, Not Hate Speech)
  • multi-label classification: (labels: Politics, Origin, Physical, Age, Gender, Religion, Race, Profanity, Not Hate Speech)

For the multi-label classification, a Hate Speech class from the binary classification, is broken down into eight classes, associated with the hate speech category. In order to reflect the social and historical context, we select the eight hate speech classes. For example, the Politics class is chosen, due to a significant influence on the style of Korean hate speech.

Languages

Korean

Dataset Structure

Data Instances

The dataset is provided with train/validation/test set in the txt format. Each instance is a news comment with a corresponding one or more hate speech classes (labels: Politics, Origin, Physical, Age, Gender, Religion, Race, Profanity) or Not Hate Speech class. The label numbers matching in both English and Korean is in the data fields section.

{'text':'수꼴틀딱시키들이 다 디져야 나라가 똑바로 될것같다..답이 없는 종자들ㅠ'
 'label': [2, 3, 4]
}

Data Fields

  • text: utterance from Korean online news comment.
  • label: the label numbers matching with 8 fine-grained hate speech classes and not hate speech class are follows.
    • 0: Origin(출신차별) hate speech based on place of origin or identity;
    • 1: Physical(외모차별) hate speech based on physical appearance (e.g. body, face) or disability;
    • 2: Politics(정치성향차별) hate speech based on political stance;
    • 3: Profanity(혐오욕설) hate speech in the form of swearing, cursing, cussing, obscene words, or expletives; or an unspecified hate speech category;
    • 4: Age(연령차별) hate speech based on age;
    • 5: Gender(성차별) hate speech based on gender or sexual orientation (e.g. woman, homosexual);
    • 6: Race(인종차별) hate speech based on ethnicity;
    • 7: Religion(종교차별) hate speech based on religion;
    • 8: Not Hate Speech(해당사항없음).

Data Splits

In our repository, we provide splitted datasets that have 78,977(train) / 8,776 (validation) / 21,939 (test) samples, preserving the class proportion.

Dataset Creation

Curation Rationale

We propose K-MHaS, a large size Korean multi-label hate speech detection dataset that represents Korean language patterns effectively. Most datasets in hate speech research are annotated using a single label classification of particular aspects, even though the subjectivity of hate speech cannot be explained with a mutually exclusive annotation scheme. We propose a multi-label hate speech annotation scheme that allows overlapping labels associated with the subjectivity and the intersectionality of hate speech.

Source Data

Initial Data Collection and Normalization

Our dataset is based on the Korean online news comments available on Kaggle and Github. The unlabeled raw data was collected between January 2018 and June 2020. Please see the details in our paper K-MHaS published at COLING2020.

Who are the source language producers?

The language producers are users who left the comments on the Korean online news platform between 2018 and 2020.

Annotations

Annotation process

We begin with the common categories of hate speech found in literature and match the keywords for each category. After the preliminary round, we investigate the results to merge or remove labels in order to provide the most representative subtype labels of hate speech contextual to the cultural background. Our annotation instructions explain a twolayered annotation to (a) distinguish hate and not hate speech, and (b) the categories of hate speech. Annotators are requested to consider given keywords or alternatives of each category within social, cultural, and historical circumstances. For more details, please refer to the paper K-MHaS.

Who are the annotators?

Five native speakers were recruited for manual annotation in both the preliminary and main rounds.

Personal and Sensitive Information

This datasets contains examples of hateful language, however, has no personal information.

Considerations for Using the Data

Social Impact of Dataset

We propose K-MHaS, a new large-sized dataset for Korean hate speech detection with a multi-label annotation scheme. We provided extensive baseline experiment results, presenting the usability of a dataset to detect Korean language patterns in hate speech.

Discussion of Biases

All annotators were recruited from a crowdsourcing platform. They were informed about hate speech before handling the data. Our instructions allowed them to feel free to leave if they were uncomfortable with the content. With respect to the potential risks, we note that the subjectivity of human annotation would impact on the quality of the dataset.

Other Known Limitations

[More Information Needed]

Additional Information

Dataset Curators

This dataset is curated by Taejun Lim, Heejun Lee and Bogeun Jo.

Licensing Information

Creative Commons Attribution-ShareAlike 4.0 International (cc-by-sa-4.0).

Citation Information

@inproceedings{lee-etal-2022-k,
    title = "K-{MH}a{S}: A Multi-label Hate Speech Detection Dataset in {K}orean Online News Comment",
    author = "Lee, Jean  and
      Lim, Taejun  and
      Lee, Heejun  and
      Jo, Bogeun  and
      Kim, Yangsok  and
      Yoon, Heegeun  and
      Han, Soyeon Caren",
    booktitle = "Proceedings of the 29th International Conference on Computational Linguistics",
    month = oct,
    year = "2022",
    address = "Gyeongju, Republic of Korea",
    publisher = "International Committee on Computational Linguistics",
    url = "https://aclanthology.org/2022.coling-1.311",
    pages = "3530--3538",
    abstract = "Online hate speech detection has become an important issue due to the growth of online content, but resources in languages other than English are extremely limited. We introduce K-MHaS, a new multi-label dataset for hate speech detection that effectively handles Korean language patterns. The dataset consists of 109k utterances from news comments and provides a multi-label classification using 1 to 4 labels, and handles subjectivity and intersectionality. We evaluate strong baselines on K-MHaS. KR-BERT with a sub-character tokenizer outperforms others, recognizing decomposed characters in each hate speech class.",
}

Contributions

The contributors of the work are:

  • Jean Lee (The University of Sydney)
  • Taejun Lim (The University of Sydney)
  • Heejun Lee (BigWave AI)
  • Bogeun Jo (BigWave AI)
  • Yangsok Kim (Keimyung University)
  • Heegeun Yoon (National Information Society Agency)
  • Soyeon Caren Han (The University of Western Australia and The University of Sydney)
Downloads last month
673

Models trained or fine-tuned on jeanlee/kmhas_korean_hate_speech