Datasets:

strombergnlp
/

shaj

Sub-tasks: hate-speech-detection

Multilinguality: monolingual

Size Categories: 10K<n<100K

Language Creators: found

Annotations Creators: expert-generated

Source Datasets: original

ArXiv:

Tags:

DOI:

Dataset card Viewer Files Files and versions Community

Dataset Viewer

Auto-converted to Parquet

View in Dataset Viewer

Viewer

Split (1)

train · 11.9k rows

id stringlengths 1 5	text stringlengths 1 1.38k	subtask_a class label 2 classes	subtask_b class label 3 classes	subtask_c class label 4 classes
0	Lule e eger e rrazikshme,ato frutat kan helm.	1NOT	2no label	3no label
1	Kungull eshte	1NOT	2no label	3no label
2	Kungull	1NOT	2no label	3no label
3	Iriq Bio	1NOT	2no label	3no label
4	Eshte kungull	1NOT	2no label	3no label
5	Ed rama me mjekerr😉	1NOT	2no label	3no label
6	kajota	1NOT	2no label	3no label
7	PYETE SPIROPALIN se ajo e di	1NOT	2no label	3no label
8	Patate	1NOT	2no label	3no label
9	Dardhe Me Stil Iriqi😂😂	1NOT	2no label	3no label
10	Kungull eshte	1NOT	2no label	3no label
11	Dhe un kam ne shpi dhe e bej burek sepse eshte shume i mire per cdo loje semundje @joqalbanian	1NOT	2no label	3no label
12	Patate kineze	1NOT	2no label	3no label
13	eshte kungull	1NOT	2no label	3no label
14	Quhet patate	1NOT	2no label	3no label
15	Kungull	1NOT	2no label	3no label
16	Kungull eshte	1NOT	2no label	3no label
17	Patate arabe	1NOT	2no label	3no label
18	patate e embek ka shije si kungulli i vogel	1NOT	2no label	3no label
19	😘	1NOT	2no label	3no label
20	Kungull	1NOT	2no label	3no label
21	dardhë me gjëmba	1NOT	2no label	3no label
22	Dardh po I ka vdek berberi	1NOT	2no label	3no label
23	Dardh me mbiemer kivi😂😂😂	1NOT	2no label	3no label
24	Patate	1NOT	2no label	3no label
25	Patate me ferra asht se kam edhe un nshpi	1NOT	2no label	3no label
26	Eshte patate e embel	1NOT	2no label	3no label
27	Se nuk e di	1NOT	2no label	3no label
28	Nuk asht frut ai	1NOT	2no label	3no label
29	Kajota	1NOT	2no label	3no label
30	Kungull o popull kungull	0OFF	0TIN	1GRP
31	Patate e embel quhet	1NOT	2no label	3no label
32	@taku_begaj ca esht ky vlla 😅😅😅 mua m duket kungull grek 😅😅	0OFF	0TIN	1GRP
33	Patate po e llojt qe rritet si kivi ose kungulli varet neper mure osht shm e mir	1NOT	2no label	3no label
34	Qershi	1NOT	2no label	3no label
35	Kajota eshte perime ka formen e nje dardhe dhe esht si patate e embel	1NOT	2no label	3no label
36	Kungull spanjoll, gatuhet tave me patate ,,etj normal.	1NOT	2no label	3no label
37	Dardha gjirokastrite	1NOT	2no label	3no label
38	kungull me gjemba ndryshe quhen kajota	1NOT	2no label	3no label
39	Kungull	1NOT	2no label	3no label
40	Patate indiane👌	1NOT	2no label	3no label
41	Kajot	1NOT	2no label	3no label
42	Qershi	1NOT	2no label	3no label
43	patate eshte	1NOT	2no label	3no label
44	Ky. Fruti. Eshte😎 kajota	1NOT	2no label	3no label
45	Dardh e pa ruarr	1NOT	2no label	3no label
46	🤣🤣🤣🤣🤣	1NOT	2no label	3no label
47	Tope me lesh.	0OFF	1UNT	3no label
48	O raka esht avokado	0OFF	0TIN	1GRP
49	Kungull	1NOT	2no label	3no label
50	kungull, shumë i shijshëm në gatim	1NOT	2no label	3no label
51	Kjo quhet rilindje. Ose koqe rame me shkurt.	0OFF	0TIN	1GRP
52	Dardh gjirokastre,qe vetem ta shohesh e mos ta hash kurr	1NOT	2no label	3no label
53	Esht patate vec se ka me shum vitamina dhe shije se patatet origjinale	1NOT	2no label	3no label
54	per sa koh eshte rama ne pushtet keto cudira do ndodhin ... prite kur me dal ndonje domatme me baluke	1NOT	2no label	3no label
55	Kungull	1NOT	2no label	3no label
56	Patate🥔🥔🥔	1NOT	2no label	3no label
57	Dardh e eger hahaah😂🤣😂	1NOT	2no label	3no label
58	Patate eger mo vlla se kam mbjell te kopshti sa te dush😂😂	1NOT	2no label	3no label
59	kungull	1NOT	2no label	3no label
60	Kungull e shartuar me kivi,. Kungullkivi RAMA IK 😜😜😂😂😂@igli_bala 😂 😂 😂 😂	1NOT	2no label	3no label
61	Kiwi austrilian me origjine greke😉	1NOT	2no label	3no label
62	😂 ka tru tjetri	1NOT	2no label	3no label
63	Quhet Kajote. Bëjë sallatë me vaj e uthull se shkon mirë me kte frut. 😂	1NOT	2no label	3no label
64	Hahet un e kam henger ne turqi	1NOT	2no label	3no label
65	Jan dardha po nuk ja depilu se asht ftoht😂😂😂	1NOT	2no label	3no label
66	Kjo quhet Patate Amerikane, edhe vetem mund te zihet, nuk behet e skuqur	1NOT	2no label	3no label
67	@ini_hysa1 oo gjigand 😂🇦🇱	1NOT	2no label	3no label
68	Patate e embel	1NOT	2no label	3no label
69	dardh adoleshent ka ngritur floket me xhel	1NOT	2no label	3no label
70	Kjo quhet patate deti.Qerohet e gjithe lekura e jashtme megjithe gjembat dhe skuqet njesoj si patatja e zakonshme,ne shije vjen pak me e embel se patatja normale	1NOT	2no label	3no label
71	Fik po nuk o fik 😂😂	1NOT	2no label	3no label
72	Koqe pylli	0OFF	1UNT	3no label
73	Edhte patate	1NOT	2no label	3no label
74	Kungulli hirushes😂😂	1NOT	2no label	3no label
75	Kungull .	1NOT	2no label	3no label
76	Kungull o kungull	0OFF	0TIN	0IND
77	Behet sallat behet byrek me qumesht	1NOT	2no label	3no label
78	Ftoj 😂😂😂	1NOT	2no label	3no label
79	😍😍😍😍👏	1NOT	2no label	3no label
80	Kungull greke behet e skuqur ne feta me pak hudhra dhe kose ose i mbushur me mish te grire e karota e djathe te grire kackavall eshte shume i mire	1NOT	2no label	3no label
81	Hyn te lloj i kungullit esht shum i shijshem ne gatim dhe ka shum vlera ushqimore	1NOT	2no label	3no label
82	Kungull eshte	1NOT	2no label	3no label
83	Patate indjane	1NOT	2no label	3no label
84	Kungull kam un ne shpi te till	1NOT	2no label	3no label
85	Patate franceze	1NOT	2no label	3no label
86	Jan te ngrenshem madje me 400 lek kilja	1NOT	2no label	3no label
87	Kungull me gjemba	1NOT	2no label	3no label
88	Patate deti	1NOT	2no label	3no label
89	Kungull i eger	1NOT	2no label	3no label
90	Dardhe e sterilizuar	1NOT	2no label	3no label
91	Shkule dhe digje se a kungull serb	0OFF	0TIN	1GRP
92	Kajote quhet italisht🔥	1NOT	2no label	3no label
93	Dardh me gjemba 👌🏻	1NOT	2no label	3no label
94	Patate e eger	1NOT	2no label	3no label
95	Kungull i eger.Sa shume jane shtuar 😁😁😁qe kur drejtimin e mori Lali fiku	1NOT	2no label	3no label
96	Kur i ben te skuqur	1NOT	2no label	3no label
97	Kungull jan dhe jan shum te mir	1NOT	2no label	3no label
98	Kungull o kunguj	0OFF	0TIN	1GRP
99	tope dashi	0OFF	1UNT	3no label

End of preview. Expand in Dataset Viewer.

YAML Metadata Error: "licenses" is deprecated. Use "license" instead.

YAML Metadata Error: "languages" is deprecated. Use "language" instead.

YAML Metadata Warning: The task_categories "text_classification" is not in the official list: text-classification, token-classification, table-question-answering, question-answering, zero-shot-classification, translation, summarization, feature-extraction, text-generation, text2text-generation, fill-mask, sentence-similarity, text-to-speech, text-to-audio, automatic-speech-recognition, audio-to-audio, audio-classification, voice-activity-detection, depth-estimation, image-classification, object-detection, image-segmentation, text-to-image, image-to-text, image-to-image, image-to-video, unconditional-image-generation, video-classification, reinforcement-learning, robotics, tabular-classification, tabular-regression, tabular-to-text, table-to-text, multiple-choice, text-retrieval, time-series-forecasting, text-to-video, image-text-to-text, visual-question-answering, document-question-answering, zero-shot-image-classification, graph-ml, mask-generation, zero-shot-object-detection, text-to-3d, image-to-3d, image-feature-extraction, other

YAML Metadata Warning: The task_ids "text-classification-other-hate-speech-detection" is not in the official list: acceptability-classification, entity-linking-classification, fact-checking, intent-classification, language-identification, multi-class-classification, multi-label-classification, multi-input-text-classification, natural-language-inference, semantic-similarity-classification, sentiment-classification, topic-classification, semantic-similarity-scoring, sentiment-scoring, sentiment-analysis, hate-speech-detection, text-scoring, named-entity-recognition, part-of-speech, parsing, lemmatization, word-sense-disambiguation, coreference-resolution, extractive-qa, open-domain-qa, closed-domain-qa, news-articles-summarization, news-articles-headline-generation, dialogue-modeling, dialogue-generation, conversational, language-modeling, text-simplification, explanation-generation, abstractive-qa, open-domain-abstractive-qa, closed-domain-qa, open-book-qa, closed-book-qa, slot-filling, masked-language-modeling, keyword-spotting, speaker-identification, audio-intent-classification, audio-emotion-recognition, audio-language-identification, multi-label-image-classification, multi-class-image-classification, face-detection, vehicle-detection, instance-segmentation, semantic-segmentation, panoptic-segmentation, image-captioning, image-inpainting, image-colorization, super-resolution, grasping, task-planning, tabular-multi-class-classification, tabular-multi-label-classification, tabular-single-column-regression, rdf-to-text, multiple-choice-qa, multiple-choice-coreference-resolution, document-retrieval, utterance-retrieval, entity-linking-retrieval, fact-checking-retrieval, univariate-time-series-forecasting, multivariate-time-series-forecasting, visual-question-answering, document-question-answering

Dataset Card for "shaj"

Dataset Summary

This is an abusive/offensive language detection dataset for Albanian. The data is formatted following the OffensEval convention, with three tasks:

Subtask A: Offensive (OFF) or not (NOT)
Subtask B: Untargeted (UNT) or targeted insult (TIN)
Subtask C: Type of target: individual (IND), group (GRP), or other (OTH)

Notes on the above:

The subtask A field should always be filled.
The subtask B field should only be filled if there's "offensive" (OFF) in A.
The subtask C field should only be filled if there's "targeted" (TIN) in B.

The dataset name is a backronym, also standing for "Spoken Hate in the Albanian Jargon"

See the paper https://arxiv.org/abs/2107.13592 for full details.

Supported Tasks and Leaderboards

Task A leaderboard at paperswithcode.com/sota/hate-speech-detection-on-shaj

Languages

Albanian (bcp47:sq-AL)

Dataset Structure

Data Instances

shaj

Size of downloaded dataset files: 769.21 KiB
Size of the generated dataset: 1.06 MiB
Total amount of disk used: 1.85 MiB

An example of 'train' looks as follows.

{
  'id': '0', 
  'text': 'PLACEHOLDER TEXT', 
  'subtask_a': 1, 
  'subtask_b': 0, 
  'subtask_c': 0
}

Data Fields

id: a string feature.
text: a string.
subtask_a: whether or not the instance is offensive; 0: OFF, 1: NOT
subtask_b: whether an offensive instance is a targeted insult; 0: TIN, 1: UNT, 2: not applicable
subtask_c: what a targeted insult is aimed at; 0: IND, 1: GRP, 2: OTH, 3: not applicable

Data Splits

name	train
shaj	11874 sentences

Dataset Creation

Curation Rationale

Collecting data for enabling offensive speech detection in Albanian

Source Data

Initial Data Collection and Normalization

The text is scraped from comments on popular Albanian YouTube and Instagram accounts. An extended discussion is given in the paper in section 3.2.

Who are the source language producers?

People who comment on a selection of high-activity Albanian instagram and youtube profiles.

Annotations

Annotation process

The annotation scheme was taken from OffensEval 2019 and applied by two native speaker authors of the paper as well as their friends and family.

Who are the annotators?

Albanian native speakers, male and female, aged 20-60.

Personal and Sensitive Information

The data was public at the time of collection. No PII removal has been performed.

Considerations for Using the Data

Social Impact of Dataset

The data definitely contains abusive language.

Discussion of Biases

Other Known Limitations

Additional Information

Dataset Curators

The dataset is curated by the paper's authors.

Licensing Information

The authors distribute this data under Creative Commons attribution license, CC-BY 4.0.

Citation Information

@article{nurce2021detecting,
  title={Detecting Abusive Albanian},
  author={Nurce, Erida and Keci, Jorgel and Derczynski, Leon},
  journal={arXiv preprint arXiv:2107.13592},
  year={2021}
}

Contributions

Author-added dataset @leondz

Downloads last month: 2

Edit dataset card Papers with Code

Size of downloaded dataset files:

788 kB

Size of the auto-converted Parquet files:

612 kB

Number of rows:

11,874