Datasets:

Languages:
English
Multilinguality:
monolingual
Size Categories:
1K<n<10K
Source Datasets:
original
Tags:
License:
Tweet ID
int64
1,552,636,108B
1,561,385,783B
Label
int64
0
1
1,556,394,598,150,422,500
0
1,556,348,956,787,019,800
1
1,556,369,996,326,084,600
0
1,556,346,056,388,010,000
1
1,556,388,136,065,740,800
0
1,556,345,497,404,579,800
1
1,556,411,641,939,370,000
0
1,556,344,123,073,986,600
1
1,556,410,426,367,385,600
0
1,556,342,567,410,643,000
1
1,556,341,903,330,885,600
0
1,556,341,290,241,065,000
1
1,556,368,172,952,854,500
0
1,556,338,715,542,184,000
1
1,556,413,096,662,405,000
0
1,556,336,938,008,789,000
1
1,556,404,636,369,621,000
0
1,556,335,954,842,886,100
1
1,556,383,890,130,014,200
0
1,556,334,597,478,224,000
1
1,556,413,332,877,459,500
0
1,556,331,405,428,363,300
1
1,556,393,724,212,646,000
0
1,556,327,194,636,566,500
1
1,556,351,636,955,709,400
0
1,556,324,226,747,048,000
1
1,556,392,003,981,545,500
0
1,556,321,569,474,941,000
1
1,556,405,022,484,840,400
0
1,556,319,481,126,944,800
1
1,556,391,334,977,626,000
0
1,556,315,945,353,764,900
1
1,556,361,950,870,864,000
0
1,556,299,876,589,240,300
1
1,556,403,556,189,552,600
0
1,556,294,469,493,723,100
1
1,556,409,789,462,233,000
0
1,556,290,988,645,224,400
1
1,556,401,226,270,576,600
0
1,556,286,402,593,280,000
1
1,556,379,716,973,727,700
0
1,556,284,129,196,421,000
1
1,556,401,820,389,544,000
0
1,556,282,992,342,990,800
1
1,556,374,520,608,329,700
0
1,556,281,557,878,337,500
1
1,556,317,772,799,410,200
0
1,556,278,437,052,022,800
1
1,556,352,469,881,471,000
0
1,556,275,882,662,441,000
1
1,556,394,837,682,929,700
0
1,556,266,428,751,306,800
1
1,556,260,888,717,541,400
0
1,556,259,636,273,111,000
1
1,556,413,243,404,365,800
0
1,556,256,753,305,628,700
1
1,556,320,179,436,470,300
0
1,556,252,909,444,014,000
1
1,556,402,563,565,731,800
0
1,556,248,165,732,520,000
1
1,556,390,120,953,086,000
0
1,556,248,015,945,535,500
1
1,556,399,614,256,898,000
0
1,556,241,410,545,836,000
1
1,556,236,662,149,111,800
0
1,556,233,449,744,334,800
1
1,556,364,224,959,582,200
0
1,556,228,196,428,832,800
1
1,556,295,116,628,938,800
0
1,556,226,866,637,410,300
1
1,556,393,324,407,398,400
0
1,556,209,188,614,934,500
1
1,556,384,819,092,738,000
0
1,556,202,173,792,387,000
1
1,556,387,682,787,393,500
0
1,556,194,325,637,353,500
1
1,556,334,910,348,005,400
0
1,556,173,748,302,381,000
1
1,556,313,815,129,677,800
0
1,556,169,994,634,944,500
1
1,556,408,821,018,497,000
0
1,556,141,126,419,112,000
1
1,556,371,111,218,233,300
0
1,556,140,016,417,575,000
1
1,556,193,929,011,564,500
0
1,556,130,707,231,350,800
1
1,556,374,748,333,936,600
0
1,556,125,707,851,292,700
1
1,556,346,095,936,102,400
0
1,556,120,395,987,705,900
1
1,556,377,012,754,960,400
0
1,556,119,568,027,885,600
1
1,556,351,591,770,427,400
0
1,556,119,015,801,634,800
1
1,556,114,452,491,247,600
0
1,556,113,660,078,162,000
1
1,556,370,801,145,745,400
0
1,556,103,574,592,991,200
1
1,556,402,250,733,342,700
0
1,556,102,966,758,613,000
1

Utilising Weak Supervision to Create S3D: A Sarcasm Annotated Dataset

This is the repository for the S3D dataset published at EMNLP 2022. The dataset can help build sarcasm detection models.

SAD

The SAD dataset is our gold standard dataset of tweets labelled for sarcasm. These tweets were scraped by observing a '#sarcasm' hashtag and then manually annotated by three annotators. There are a total of 1170 pairs of a sarcastic and non-sarcastic tweets which were both posted by the same user, resulting in a total of 2340 tweets annotated for sarcasm. These tweets can be accessed by using the Twitter API so that they can be used for other experiments.

Data Fields

  • Tweet ID: The ID of the labelled tweet
  • Label: A label to denote if a given tweet is sarcastic

Data Splits

  • Train: 1638
  • Valid: 351
  • Test: 351
Downloads last month
2
Edit dataset card