Datasets:
Paul
/

Modalities:
Tabular
Text
Formats:
csv
Languages:
English
ArXiv:
Libraries:
Datasets
pandas
License:
Paul Röttger commited on
Commit
65cd5e7
1 Parent(s): 2fefa23

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +51 -68
README.md CHANGED
@@ -24,94 +24,77 @@ task_ids:
24
 
25
  ## Dataset Description
26
 
27
- - **Paper:**
28
- - **Repository:**
29
- - **Point of Contact:**
30
 
31
- ### Dataset Summary
32
 
33
- [More Information Needed]
 
 
34
 
35
- ### Supported Tasks and Leaderboards
36
-
37
- [More Information Needed]
38
-
39
- ### Languages
40
-
41
- [More Information Needed]
42
 
43
  ## Dataset Structure
44
 
45
- ### Data Instances
46
-
47
- [More Information Needed]
48
-
49
- ### Data Fields
50
-
51
- [More Information Needed]
52
-
53
- ### Data Splits
54
-
55
- [More Information Needed]
56
-
57
- ## Dataset Creation
58
-
59
- ### Curation Rationale
60
-
61
- [More Information Needed]
62
-
63
- ### Source Data
64
 
65
- #### Initial Data Collection and Normalization
 
66
 
67
- [More Information Needed]
 
68
 
69
- #### Who are the source language producers?
 
70
 
71
- [More Information Needed]
 
72
 
73
- ### Annotations
 
74
 
75
- #### Annotation process
 
76
 
77
- [More Information Needed]
 
78
 
79
- #### Who are the annotators?
 
80
 
81
- [More Information Needed]
 
 
82
 
83
- ### Personal and Sensitive Information
 
84
 
85
- [More Information Needed]
 
86
 
87
- ## Considerations for Using the Data
88
-
89
- ### Social Impact of Dataset
90
-
91
- [More Information Needed]
92
-
93
- ### Discussion of Biases
94
-
95
- [More Information Needed]
96
-
97
- ### Other Known Limitations
98
-
99
- [More Information Needed]
100
-
101
- ## Additional Information
102
-
103
- ### Dataset Curators
104
-
105
- [More Information Needed]
106
-
107
- ### Licensing Information
108
-
109
- [More Information Needed]
110
 
111
  ### Citation Information
112
 
113
- [More Information Needed]
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
114
 
115
- ### Contributions
116
 
117
- Thanks to [@github-username](https://github.com/<github-username>) for adding this dataset.
 
24
 
25
  ## Dataset Description
26
 
27
+ HateCheck is a suite of functional test for hate speech detection models.
28
+ 3
 
29
 
30
+ t contains ,728 test cases in 29 functional tests
31
 
32
+ - **Paper:** Röttger et al. (2021) - HateCheck: Functional Tests for Hate Speech Detection Model. https://aclanthology.org/2021.acl-long.4/
33
+ - **Repository:** https://github.com/paul-rottger/hatecheck-data
34
+ - **Point of Contact:** [email protected]
35
 
 
 
 
 
 
 
 
36
 
37
  ## Dataset Structure
38
 
39
+ "test.csv" contains all 3,728 validated test cases. Each test case (row) has the following attributes:
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
40
 
41
+ **functionality**
42
+ The shorthand for the functionality tested by the test case.
43
 
44
+ **case_id**
45
+ The unique ID of the test case (assigned to each of the 3,901 cases we initially generated)
46
 
47
+ **test_case**
48
+ The text of the test case.
49
 
50
+ **label_gold**
51
+ The gold standard label (hateful/non-hateful) of the test case. All test cases within a given functionality have the same gold standard label.
52
 
53
+ **target_ident**
54
+ Where applicable, the protected group targeted or referenced by the test case. We cover seven protected groups in the test suite: women, trans people, gay people, black people, disabled people, Muslims and immigrants.
55
 
56
+ **direction**
57
+ For hateful cases, the binary secondary label indicating whether they are *directed* at an individual as part of a protected group or aimed at the group in *general*.
58
 
59
+ **focus_words**
60
+ Where applicable, the key word or phrase in a given test case (e.g. "cut their throats").
61
 
62
+ **focus_lemma**
63
+ Where applicable, the corresponding lemma (e.g. "cut sb. throat").
64
 
65
+ **ref_case_id**
66
+ For hateful cases, where applicable, the ID of the simpler hateful case which was perturbed to generate them.
67
+ For non-hateful cases, where applicable, the ID of the hateful case which is contrasted.
68
 
69
+ **ref_templ_id**
70
+ The equivalent, but for template IDs.
71
 
72
+ **templ_id**
73
+ The unique ID of the template from which the test case was generated (assigned to each of the 866 cases and templates from which we generated the 3,901 initial cases).
74
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
75
 
76
  ### Citation Information
77
 
78
+ When using HateCheck, please cite our ACL paper:
79
+
80
+ @inproceedings{rottger-etal-2021-hatecheck,
81
+ title = "{H}ate{C}heck: Functional Tests for Hate Speech Detection Models",
82
+ author = {R{\"o}ttger, Paul and
83
+ Vidgen, Bertie and
84
+ Nguyen, Dong and
85
+ Waseem, Zeerak and
86
+ Margetts, Helen and
87
+ Pierrehumbert, Janet},
88
+ booktitle = "Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers)",
89
+ month = aug,
90
+ year = "2021",
91
+ address = "Online",
92
+ publisher = "Association for Computational Linguistics",
93
+ url = "https://aclanthology.org/2021.acl-long.4",
94
+ doi = "10.18653/v1/2021.acl-long.4",
95
+ pages = "41--58",
96
+ abstract = "Detecting online hate is a difficult task that even state-of-the-art models struggle with. Typically, hate speech detection models are evaluated by measuring their performance on held-out test data using metrics such as accuracy and F1 score. However, this approach makes it difficult to identify specific model weak points. It also risks overestimating generalisable model performance due to increasingly well-evidenced systematic gaps and biases in hate speech datasets. To enable more targeted diagnostic insights, we introduce HateCheck, a suite of functional tests for hate speech detection models. We specify 29 model functionalities motivated by a review of previous research and a series of interviews with civil society stakeholders. We craft test cases for each functionality and validate their quality through a structured annotation process. To illustrate HateCheck{'}s utility, we test near-state-of-the-art transformer models as well as two popular commercial models, revealing critical model weaknesses.",
97
+ }
98
+
99
 
 
100