system HF staff commited on
Commit
aac70a7
1 Parent(s): 9999b3f

Update files from the datasets library (from 1.6.0)

Browse files

Release notes: https://github.com/huggingface/datasets/releases/tag/1.6.0

README.md CHANGED
@@ -10,15 +10,13 @@ licenses:
10
  multilinguality:
11
  - monolingual
12
  size_categories:
13
- - 10K<n<100K
14
  source_datasets:
15
  - original
16
  task_categories:
17
  - structure-prediction
18
  task_ids:
19
  - coreference-resolution
20
- - named-entity-recognition
21
- - part-of-speech-tagging
22
  ---
23
 
24
  # Dataset Card for Wino_Bias dataset
@@ -63,8 +61,7 @@ referred by their occupation (e.g. the nurse, the doctor, the carpenter).
63
 
64
  ### Supported Tasks and Leaderboards
65
 
66
- The underlying task is coreference resolution. But it also support NER and POS tasks.
67
-
68
  ### Languages
69
 
70
  English
@@ -92,7 +89,7 @@ English
92
 
93
  ### Data Splits
94
 
95
- Single Split available
96
 
97
  ## Dataset Creation
98
 
@@ -170,4 +167,4 @@ MIT Licence
170
 
171
  ### Contributions
172
 
173
- Thanks to [@akshayb7](https://github.com/akshayb7) for adding this dataset.
10
  multilinguality:
11
  - monolingual
12
  size_categories:
13
+ - 1K<n<10K
14
  source_datasets:
15
  - original
16
  task_categories:
17
  - structure-prediction
18
  task_ids:
19
  - coreference-resolution
 
 
20
  ---
21
 
22
  # Dataset Card for Wino_Bias dataset
61
 
62
  ### Supported Tasks and Leaderboards
63
 
64
+ The underlying task is coreference resolution.
 
65
  ### Languages
66
 
67
  English
89
 
90
  ### Data Splits
91
 
92
+ Dev and Test Split available
93
 
94
  ## Dataset Creation
95
 
167
 
168
  ### Contributions
169
 
170
+ Thanks to [@akshayb7](https://github.com/akshayb7) for adding this dataset. Updated by [@JieyuZhao](https://github.com/JieyuZhao).
dataset_infos.json CHANGED
@@ -1 +1,1315 @@
1
- {"wino_bias": {"description": "WinoBias, a Winograd-schema dataset for coreference resolution focused on gender bias.\nThe corpus contains Winograd-schema style sentences with entities corresponding to people\nreferred by their occupation (e.g. the nurse, the doctor, the carpenter).\n", "citation": "@article{DBLP:journals/corr/abs-1804-06876,\n author = {Jieyu Zhao and\n Tianlu Wang and\n Mark Yatskar and\n Vicente Ordonez and\n Kai{-}Wei Chang},\n title = {Gender Bias in Coreference Resolution: Evaluation and Debiasing Methods},\n journal = {CoRR},\n volume = {abs/1804.06876},\n year = {2018},\n url = {http://arxiv.org/abs/1804.06876},\n archivePrefix = {arXiv},\n eprint = {1804.06876},\n timestamp = {Mon, 13 Aug 2018 16:47:01 +0200},\n biburl = {https://dblp.org/rec/journals/corr/abs-1804-06876.bib},\n bibsource = {dblp computer science bibliography, https://dblp.org}\n}\n", "homepage": "https://uclanlp.github.io/corefBias/overview", "license": "MIT License (https://github.com/uclanlp/corefBias/blob/master/LICENSE)", "features": {"document_id": {"dtype": "string", "id": null, "_type": "Value"}, "part_number": {"dtype": "string", "id": null, "_type": "Value"}, "word_number": {"feature": {"dtype": "int32", "id": null, "_type": "Value"}, "length": -1, "id": null, "_type": "Sequence"}, "tokens": {"feature": {"dtype": "string", "id": null, "_type": "Value"}, "length": -1, "id": null, "_type": "Sequence"}, "pos_tags": {"feature": {"num_classes": 54, "names": ["\"", "''", "#", "$", "(", ")", ",", ".", ":", "``", "CC", "CD", "DT", "EX", "FW", "IN", "JJ", "JJR", "JJS", "LS", "MD", "NN", "NNP", "NNPS", "NNS", "NN|SYM", "PDT", "POS", "PRP", "PRP$", "RB", "RBR", "RBS", "RP", "SYM", "TO", "UH", "VB", "VBD", "VBG", "VBN", "VBP", "VBZ", "WDT", "WP", "WP$", "WRB", "HYPH", "XX", "NFP", "AFX", "ADD", "-LRB-", "-RRB-"], "names_file": null, "id": null, "_type": "ClassLabel"}, "length": -1, "id": null, "_type": "Sequence"}, "parse_bit": {"feature": {"dtype": "string", "id": null, "_type": "Value"}, "length": -1, "id": null, "_type": "Sequence"}, "predicate_lemma": {"feature": {"dtype": "string", "id": null, "_type": "Value"}, "length": -1, "id": null, "_type": "Sequence"}, "predicate_framenet_id": {"feature": {"dtype": "string", "id": null, "_type": "Value"}, "length": -1, "id": null, "_type": "Sequence"}, "word_sense": {"feature": {"dtype": "string", "id": null, "_type": "Value"}, "length": -1, "id": null, "_type": "Sequence"}, "speaker": {"feature": {"dtype": "string", "id": null, "_type": "Value"}, "length": -1, "id": null, "_type": "Sequence"}, "ner_tags": {"feature": {"num_classes": 38, "names": ["B-PERSON", "I-PERSON", "B-NORP", "I-NORP", "B-FAC", "I-FAC", "B-ORG", "I-ORG", "B-GPE", "I-GPE", "B-LOC", "I-LOC", "B-PRODUCT", "I-PRODUCT", "B-EVENT", "I-EVENT", "B-WORK_OF_ART", "I-WORK_OF_ART", "B-LAW", "I-LAW", "B-LANGUAGE", "I-LANGUAGE", "B-DATE", "I-DATE", "B-TIME", "I-TIME", "B-PERCENT", "I-PERCENT", "B-MONEY", "I-MONEY", "B-QUANTITY", "I-QUANTITY", "B-ORDINAL", "I-ORDINAL", "B-CARDINAL", "I-CARDINAL", "*", "0"], "names_file": null, "id": null, "_type": "ClassLabel"}, "length": -1, "id": null, "_type": "Sequence"}, "verbal_predicates": {"feature": {"dtype": "string", "id": null, "_type": "Value"}, "length": -1, "id": null, "_type": "Sequence"}}, "post_processed": null, "supervised_keys": null, "builder_name": "wino_bias", "config_name": "wino_bias", "version": {"version_str": "4.0.0", "description": null, "major": 4, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 173899234, "num_examples": 150335, "dataset_name": "wino_bias"}}, "download_checksums": {"https://drive.google.com/uc?export=download&id=14Im3BnNl-d2fYETYmiH5yq6eFGLVC3g0": {"num_bytes": 268725744, "checksum": "139a6511bcb9761f6306bba7d151bde6ec7ec82cc9c593127b848df8af5f68a1"}}, "download_size": 268725744, "post_processing_size": null, "dataset_size": 173899234, "size_in_bytes": 442624978}}
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "wino_bias": {
3
+ "description": "WinoBias, a Winograd-schema dataset for coreference resolution focused on gender bias.\nThe corpus contains Winograd-schema style sentences with entities corresponding to people\nreferred by their occupation (e.g. the nurse, the doctor, the carpenter).\n",
4
+ "citation": "@article{DBLP:journals/corr/abs-1804-06876,\n author = {Jieyu Zhao and\n Tianlu Wang and\n Mark Yatskar and\n Vicente Ordonez and\n Kai{-}Wei Chang},\n title = {Gender Bias in Coreference Resolution: Evaluation and Debiasing Methods},\n journal = {CoRR},\n volume = {abs/1804.06876},\n year = {2018},\n url = {http://arxiv.org/abs/1804.06876},\n archivePrefix = {arXiv},\n eprint = {1804.06876},\n timestamp = {Mon, 13 Aug 2018 16:47:01 +0200},\n biburl = {https://dblp.org/rec/journals/corr/abs-1804-06876.bib},\n bibsource = {dblp computer science bibliography, https://dblp.org}\n}\n",
5
+ "homepage": "https://uclanlp.github.io/corefBias/overview",
6
+ "license": "MIT License (https://github.com/uclanlp/corefBias/blob/master/LICENSE)",
7
+ "features": {
8
+ "document_id": {
9
+ "dtype": "string",
10
+ "id": null,
11
+ "_type": "Value"
12
+ },
13
+ "part_number": {
14
+ "dtype": "string",
15
+ "id": null,
16
+ "_type": "Value"
17
+ },
18
+ "word_number": {
19
+ "feature": {
20
+ "dtype": "int32",
21
+ "id": null,
22
+ "_type": "Value"
23
+ },
24
+ "length": -1,
25
+ "id": null,
26
+ "_type": "Sequence"
27
+ },
28
+ "tokens": {
29
+ "feature": {
30
+ "dtype": "string",
31
+ "id": null,
32
+ "_type": "Value"
33
+ },
34
+ "length": -1,
35
+ "id": null,
36
+ "_type": "Sequence"
37
+ },
38
+ "pos_tags": {
39
+ "feature": {
40
+ "num_classes": 54,
41
+ "names": [
42
+ "\"",
43
+ "''",
44
+ "#",
45
+ "$",
46
+ "(",
47
+ ")",
48
+ ",",
49
+ ".",
50
+ ":",
51
+ "``",
52
+ "CC",
53
+ "CD",
54
+ "DT",
55
+ "EX",
56
+ "FW",
57
+ "IN",
58
+ "JJ",
59
+ "JJR",
60
+ "JJS",
61
+ "LS",
62
+ "MD",
63
+ "NN",
64
+ "NNP",
65
+ "NNPS",
66
+ "NNS",
67
+ "NN|SYM",
68
+ "PDT",
69
+ "POS",
70
+ "PRP",
71
+ "PRP$",
72
+ "RB",
73
+ "RBR",
74
+ "RBS",
75
+ "RP",
76
+ "SYM",
77
+ "TO",
78
+ "UH",
79
+ "VB",
80
+ "VBD",
81
+ "VBG",
82
+ "VBN",
83
+ "VBP",
84
+ "VBZ",
85
+ "WDT",
86
+ "WP",
87
+ "WP$",
88
+ "WRB",
89
+ "HYPH",
90
+ "XX",
91
+ "NFP",
92
+ "AFX",
93
+ "ADD",
94
+ "-LRB-",
95
+ "-RRB-"
96
+ ],
97
+ "names_file": null,
98
+ "id": null,
99
+ "_type": "ClassLabel"
100
+ },
101
+ "length": -1,
102
+ "id": null,
103
+ "_type": "Sequence"
104
+ },
105
+ "parse_bit": {
106
+ "feature": {
107
+ "dtype": "string",
108
+ "id": null,
109
+ "_type": "Value"
110
+ },
111
+ "length": -1,
112
+ "id": null,
113
+ "_type": "Sequence"
114
+ },
115
+ "predicate_lemma": {
116
+ "feature": {
117
+ "dtype": "string",
118
+ "id": null,
119
+ "_type": "Value"
120
+ },
121
+ "length": -1,
122
+ "id": null,
123
+ "_type": "Sequence"
124
+ },
125
+ "predicate_framenet_id": {
126
+ "feature": {
127
+ "dtype": "string",
128
+ "id": null,
129
+ "_type": "Value"
130
+ },
131
+ "length": -1,
132
+ "id": null,
133
+ "_type": "Sequence"
134
+ },
135
+ "word_sense": {
136
+ "feature": {
137
+ "dtype": "string",
138
+ "id": null,
139
+ "_type": "Value"
140
+ },
141
+ "length": -1,
142
+ "id": null,
143
+ "_type": "Sequence"
144
+ },
145
+ "speaker": {
146
+ "feature": {
147
+ "dtype": "string",
148
+ "id": null,
149
+ "_type": "Value"
150
+ },
151
+ "length": -1,
152
+ "id": null,
153
+ "_type": "Sequence"
154
+ },
155
+ "ner_tags": {
156
+ "feature": {
157
+ "num_classes": 38,
158
+ "names": [
159
+ "B-PERSON",
160
+ "I-PERSON",
161
+ "B-NORP",
162
+ "I-NORP",
163
+ "B-FAC",
164
+ "I-FAC",
165
+ "B-ORG",
166
+ "I-ORG",
167
+ "B-GPE",
168
+ "I-GPE",
169
+ "B-LOC",
170
+ "I-LOC",
171
+ "B-PRODUCT",
172
+ "I-PRODUCT",
173
+ "B-EVENT",
174
+ "I-EVENT",
175
+ "B-WORK_OF_ART",
176
+ "I-WORK_OF_ART",
177
+ "B-LAW",
178
+ "I-LAW",
179
+ "B-LANGUAGE",
180
+ "I-LANGUAGE",
181
+ "B-DATE",
182
+ "I-DATE",
183
+ "B-TIME",
184
+ "I-TIME",
185
+ "B-PERCENT",
186
+ "I-PERCENT",
187
+ "B-MONEY",
188
+ "I-MONEY",
189
+ "B-QUANTITY",
190
+ "I-QUANTITY",
191
+ "B-ORDINAL",
192
+ "I-ORDINAL",
193
+ "B-CARDINAL",
194
+ "I-CARDINAL",
195
+ "*",
196
+ "0"
197
+ ],
198
+ "names_file": null,
199
+ "id": null,
200
+ "_type": "ClassLabel"
201
+ },
202
+ "length": -1,
203
+ "id": null,
204
+ "_type": "Sequence"
205
+ },
206
+ "verbal_predicates": {
207
+ "feature": {
208
+ "dtype": "string",
209
+ "id": null,
210
+ "_type": "Value"
211
+ },
212
+ "length": -1,
213
+ "id": null,
214
+ "_type": "Sequence"
215
+ }
216
+ },
217
+ "post_processed": null,
218
+ "supervised_keys": null,
219
+ "builder_name": "wino_bias",
220
+ "config_name": "wino_bias",
221
+ "version": {
222
+ "version_str": "4.0.0",
223
+ "description": null,
224
+ "major": 4,
225
+ "minor": 0,
226
+ "patch": 0
227
+ },
228
+ "splits": {
229
+ "train": {
230
+ "name": "train",
231
+ "num_bytes": 173899234,
232
+ "num_examples": 150335,
233
+ "dataset_name": "wino_bias"
234
+ }
235
+ },
236
+ "download_checksums": {
237
+ "https://drive.google.com/uc?export=download&id=14Im3BnNl-d2fYETYmiH5yq6eFGLVC3g0": {
238
+ "num_bytes": 268725744,
239
+ "checksum": "139a6511bcb9761f6306bba7d151bde6ec7ec82cc9c593127b848df8af5f68a1"
240
+ }
241
+ },
242
+ "download_size": 268725744,
243
+ "post_processing_size": null,
244
+ "dataset_size": 173899234,
245
+ "size_in_bytes": 442624978
246
+ },
247
+ "type1_pro": {
248
+ "description": "WinoBias, a Winograd-schema dataset for coreference resolution focused on gender bias.\nThe corpus contains Winograd-schema style sentences with entities corresponding to people\nreferred by their occupation (e.g. the nurse, the doctor, the carpenter).\n",
249
+ "citation": "@article{DBLP:journals/corr/abs-1804-06876,\n author = {Jieyu Zhao and\n Tianlu Wang and\n Mark Yatskar and\n Vicente Ordonez and\n Kai{-}Wei Chang},\n title = {Gender Bias in Coreference Resolution: Evaluation and Debiasing Methods},\n journal = {CoRR},\n volume = {abs/1804.06876},\n year = {2018},\n url = {http://arxiv.org/abs/1804.06876},\n archivePrefix = {arXiv},\n eprint = {1804.06876},\n timestamp = {Mon, 13 Aug 2018 16:47:01 +0200},\n biburl = {https://dblp.org/rec/journals/corr/abs-1804-06876.bib},\n bibsource = {dblp computer science bibliography, https://dblp.org}\n}\n",
250
+ "homepage": "https://uclanlp.github.io/corefBias/overview",
251
+ "license": "MIT License (https://github.com/uclanlp/corefBias/blob/master/LICENSE)",
252
+ "features": {
253
+ "document_id": {
254
+ "dtype": "string",
255
+ "id": null,
256
+ "_type": "Value"
257
+ },
258
+ "part_number": {
259
+ "dtype": "string",
260
+ "id": null,
261
+ "_type": "Value"
262
+ },
263
+ "word_number": {
264
+ "feature": {
265
+ "dtype": "int32",
266
+ "id": null,
267
+ "_type": "Value"
268
+ },
269
+ "length": -1,
270
+ "id": null,
271
+ "_type": "Sequence"
272
+ },
273
+ "tokens": {
274
+ "feature": {
275
+ "dtype": "string",
276
+ "id": null,
277
+ "_type": "Value"
278
+ },
279
+ "length": -1,
280
+ "id": null,
281
+ "_type": "Sequence"
282
+ },
283
+ "pos_tags": {
284
+ "feature": {
285
+ "num_classes": 55,
286
+ "names": [
287
+ "\"",
288
+ "''",
289
+ "#",
290
+ "$",
291
+ "(",
292
+ ")",
293
+ ",",
294
+ ".",
295
+ ":",
296
+ "``",
297
+ "CC",
298
+ "CD",
299
+ "DT",
300
+ "EX",
301
+ "FW",
302
+ "IN",
303
+ "JJ",
304
+ "JJR",
305
+ "JJS",
306
+ "LS",
307
+ "MD",
308
+ "NN",
309
+ "NNP",
310
+ "NNPS",
311
+ "NNS",
312
+ "NN|SYM",
313
+ "PDT",
314
+ "POS",
315
+ "PRP",
316
+ "PRP$",
317
+ "RB",
318
+ "RBR",
319
+ "RBS",
320
+ "RP",
321
+ "SYM",
322
+ "TO",
323
+ "UH",
324
+ "VB",
325
+ "VBD",
326
+ "VBG",
327
+ "VBN",
328
+ "VBP",
329
+ "VBZ",
330
+ "WDT",
331
+ "WP",
332
+ "WP$",
333
+ "WRB",
334
+ "HYPH",
335
+ "XX",
336
+ "NFP",
337
+ "AFX",
338
+ "ADD",
339
+ "-LRB-",
340
+ "-RRB-",
341
+ "-"
342
+ ],
343
+ "names_file": null,
344
+ "id": null,
345
+ "_type": "ClassLabel"
346
+ },
347
+ "length": -1,
348
+ "id": null,
349
+ "_type": "Sequence"
350
+ },
351
+ "parse_bit": {
352
+ "feature": {
353
+ "dtype": "string",
354
+ "id": null,
355
+ "_type": "Value"
356
+ },
357
+ "length": -1,
358
+ "id": null,
359
+ "_type": "Sequence"
360
+ },
361
+ "predicate_lemma": {
362
+ "feature": {
363
+ "dtype": "string",
364
+ "id": null,
365
+ "_type": "Value"
366
+ },
367
+ "length": -1,
368
+ "id": null,
369
+ "_type": "Sequence"
370
+ },
371
+ "predicate_framenet_id": {
372
+ "feature": {
373
+ "dtype": "string",
374
+ "id": null,
375
+ "_type": "Value"
376
+ },
377
+ "length": -1,
378
+ "id": null,
379
+ "_type": "Sequence"
380
+ },
381
+ "word_sense": {
382
+ "feature": {
383
+ "dtype": "string",
384
+ "id": null,
385
+ "_type": "Value"
386
+ },
387
+ "length": -1,
388
+ "id": null,
389
+ "_type": "Sequence"
390
+ },
391
+ "speaker": {
392
+ "feature": {
393
+ "dtype": "string",
394
+ "id": null,
395
+ "_type": "Value"
396
+ },
397
+ "length": -1,
398
+ "id": null,
399
+ "_type": "Sequence"
400
+ },
401
+ "ner_tags": {
402
+ "feature": {
403
+ "num_classes": 39,
404
+ "names": [
405
+ "B-PERSON",
406
+ "I-PERSON",
407
+ "B-NORP",
408
+ "I-NORP",
409
+ "B-FAC",
410
+ "I-FAC",
411
+ "B-ORG",
412
+ "I-ORG",
413
+ "B-GPE",
414
+ "I-GPE",
415
+ "B-LOC",
416
+ "I-LOC",
417
+ "B-PRODUCT",
418
+ "I-PRODUCT",
419
+ "B-EVENT",
420
+ "I-EVENT",
421
+ "B-WORK_OF_ART",
422
+ "I-WORK_OF_ART",
423
+ "B-LAW",
424
+ "I-LAW",
425
+ "B-LANGUAGE",
426
+ "I-LANGUAGE",
427
+ "B-DATE",
428
+ "I-DATE",
429
+ "B-TIME",
430
+ "I-TIME",
431
+ "B-PERCENT",
432
+ "I-PERCENT",
433
+ "B-MONEY",
434
+ "I-MONEY",
435
+ "B-QUANTITY",
436
+ "I-QUANTITY",
437
+ "B-ORDINAL",
438
+ "I-ORDINAL",
439
+ "B-CARDINAL",
440
+ "I-CARDINAL",
441
+ "*",
442
+ "0",
443
+ "-"
444
+ ],
445
+ "names_file": null,
446
+ "id": null,
447
+ "_type": "ClassLabel"
448
+ },
449
+ "length": -1,
450
+ "id": null,
451
+ "_type": "Sequence"
452
+ },
453
+ "verbal_predicates": {
454
+ "feature": {
455
+ "dtype": "string",
456
+ "id": null,
457
+ "_type": "Value"
458
+ },
459
+ "length": -1,
460
+ "id": null,
461
+ "_type": "Sequence"
462
+ },
463
+ "coreference_clusters": {
464
+ "feature": {
465
+ "dtype": "string",
466
+ "id": null,
467
+ "_type": "Value"
468
+ },
469
+ "length": -1,
470
+ "id": null,
471
+ "_type": "Sequence"
472
+ }
473
+ },
474
+ "post_processed": null,
475
+ "supervised_keys": null,
476
+ "builder_name": "wino_bias",
477
+ "config_name": "type1_pro",
478
+ "version": {
479
+ "version_str": "1.0.0",
480
+ "description": "",
481
+ "major": 1,
482
+ "minor": 0,
483
+ "patch": 0
484
+ },
485
+ "splits": {
486
+ "validation": {
487
+ "name": "validation",
488
+ "num_bytes": 379380,
489
+ "num_examples": 396,
490
+ "dataset_name": "wino_bias"
491
+ },
492
+ "test": {
493
+ "name": "test",
494
+ "num_bytes": 402041,
495
+ "num_examples": 396,
496
+ "dataset_name": "wino_bias"
497
+ }
498
+ },
499
+ "download_checksums": {
500
+ "https://raw.githubusercontent.com/uclanlp/corefBias/master/WinoBias/wino/data/conll_format/dev_type1_pro_stereotype.v4_auto_conll": {
501
+ "num_bytes": 407542,
502
+ "checksum": "21b5ca0cfb6acbb138ff37ea5deefd03783f8b7e65e296ead081d70cec9900c1"
503
+ },
504
+ "https://raw.githubusercontent.com/uclanlp/corefBias/master/WinoBias/wino/data/conll_format/test_type1_pro_stereotype.v4_auto_conll": {
505
+ "num_bytes": 438656,
506
+ "checksum": "419c0f4ecff4c6b1d81f0ce646e8ce1ebff9bde4729295064049ed9f97ad25d1"
507
+ }
508
+ },
509
+ "download_size": 846198,
510
+ "post_processing_size": null,
511
+ "dataset_size": 781421,
512
+ "size_in_bytes": 1627619
513
+ },
514
+ "type1_anti": {
515
+ "description": "WinoBias, a Winograd-schema dataset for coreference resolution focused on gender bias.\nThe corpus contains Winograd-schema style sentences with entities corresponding to people\nreferred by their occupation (e.g. the nurse, the doctor, the carpenter).\n",
516
+ "citation": "@article{DBLP:journals/corr/abs-1804-06876,\n author = {Jieyu Zhao and\n Tianlu Wang and\n Mark Yatskar and\n Vicente Ordonez and\n Kai{-}Wei Chang},\n title = {Gender Bias in Coreference Resolution: Evaluation and Debiasing Methods},\n journal = {CoRR},\n volume = {abs/1804.06876},\n year = {2018},\n url = {http://arxiv.org/abs/1804.06876},\n archivePrefix = {arXiv},\n eprint = {1804.06876},\n timestamp = {Mon, 13 Aug 2018 16:47:01 +0200},\n biburl = {https://dblp.org/rec/journals/corr/abs-1804-06876.bib},\n bibsource = {dblp computer science bibliography, https://dblp.org}\n}\n",
517
+ "homepage": "https://uclanlp.github.io/corefBias/overview",
518
+ "license": "MIT License (https://github.com/uclanlp/corefBias/blob/master/LICENSE)",
519
+ "features": {
520
+ "document_id": {
521
+ "dtype": "string",
522
+ "id": null,
523
+ "_type": "Value"
524
+ },
525
+ "part_number": {
526
+ "dtype": "string",
527
+ "id": null,
528
+ "_type": "Value"
529
+ },
530
+ "word_number": {
531
+ "feature": {
532
+ "dtype": "int32",
533
+ "id": null,
534
+ "_type": "Value"
535
+ },
536
+ "length": -1,
537
+ "id": null,
538
+ "_type": "Sequence"
539
+ },
540
+ "tokens": {
541
+ "feature": {
542
+ "dtype": "string",
543
+ "id": null,
544
+ "_type": "Value"
545
+ },
546
+ "length": -1,
547
+ "id": null,
548
+ "_type": "Sequence"
549
+ },
550
+ "pos_tags": {
551
+ "feature": {
552
+ "num_classes": 55,
553
+ "names": [
554
+ "\"",
555
+ "''",
556
+ "#",
557
+ "$",
558
+ "(",
559
+ ")",
560
+ ",",
561
+ ".",
562
+ ":",
563
+ "``",
564
+ "CC",
565
+ "CD",
566
+ "DT",
567
+ "EX",
568
+ "FW",
569
+ "IN",
570
+ "JJ",
571
+ "JJR",
572
+ "JJS",
573
+ "LS",
574
+ "MD",
575
+ "NN",
576
+ "NNP",
577
+ "NNPS",
578
+ "NNS",
579
+ "NN|SYM",
580
+ "PDT",
581
+ "POS",
582
+ "PRP",
583
+ "PRP$",
584
+ "RB",
585
+ "RBR",
586
+ "RBS",
587
+ "RP",
588
+ "SYM",
589
+ "TO",
590
+ "UH",
591
+ "VB",
592
+ "VBD",
593
+ "VBG",
594
+ "VBN",
595
+ "VBP",
596
+ "VBZ",
597
+ "WDT",
598
+ "WP",
599
+ "WP$",
600
+ "WRB",
601
+ "HYPH",
602
+ "XX",
603
+ "NFP",
604
+ "AFX",
605
+ "ADD",
606
+ "-LRB-",
607
+ "-RRB-",
608
+ "-"
609
+ ],
610
+ "names_file": null,
611
+ "id": null,
612
+ "_type": "ClassLabel"
613
+ },
614
+ "length": -1,
615
+ "id": null,
616
+ "_type": "Sequence"
617
+ },
618
+ "parse_bit": {
619
+ "feature": {
620
+ "dtype": "string",
621
+ "id": null,
622
+ "_type": "Value"
623
+ },
624
+ "length": -1,
625
+ "id": null,
626
+ "_type": "Sequence"
627
+ },
628
+ "predicate_lemma": {
629
+ "feature": {
630
+ "dtype": "string",
631
+ "id": null,
632
+ "_type": "Value"
633
+ },
634
+ "length": -1,
635
+ "id": null,
636
+ "_type": "Sequence"
637
+ },
638
+ "predicate_framenet_id": {
639
+ "feature": {
640
+ "dtype": "string",
641
+ "id": null,
642
+ "_type": "Value"
643
+ },
644
+ "length": -1,
645
+ "id": null,
646
+ "_type": "Sequence"
647
+ },
648
+ "word_sense": {
649
+ "feature": {
650
+ "dtype": "string",
651
+ "id": null,
652
+ "_type": "Value"
653
+ },
654
+ "length": -1,
655
+ "id": null,
656
+ "_type": "Sequence"
657
+ },
658
+ "speaker": {
659
+ "feature": {
660
+ "dtype": "string",
661
+ "id": null,
662
+ "_type": "Value"
663
+ },
664
+ "length": -1,
665
+ "id": null,
666
+ "_type": "Sequence"
667
+ },
668
+ "ner_tags": {
669
+ "feature": {
670
+ "num_classes": 39,
671
+ "names": [
672
+ "B-PERSON",
673
+ "I-PERSON",
674
+ "B-NORP",
675
+ "I-NORP",
676
+ "B-FAC",
677
+ "I-FAC",
678
+ "B-ORG",
679
+ "I-ORG",
680
+ "B-GPE",
681
+ "I-GPE",
682
+ "B-LOC",
683
+ "I-LOC",
684
+ "B-PRODUCT",
685
+ "I-PRODUCT",
686
+ "B-EVENT",
687
+ "I-EVENT",
688
+ "B-WORK_OF_ART",
689
+ "I-WORK_OF_ART",
690
+ "B-LAW",
691
+ "I-LAW",
692
+ "B-LANGUAGE",
693
+ "I-LANGUAGE",
694
+ "B-DATE",
695
+ "I-DATE",
696
+ "B-TIME",
697
+ "I-TIME",
698
+ "B-PERCENT",
699
+ "I-PERCENT",
700
+ "B-MONEY",
701
+ "I-MONEY",
702
+ "B-QUANTITY",
703
+ "I-QUANTITY",
704
+ "B-ORDINAL",
705
+ "I-ORDINAL",
706
+ "B-CARDINAL",
707
+ "I-CARDINAL",
708
+ "*",
709
+ "0",
710
+ "-"
711
+ ],
712
+ "names_file": null,
713
+ "id": null,
714
+ "_type": "ClassLabel"
715
+ },
716
+ "length": -1,
717
+ "id": null,
718
+ "_type": "Sequence"
719
+ },
720
+ "verbal_predicates": {
721
+ "feature": {
722
+ "dtype": "string",
723
+ "id": null,
724
+ "_type": "Value"
725
+ },
726
+ "length": -1,
727
+ "id": null,
728
+ "_type": "Sequence"
729
+ },
730
+ "coreference_clusters": {
731
+ "feature": {
732
+ "dtype": "string",
733
+ "id": null,
734
+ "_type": "Value"
735
+ },
736
+ "length": -1,
737
+ "id": null,
738
+ "_type": "Sequence"
739
+ }
740
+ },
741
+ "post_processed": null,
742
+ "supervised_keys": null,
743
+ "builder_name": "wino_bias",
744
+ "config_name": "type1_anti",
745
+ "version": {
746
+ "version_str": "1.0.0",
747
+ "description": "",
748
+ "major": 1,
749
+ "minor": 0,
750
+ "patch": 0
751
+ },
752
+ "splits": {
753
+ "validation": {
754
+ "name": "validation",
755
+ "num_bytes": 380846,
756
+ "num_examples": 396,
757
+ "dataset_name": "wino_bias"
758
+ },
759
+ "test": {
760
+ "name": "test",
761
+ "num_bytes": 403229,
762
+ "num_examples": 396,
763
+ "dataset_name": "wino_bias"
764
+ }
765
+ },
766
+ "download_checksums": {
767
+ "https://raw.githubusercontent.com/uclanlp/corefBias/master/WinoBias/wino/data/conll_format/dev_type1_anti_stereotype.v4_auto_conll": {
768
+ "num_bytes": 431039,
769
+ "checksum": "5c9565b41c546d04172190b92009a7a02375183c9232522fdaf195b27bee45a4"
770
+ },
771
+ "https://raw.githubusercontent.com/uclanlp/corefBias/master/WinoBias/wino/data/conll_format/test_type1_anti_stereotype.v4_auto_conll": {
772
+ "num_bytes": 463272,
773
+ "checksum": "68a7f4f0f055f5cddeed3b0a07a572fded2b2077c1afd2bb596c02e5fe0c9476"
774
+ }
775
+ },
776
+ "download_size": 894311,
777
+ "post_processing_size": null,
778
+ "dataset_size": 784075,
779
+ "size_in_bytes": 1678386
780
+ },
781
+ "type2_pro": {
782
+ "description": "WinoBias, a Winograd-schema dataset for coreference resolution focused on gender bias.\nThe corpus contains Winograd-schema style sentences with entities corresponding to people\nreferred by their occupation (e.g. the nurse, the doctor, the carpenter).\n",
783
+ "citation": "@article{DBLP:journals/corr/abs-1804-06876,\n author = {Jieyu Zhao and\n Tianlu Wang and\n Mark Yatskar and\n Vicente Ordonez and\n Kai{-}Wei Chang},\n title = {Gender Bias in Coreference Resolution: Evaluation and Debiasing Methods},\n journal = {CoRR},\n volume = {abs/1804.06876},\n year = {2018},\n url = {http://arxiv.org/abs/1804.06876},\n archivePrefix = {arXiv},\n eprint = {1804.06876},\n timestamp = {Mon, 13 Aug 2018 16:47:01 +0200},\n biburl = {https://dblp.org/rec/journals/corr/abs-1804-06876.bib},\n bibsource = {dblp computer science bibliography, https://dblp.org}\n}\n",
784
+ "homepage": "https://uclanlp.github.io/corefBias/overview",
785
+ "license": "MIT License (https://github.com/uclanlp/corefBias/blob/master/LICENSE)",
786
+ "features": {
787
+ "document_id": {
788
+ "dtype": "string",
789
+ "id": null,
790
+ "_type": "Value"
791
+ },
792
+ "part_number": {
793
+ "dtype": "string",
794
+ "id": null,
795
+ "_type": "Value"
796
+ },
797
+ "word_number": {
798
+ "feature": {
799
+ "dtype": "int32",
800
+ "id": null,
801
+ "_type": "Value"
802
+ },
803
+ "length": -1,
804
+ "id": null,
805
+ "_type": "Sequence"
806
+ },
807
+ "tokens": {
808
+ "feature": {
809
+ "dtype": "string",
810
+ "id": null,
811
+ "_type": "Value"
812
+ },
813
+ "length": -1,
814
+ "id": null,
815
+ "_type": "Sequence"
816
+ },
817
+ "pos_tags": {
818
+ "feature": {
819
+ "num_classes": 55,
820
+ "names": [
821
+ "\"",
822
+ "''",
823
+ "#",
824
+ "$",
825
+ "(",
826
+ ")",
827
+ ",",
828
+ ".",
829
+ ":",
830
+ "``",
831
+ "CC",
832
+ "CD",
833
+ "DT",
834
+ "EX",
835
+ "FW",
836
+ "IN",
837
+ "JJ",
838
+ "JJR",
839
+ "JJS",
840
+ "LS",
841
+ "MD",
842
+ "NN",
843
+ "NNP",
844
+ "NNPS",
845
+ "NNS",
846
+ "NN|SYM",
847
+ "PDT",
848
+ "POS",
849
+ "PRP",
850
+ "PRP$",
851
+ "RB",
852
+ "RBR",
853
+ "RBS",
854
+ "RP",
855
+ "SYM",
856
+ "TO",
857
+ "UH",
858
+ "VB",
859
+ "VBD",
860
+ "VBG",
861
+ "VBN",
862
+ "VBP",
863
+ "VBZ",
864
+ "WDT",
865
+ "WP",
866
+ "WP$",
867
+ "WRB",
868
+ "HYPH",
869
+ "XX",
870
+ "NFP",
871
+ "AFX",
872
+ "ADD",
873
+ "-LRB-",
874
+ "-RRB-",
875
+ "-"
876
+ ],
877
+ "names_file": null,
878
+ "id": null,
879
+ "_type": "ClassLabel"
880
+ },
881
+ "length": -1,
882
+ "id": null,
883
+ "_type": "Sequence"
884
+ },
885
+ "parse_bit": {
886
+ "feature": {
887
+ "dtype": "string",
888
+ "id": null,
889
+ "_type": "Value"
890
+ },
891
+ "length": -1,
892
+ "id": null,
893
+ "_type": "Sequence"
894
+ },
895
+ "predicate_lemma": {
896
+ "feature": {
897
+ "dtype": "string",
898
+ "id": null,
899
+ "_type": "Value"
900
+ },
901
+ "length": -1,
902
+ "id": null,
903
+ "_type": "Sequence"
904
+ },
905
+ "predicate_framenet_id": {
906
+ "feature": {
907
+ "dtype": "string",
908
+ "id": null,
909
+ "_type": "Value"
910
+ },
911
+ "length": -1,
912
+ "id": null,
913
+ "_type": "Sequence"
914
+ },
915
+ "word_sense": {
916
+ "feature": {
917
+ "dtype": "string",
918
+ "id": null,
919
+ "_type": "Value"
920
+ },
921
+ "length": -1,
922
+ "id": null,
923
+ "_type": "Sequence"
924
+ },
925
+ "speaker": {
926
+ "feature": {
927
+ "dtype": "string",
928
+ "id": null,
929
+ "_type": "Value"
930
+ },
931
+ "length": -1,
932
+ "id": null,
933
+ "_type": "Sequence"
934
+ },
935
+ "ner_tags": {
936
+ "feature": {
937
+ "num_classes": 39,
938
+ "names": [
939
+ "B-PERSON",
940
+ "I-PERSON",
941
+ "B-NORP",
942
+ "I-NORP",
943
+ "B-FAC",
944
+ "I-FAC",
945
+ "B-ORG",
946
+ "I-ORG",
947
+ "B-GPE",
948
+ "I-GPE",
949
+ "B-LOC",
950
+ "I-LOC",
951
+ "B-PRODUCT",
952
+ "I-PRODUCT",
953
+ "B-EVENT",
954
+ "I-EVENT",
955
+ "B-WORK_OF_ART",
956
+ "I-WORK_OF_ART",
957
+ "B-LAW",
958
+ "I-LAW",
959
+ "B-LANGUAGE",
960
+ "I-LANGUAGE",
961
+ "B-DATE",
962
+ "I-DATE",
963
+ "B-TIME",
964
+ "I-TIME",
965
+ "B-PERCENT",
966
+ "I-PERCENT",
967
+ "B-MONEY",
968
+ "I-MONEY",
969
+ "B-QUANTITY",
970
+ "I-QUANTITY",
971
+ "B-ORDINAL",
972
+ "I-ORDINAL",
973
+ "B-CARDINAL",
974
+ "I-CARDINAL",
975
+ "*",
976
+ "0",
977
+ "-"
978
+ ],
979
+ "names_file": null,
980
+ "id": null,
981
+ "_type": "ClassLabel"
982
+ },
983
+ "length": -1,
984
+ "id": null,
985
+ "_type": "Sequence"
986
+ },
987
+ "verbal_predicates": {
988
+ "feature": {
989
+ "dtype": "string",
990
+ "id": null,
991
+ "_type": "Value"
992
+ },
993
+ "length": -1,
994
+ "id": null,
995
+ "_type": "Sequence"
996
+ },
997
+ "coreference_clusters": {
998
+ "feature": {
999
+ "dtype": "string",
1000
+ "id": null,
1001
+ "_type": "Value"
1002
+ },
1003
+ "length": -1,
1004
+ "id": null,
1005
+ "_type": "Sequence"
1006
+ }
1007
+ },
1008
+ "post_processed": null,
1009
+ "supervised_keys": null,
1010
+ "builder_name": "wino_bias",
1011
+ "config_name": "type2_pro",
1012
+ "version": {
1013
+ "version_str": "1.0.0",
1014
+ "description": "",
1015
+ "major": 1,
1016
+ "minor": 0,
1017
+ "patch": 0
1018
+ },
1019
+ "splits": {
1020
+ "validation": {
1021
+ "name": "validation",
1022
+ "num_bytes": 367293,
1023
+ "num_examples": 396,
1024
+ "dataset_name": "wino_bias"
1025
+ },
1026
+ "test": {
1027
+ "name": "test",
1028
+ "num_bytes": 375480,
1029
+ "num_examples": 396,
1030
+ "dataset_name": "wino_bias"
1031
+ }
1032
+ },
1033
+ "download_checksums": {
1034
+ "https://raw.githubusercontent.com/uclanlp/corefBias/master/WinoBias/wino/data/conll_format/dev_type2_pro_stereotype.v4_auto_conll": {
1035
+ "num_bytes": 393889,
1036
+ "checksum": "8f96017dee3a1cfeda74da3b2e8b7c794f48680f21d83ae4118d0f1e22692ea0"
1037
+ },
1038
+ "https://raw.githubusercontent.com/uclanlp/corefBias/master/WinoBias/wino/data/conll_format/test_type2_pro_stereotype.v4_auto_conll": {
1039
+ "num_bytes": 408536,
1040
+ "checksum": "7720dee1432dc6a3d6077c442ddea37c63773e80f4f75c79b27e00087f88d574"
1041
+ }
1042
+ },
1043
+ "download_size": 802425,
1044
+ "post_processing_size": null,
1045
+ "dataset_size": 742773,
1046
+ "size_in_bytes": 1545198
1047
+ },
1048
+ "type2_anti": {
1049
+ "description": "WinoBias, a Winograd-schema dataset for coreference resolution focused on gender bias.\nThe corpus contains Winograd-schema style sentences with entities corresponding to people\nreferred by their occupation (e.g. the nurse, the doctor, the carpenter).\n",
1050
+ "citation": "@article{DBLP:journals/corr/abs-1804-06876,\n author = {Jieyu Zhao and\n Tianlu Wang and\n Mark Yatskar and\n Vicente Ordonez and\n Kai{-}Wei Chang},\n title = {Gender Bias in Coreference Resolution: Evaluation and Debiasing Methods},\n journal = {CoRR},\n volume = {abs/1804.06876},\n year = {2018},\n url = {http://arxiv.org/abs/1804.06876},\n archivePrefix = {arXiv},\n eprint = {1804.06876},\n timestamp = {Mon, 13 Aug 2018 16:47:01 +0200},\n biburl = {https://dblp.org/rec/journals/corr/abs-1804-06876.bib},\n bibsource = {dblp computer science bibliography, https://dblp.org}\n}\n",
1051
+ "homepage": "https://uclanlp.github.io/corefBias/overview",
1052
+ "license": "MIT License (https://github.com/uclanlp/corefBias/blob/master/LICENSE)",
1053
+ "features": {
1054
+ "document_id": {
1055
+ "dtype": "string",
1056
+ "id": null,
1057
+ "_type": "Value"
1058
+ },
1059
+ "part_number": {
1060
+ "dtype": "string",
1061
+ "id": null,
1062
+ "_type": "Value"
1063
+ },
1064
+ "word_number": {
1065
+ "feature": {
1066
+ "dtype": "int32",
1067
+ "id": null,
1068
+ "_type": "Value"
1069
+ },
1070
+ "length": -1,
1071
+ "id": null,
1072
+ "_type": "Sequence"
1073
+ },
1074
+ "tokens": {
1075
+ "feature": {
1076
+ "dtype": "string",
1077
+ "id": null,
1078
+ "_type": "Value"
1079
+ },
1080
+ "length": -1,
1081
+ "id": null,
1082
+ "_type": "Sequence"
1083
+ },
1084
+ "pos_tags": {
1085
+ "feature": {
1086
+ "num_classes": 55,
1087
+ "names": [
1088
+ "\"",
1089
+ "''",
1090
+ "#",
1091
+ "$",
1092
+ "(",
1093
+ ")",
1094
+ ",",
1095
+ ".",
1096
+ ":",
1097
+ "``",
1098
+ "CC",
1099
+ "CD",
1100
+ "DT",
1101
+ "EX",
1102
+ "FW",
1103
+ "IN",
1104
+ "JJ",
1105
+ "JJR",
1106
+ "JJS",
1107
+ "LS",
1108
+ "MD",
1109
+ "NN",
1110
+ "NNP",
1111
+ "NNPS",
1112
+ "NNS",
1113
+ "NN|SYM",
1114
+ "PDT",
1115
+ "POS",
1116
+ "PRP",
1117
+ "PRP$",
1118
+ "RB",
1119
+ "RBR",
1120
+ "RBS",
1121
+ "RP",
1122
+ "SYM",
1123
+ "TO",
1124
+ "UH",
1125
+ "VB",
1126
+ "VBD",
1127
+ "VBG",
1128
+ "VBN",
1129
+ "VBP",
1130
+ "VBZ",
1131
+ "WDT",
1132
+ "WP",
1133
+ "WP$",
1134
+ "WRB",
1135
+ "HYPH",
1136
+ "XX",
1137
+ "NFP",
1138
+ "AFX",
1139
+ "ADD",
1140
+ "-LRB-",
1141
+ "-RRB-",
1142
+ "-"
1143
+ ],
1144
+ "names_file": null,
1145
+ "id": null,
1146
+ "_type": "ClassLabel"
1147
+ },
1148
+ "length": -1,
1149
+ "id": null,
1150
+ "_type": "Sequence"
1151
+ },
1152
+ "parse_bit": {
1153
+ "feature": {
1154
+ "dtype": "string",
1155
+ "id": null,
1156
+ "_type": "Value"
1157
+ },
1158
+ "length": -1,
1159
+ "id": null,
1160
+ "_type": "Sequence"
1161
+ },
1162
+ "predicate_lemma": {
1163
+ "feature": {
1164
+ "dtype": "string",
1165
+ "id": null,
1166
+ "_type": "Value"
1167
+ },
1168
+ "length": -1,
1169
+ "id": null,
1170
+ "_type": "Sequence"
1171
+ },
1172
+ "predicate_framenet_id": {
1173
+ "feature": {
1174
+ "dtype": "string",
1175
+ "id": null,
1176
+ "_type": "Value"
1177
+ },
1178
+ "length": -1,
1179
+ "id": null,
1180
+ "_type": "Sequence"
1181
+ },
1182
+ "word_sense": {
1183
+ "feature": {
1184
+ "dtype": "string",
1185
+ "id": null,
1186
+ "_type": "Value"
1187
+ },
1188
+ "length": -1,
1189
+ "id": null,
1190
+ "_type": "Sequence"
1191
+ },
1192
+ "speaker": {
1193
+ "feature": {
1194
+ "dtype": "string",
1195
+ "id": null,
1196
+ "_type": "Value"
1197
+ },
1198
+ "length": -1,
1199
+ "id": null,
1200
+ "_type": "Sequence"
1201
+ },
1202
+ "ner_tags": {
1203
+ "feature": {
1204
+ "num_classes": 39,
1205
+ "names": [
1206
+ "B-PERSON",
1207
+ "I-PERSON",
1208
+ "B-NORP",
1209
+ "I-NORP",
1210
+ "B-FAC",
1211
+ "I-FAC",
1212
+ "B-ORG",
1213
+ "I-ORG",
1214
+ "B-GPE",
1215
+ "I-GPE",
1216
+ "B-LOC",
1217
+ "I-LOC",
1218
+ "B-PRODUCT",
1219
+ "I-PRODUCT",
1220
+ "B-EVENT",
1221
+ "I-EVENT",
1222
+ "B-WORK_OF_ART",
1223
+ "I-WORK_OF_ART",
1224
+ "B-LAW",
1225
+ "I-LAW",
1226
+ "B-LANGUAGE",
1227
+ "I-LANGUAGE",
1228
+ "B-DATE",
1229
+ "I-DATE",
1230
+ "B-TIME",
1231
+ "I-TIME",
1232
+ "B-PERCENT",
1233
+ "I-PERCENT",
1234
+ "B-MONEY",
1235
+ "I-MONEY",
1236
+ "B-QUANTITY",
1237
+ "I-QUANTITY",
1238
+ "B-ORDINAL",
1239
+ "I-ORDINAL",
1240
+ "B-CARDINAL",
1241
+ "I-CARDINAL",
1242
+ "*",
1243
+ "0",
1244
+ "-"
1245
+ ],
1246
+ "names_file": null,
1247
+ "id": null,
1248
+ "_type": "ClassLabel"
1249
+ },
1250
+ "length": -1,
1251
+ "id": null,
1252
+ "_type": "Sequence"
1253
+ },
1254
+ "verbal_predicates": {
1255
+ "feature": {
1256
+ "dtype": "string",
1257
+ "id": null,
1258
+ "_type": "Value"
1259
+ },
1260
+ "length": -1,
1261
+ "id": null,
1262
+ "_type": "Sequence"
1263
+ },
1264
+ "coreference_clusters": {
1265
+ "feature": {
1266
+ "dtype": "string",
1267
+ "id": null,
1268
+ "_type": "Value"
1269
+ },
1270
+ "length": -1,
1271
+ "id": null,
1272
+ "_type": "Sequence"
1273
+ }
1274
+ },
1275
+ "post_processed": null,
1276
+ "supervised_keys": null,
1277
+ "builder_name": "wino_bias",
1278
+ "config_name": "type2_anti",
1279
+ "version": {
1280
+ "version_str": "1.0.0",
1281
+ "description": "",
1282
+ "major": 1,
1283
+ "minor": 0,
1284
+ "patch": 0
1285
+ },
1286
+ "splits": {
1287
+ "validation": {
1288
+ "name": "validation",
1289
+ "num_bytes": 368757,
1290
+ "num_examples": 396,
1291
+ "dataset_name": "wino_bias"
1292
+ },
1293
+ "test": {
1294
+ "name": "test",
1295
+ "num_bytes": 377262,
1296
+ "num_examples": 396,
1297
+ "dataset_name": "wino_bias"
1298
+ }
1299
+ },
1300
+ "download_checksums": {
1301
+ "https://raw.githubusercontent.com/uclanlp/corefBias/master/WinoBias/wino/data/conll_format/dev_type2_anti_stereotype.v4_auto_conll": {
1302
+ "num_bytes": 416645,
1303
+ "checksum": "022c18ce9fa07161b24daa6d79465190f127fde5c15e796778b12d950966fb03"
1304
+ },
1305
+ "https://raw.githubusercontent.com/uclanlp/corefBias/master/WinoBias/wino/data/conll_format/test_type2_anti_stereotype.v4_auto_conll": {
1306
+ "num_bytes": 432159,
1307
+ "checksum": "c9a6b66d60ce6a008f0d8cb619786042454abc6e9502d134cf88352aed7da57b"
1308
+ }
1309
+ },
1310
+ "download_size": 848804,
1311
+ "post_processing_size": null,
1312
+ "dataset_size": 746019,
1313
+ "size_in_bytes": 1594823
1314
+ }
1315
+ }
dummy/{wino_bias/4.0.0 → type1_anti/1.0.0}/dummy_data.zip RENAMED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a5c20b4ec3916367f3f8b88384e3c6bde4aefdedfa670cbfb241f05e0156411c
3
- size 764
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:49caed7f2dc4ade7cd16916f5d661530392938b5a82d1bdb268980e04aa8d306
3
+ size 854
dummy/type1_pro/1.0.0/dummy_data.zip ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:03d276a6dff6bbf859228dbbe16221620a359e7146bd61f0e7a32b18877e39ec
3
+ size 843
dummy/type2_anti/1.0.0/dummy_data.zip ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8fa16711b9beb853b87cc6af7881ef31ad23f7bcc183d44b0e0ec3e0daa94fd1
3
+ size 836
dummy/type2_pro/1.0.0/dummy_data.zip ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d80d4c4bf2bce5c643e9bfb681b508be290611a86c3a6753171301ddd4811d10
3
+ size 828
wino_bias.py CHANGED
@@ -14,7 +14,8 @@
14
  # limitations under the License.
15
  """WinoBias: Winograd-schema dataset for detecting gender bias"""
16
 
17
- from __future__ import absolute_import, division, print_function
 
18
 
19
  import datasets
20
 
@@ -49,14 +50,17 @@ _HOMEPAGE = "https://uclanlp.github.io/corefBias/overview"
49
 
50
  _LICENSE = "MIT License (https://github.com/uclanlp/corefBias/blob/master/LICENSE)"
51
 
52
- _URL = "https://drive.google.com/uc?export=download&id=14Im3BnNl-d2fYETYmiH5yq6eFGLVC3g0"
 
 
 
 
 
53
 
54
 
55
  class WinoBias(datasets.GeneratorBasedBuilder):
56
  """WinoBias: Winograd-schema dataset for detecting gender bias"""
57
 
58
- VERSION = datasets.Version("4.0.0")
59
-
60
  # This is an example of a dataset with multiple configurations.
61
  # If you don't want/need to define several sub-sets in your dataset,
62
  # just remove the BUILDER_CONFIG_CLASS and the BUILDER_CONFIGS attributes.
@@ -68,11 +72,30 @@ class WinoBias(datasets.GeneratorBasedBuilder):
68
  # You will be able to load one or the other configurations in the following list with
69
  # data = datasets.load_dataset('my_dataset', 'first_domain')
70
  # data = datasets.load_dataset('my_dataset', 'second_domain')
 
 
 
 
 
 
 
 
71
  BUILDER_CONFIGS = [
72
- datasets.BuilderConfig(
73
- name="wino_bias",
74
- version=VERSION,
75
- description="WinoBias: Winograd-schema dataset for detecting gender bias",
 
 
 
 
 
 
 
 
 
 
 
76
  ),
77
  ]
78
 
@@ -145,6 +168,7 @@ class WinoBias(datasets.GeneratorBasedBuilder):
145
  "ADD",
146
  "-LRB-",
147
  "-RRB-",
 
148
  ]
149
  )
150
  ),
@@ -194,10 +218,12 @@ class WinoBias(datasets.GeneratorBasedBuilder):
194
  "I-CARDINAL",
195
  "*",
196
  "0",
 
197
  ]
198
  )
199
  ),
200
  "verbal_predicates": datasets.Sequence(datasets.Value("string")),
 
201
  }
202
  ),
203
  supervised_keys=None,
@@ -211,13 +237,20 @@ class WinoBias(datasets.GeneratorBasedBuilder):
211
 
212
  def _split_generators(self, dl_manager):
213
  """Returns SplitGenerators."""
214
- data_dir = dl_manager.download_and_extract(_URL)
 
 
215
  return [
216
  datasets.SplitGenerator(
217
- name=datasets.Split.TRAIN,
 
 
 
 
 
218
  # These kwargs will be passed to _generate_examples
219
- gen_kwargs={"filepath": data_dir},
220
- )
221
  ]
222
 
223
  def _generate_examples(self, filepath):
@@ -237,6 +270,9 @@ class WinoBias(datasets.GeneratorBasedBuilder):
237
  ner_tags = []
238
  ner_start = False
239
  verbal_predicates = []
 
 
 
240
  for line in f:
241
  if line.startswith("#begin") or line.startswith("#end"):
242
  continue
@@ -255,7 +291,11 @@ class WinoBias(datasets.GeneratorBasedBuilder):
255
  "speaker": speaker,
256
  "ner_tags": ner_tags,
257
  "verbal_predicates": verbal_predicates,
 
 
 
258
  }
 
259
  word_num = []
260
  tokens = []
261
  pos_tags = []
@@ -266,8 +306,11 @@ class WinoBias(datasets.GeneratorBasedBuilder):
266
  speaker = []
267
  ner_tags = []
268
  verbal_predicates = []
 
 
 
269
  else:
270
- splits = [s for s in line.split(" ") if s]
271
  if len(splits) > 7:
272
  document_id = splits[0]
273
  part_number = splits[1]
@@ -280,6 +323,7 @@ class WinoBias(datasets.GeneratorBasedBuilder):
280
  word_sense.append(splits[8])
281
  speaker.append(splits[9])
282
  ner_word = splits[10]
 
283
  if ")" in ner_word and ner_start:
284
  ner_start = False
285
  ner_word = "0"
@@ -294,6 +338,20 @@ class WinoBias(datasets.GeneratorBasedBuilder):
294
  word_is_verbal_predicate = any(["(V" in x for x in splits[11:-1]])
295
  if word_is_verbal_predicate:
296
  verbal_predicates.append(splits[3])
 
 
 
 
 
 
 
 
 
 
 
 
 
 
297
  if tokens:
298
  # add the last one
299
  id_ += 1
@@ -310,4 +368,5 @@ class WinoBias(datasets.GeneratorBasedBuilder):
310
  "speaker": speaker,
311
  "ner_tags": ner_tags,
312
  "verbal_predicates": verbal_predicates,
 
313
  }
14
  # limitations under the License.
15
  """WinoBias: Winograd-schema dataset for detecting gender bias"""
16
 
17
+
18
+ import collections
19
 
20
  import datasets
21
 
50
 
51
  _LICENSE = "MIT License (https://github.com/uclanlp/corefBias/blob/master/LICENSE)"
52
 
53
+ _URL = "https://raw.githubusercontent.com/uclanlp/corefBias/master/WinoBias/wino/data/conll_format"
54
+
55
+
56
+ class WinoBiasConfig(datasets.BuilderConfig):
57
+ def __init__(self, **kwargs):
58
+ super(WinoBiasConfig, self).__init__(version=datasets.Version("1.0.0", ""), **kwargs)
59
 
60
 
61
  class WinoBias(datasets.GeneratorBasedBuilder):
62
  """WinoBias: Winograd-schema dataset for detecting gender bias"""
63
 
 
 
64
  # This is an example of a dataset with multiple configurations.
65
  # If you don't want/need to define several sub-sets in your dataset,
66
  # just remove the BUILDER_CONFIG_CLASS and the BUILDER_CONFIGS attributes.
72
  # You will be able to load one or the other configurations in the following list with
73
  # data = datasets.load_dataset('my_dataset', 'first_domain')
74
  # data = datasets.load_dataset('my_dataset', 'second_domain')
75
+ def __init__(self, *args, writer_batch_size=None, **kwargs):
76
+ super(WinoBias, self).__init__(*args, **kwargs)
77
+ # Batch size used by the ArrowWriter
78
+ # It defines the number of samples that are kept in memory before writing them
79
+ # and also the length of the arrow chunks
80
+ # None means that the ArrowWriter will use its default value
81
+ self._writer_batch_size = writer_batch_size or 100
82
+
83
  BUILDER_CONFIGS = [
84
+ WinoBiasConfig(
85
+ name="type1_pro",
86
+ description="winoBias type1_pro_stereotype data in cornll format",
87
+ ),
88
+ WinoBiasConfig(
89
+ name="type1_anti",
90
+ description="winoBias type1_anti_stereotype data in cornll format",
91
+ ),
92
+ WinoBiasConfig(
93
+ name="type2_pro",
94
+ description="winoBias type2_pro_stereotype data in cornll format",
95
+ ),
96
+ WinoBiasConfig(
97
+ name="type2_anti",
98
+ description="winoBias type2_anti_stereotype data in cornll format",
99
  ),
100
  ]
101
 
168
  "ADD",
169
  "-LRB-",
170
  "-RRB-",
171
+ "-",
172
  ]
173
  )
174
  ),
218
  "I-CARDINAL",
219
  "*",
220
  "0",
221
+ "-",
222
  ]
223
  )
224
  ),
225
  "verbal_predicates": datasets.Sequence(datasets.Value("string")),
226
+ "coreference_clusters": datasets.Sequence(datasets.Value("string")),
227
  }
228
  ),
229
  supervised_keys=None,
237
 
238
  def _split_generators(self, dl_manager):
239
  """Returns SplitGenerators."""
240
+
241
+ dev_data_dir = dl_manager.download(_URL + "/dev_" + self.config.name + "_stereotype.v4_auto_conll")
242
+ test_data_dir = dl_manager.download(_URL + "/test_" + self.config.name + "_stereotype.v4_auto_conll")
243
  return [
244
  datasets.SplitGenerator(
245
+ name=datasets.Split.VALIDATION,
246
+ # These kwargs will be passed to _generate_examples
247
+ gen_kwargs={"filepath": dev_data_dir},
248
+ ),
249
+ datasets.SplitGenerator(
250
+ name=datasets.Split.TEST,
251
  # These kwargs will be passed to _generate_examples
252
+ gen_kwargs={"filepath": test_data_dir},
253
+ ),
254
  ]
255
 
256
  def _generate_examples(self, filepath):
270
  ner_tags = []
271
  ner_start = False
272
  verbal_predicates = []
273
+ coreference = []
274
+ clusters = collections.defaultdict(list)
275
+ coref_stacks = collections.defaultdict(list)
276
  for line in f:
277
  if line.startswith("#begin") or line.startswith("#end"):
278
  continue
291
  "speaker": speaker,
292
  "ner_tags": ner_tags,
293
  "verbal_predicates": verbal_predicates,
294
+ "coreference_clusters": sum(
295
+ clusters[1], []
296
+ ), # flatten the list as writing the exmaples needs an array.
297
  }
298
+
299
  word_num = []
300
  tokens = []
301
  pos_tags = []
306
  speaker = []
307
  ner_tags = []
308
  verbal_predicates = []
309
+ coreference = []
310
+ clusters = collections.defaultdict(list)
311
+ coref_stacks = collections.defaultdict(list)
312
  else:
313
+ splits = [s for s in line.split() if s]
314
  if len(splits) > 7:
315
  document_id = splits[0]
316
  part_number = splits[1]
323
  word_sense.append(splits[8])
324
  speaker.append(splits[9])
325
  ner_word = splits[10]
326
+ coreference = splits[-1]
327
  if ")" in ner_word and ner_start:
328
  ner_start = False
329
  ner_word = "0"
338
  word_is_verbal_predicate = any(["(V" in x for x in splits[11:-1]])
339
  if word_is_verbal_predicate:
340
  verbal_predicates.append(splits[3])
341
+ if coreference != "-":
342
+ for segment in coreference.split("|"):
343
+ if segment[0] == "(":
344
+ if segment[-1] == ")":
345
+ cluster_id = int(segment[1:-1])
346
+ clusters[cluster_id].append([splits[2], splits[2]])
347
+ else:
348
+ cluster_id = int(segment[1:])
349
+ coref_stacks[cluster_id].append(splits[2])
350
+ else:
351
+ cluster_id = int(segment[:-1])
352
+ start = coref_stacks[cluster_id].pop()
353
+ clusters[cluster_id].append([start, splits[2]])
354
+
355
  if tokens:
356
  # add the last one
357
  id_ += 1
368
  "speaker": speaker,
369
  "ner_tags": ner_tags,
370
  "verbal_predicates": verbal_predicates,
371
+ "coreference_clusters": sum(clusters[1], []),
372
  }