word
stringlengths
1
21
id_example
int64
0
299
text_id
int64
0
149
position_id
int64
1
65
annotator_id
stringclasses
99 values
reading_time
int64
0
11.8k
gaze_duration
float64
0
5.37k
fixations
int64
0
34
first_fixation_duration
float64
0
2.81k
x_coordinate_first_fixation
float64
0
1.79k
y_coordinate_first_fixation
float64
0
818
amplitude_first_saccade
float64
0
34.9
is_answer_correct
int64
0
1
pronoun
stringclasses
16 values
label
int64
0
1
correct_antecedent
stringclasses
142 values
incorrect_antecedent
stringclasses
146 values
is_pronoun
int64
0
1
Еще
0
0
1
ad01
177
177
1
177
221.7
128
1.2
1
их
0
камни
особенности
0
Еще
0
0
1
ad03
217
217
1
217
226.4
129.2
7.33
1
их
0
камни
особенности
0
Еще
0
0
1
ad04
162
162
1
162
190.9
171
6.59
1
их
0
камни
особенности
0
Еще
0
0
1
ad05
479
479
2
192
182.3
124.6
1.11
1
их
0
камни
особенности
0
Еще
0
0
1
ad06
129
129
1
129
208.5
161.6
5.98
1
их
0
камни
особенности
0
Еще
0
0
1
ad07
199
199
1
199
229
172.4
5.73
1
их
0
камни
особенности
0
Еще
0
0
1
ad08
150
150
1
150
193.3
98.3
6.93
1
их
0
камни
особенности
0
Еще
0
0
1
ad09
162
162
1
162
227
158.1
6.46
1
их
0
камни
особенности
0
Еще
0
0
1
ad10
157
157
1
157
197.1
136.2
6.91
1
их
0
камни
особенности
0
Еще
0
0
1
ad11
139
139
1
139
187.5
165.4
6.65
1
их
0
камни
особенности
0
Еще
0
0
1
ad12
136
136
1
136
164.7
84.5
7.7
1
их
0
камни
особенности
0
Еще
0
0
1
ad13
194
194
1
194
245.7
105.5
7.45
1
их
0
камни
особенности
0
Еще
0
0
1
ad14
161
161
1
161
216.1
126.4
7.14
1
их
0
камни
особенности
0
Еще
0
0
1
ad15
0
0
0
0
0
0
0
1
их
0
камни
особенности
0
Еще
0
0
1
ad16
247
247
2
116
160.2
157.3
5.93
1
их
0
камни
особенности
0
Еще
0
0
1
ad17
235
235
1
235
238.8
103.6
7.67
1
их
0
камни
особенности
0
Еще
0
0
1
ad18
0
0
0
0
0
0
0
1
их
0
камни
особенности
0
Еще
0
0
1
ad19
554
554
2
232
156.6
130.7
6.75
1
их
0
камни
особенности
0
Еще
0
0
1
ad20
412
412
2
150
167.6
160.9
6.99
1
их
0
камни
особенности
0
Еще
0
0
1
ad21
146
146
1
146
188.1
144.9
5.71
1
их
0
камни
особенности
0
Еще
0
0
1
ad22
134
134
1
134
167.9
157.9
6.99
1
их
0
камни
особенности
0
Еще
0
0
1
ad23
130
130
1
130
197
145.2
1.37
1
их
0
камни
особенности
0
Еще
0
0
1
ad24
0
0
0
0
0
0
0
1
их
0
камни
особенности
0
Еще
0
0
1
ad25
104
104
1
104
247.8
123
1.23
1
их
0
камни
особенности
0
Еще
0
0
1
ad26
163
163
1
163
209.8
180.2
5.83
1
их
0
камни
особенности
0
Еще
0
0
1
ad27
302
214
2
214
252.7
162.9
7.06
1
их
0
камни
особенности
0
Еще
0
0
1
ad28
215
215
2
95
151.8
140.1
4.15
1
их
0
камни
особенности
0
Еще
0
0
1
ad29
307
307
2
170
207.8
153.4
4.27
1
их
0
камни
особенности
0
Еще
0
0
1
ad30
230
230
1
230
221.5
173.9
6.71
1
их
0
камни
особенности
0
Еще
0
0
1
ad31
140
140
1
140
200.3
135.9
0.31
1
их
0
камни
особенности
0
Еще
0
0
1
ad32
1,015
193
5
193
200.9
140.5
7.66
1
их
0
камни
особенности
0
Еще
0
0
1
ad33
148
148
1
148
266.6
97.7
7.02
1
их
0
камни
особенности
0
Еще
0
0
1
ad34
200
200
1
200
216.9
157.4
6.37
1
их
0
камни
особенности
0
Еще
0
0
1
ad35
469
186
3
186
203.7
124.4
6.4
1
их
0
камни
особенности
0
Еще
0
0
1
ad36
489
238
2
238
205.3
152.4
6.52
1
их
0
камни
особенности
0
Еще
0
0
1
ad37
398
158
2
158
250.6
146.7
7.12
1
их
0
камни
особенности
0
Еще
0
0
1
ad38
0
0
0
0
0
0
0
1
их
0
камни
особенности
0
Еще
0
0
1
ad39
183
183
1
183
218.9
150.5
7.05
1
их
0
камни
особенности
0
Еще
0
0
1
ad40
142
142
1
142
217.2
172.1
11.69
1
их
0
камни
особенности
0
Еще
0
0
1
ad41
173
173
1
173
259.4
126.7
0.63
1
их
0
камни
особенности
0
Еще
0
0
1
ad42
490
211
2
211
217.9
92.2
8.16
1
их
0
камни
особенности
0
Еще
0
0
1
ad43
192
192
1
192
206.3
135.1
6.76
1
их
0
камни
особенности
0
Еще
0
0
1
ad44
177
177
1
177
171
143.4
6.68
1
их
0
камни
особенности
0
Еще
0
0
1
ad45
160
160
1
160
171.4
100.5
6.67
1
их
0
камни
особенности
0
Еще
0
0
1
ad46
127
127
1
127
240.1
133.6
0.85
1
их
0
камни
особенности
0
Еще
0
0
1
ad47
192
192
1
192
241
131.7
7.55
1
их
0
камни
особенности
0
Еще
0
0
1
ad48
0
0
0
0
0
0
0
1
их
0
камни
особенности
0
Еще
0
0
1
ad49
264
150
2
150
211.6
124.5
0.86
1
их
0
камни
особенности
0
Еще
0
0
1
ad50
134
134
1
134
194.5
130.5
6.88
1
их
0
камни
особенности
0
Еще
150
0
1
ad100
163
163
1
163
320.9
127.8
7.35
1
их
1
камни
особенности
0
Еще
150
0
1
ad51
162
162
1
162
285.6
135.9
7.9
1
их
1
камни
особенности
0
Еще
150
0
1
ad52
304
304
2
132
229.1
123.4
7.37
1
их
1
камни
особенности
0
Еще
150
0
1
ad53
761
512
3
170
286
60.5
8.74
1
их
1
камни
особенности
0
Еще
150
0
1
ad54
0
0
0
0
0
0
0
1
их
1
камни
особенности
0
Еще
150
0
1
ad55
411
164
2
164
270.2
114.4
7.58
1
их
1
камни
особенности
0
Еще
150
0
1
ad56
0
0
0
0
0
0
0
1
их
1
камни
особенности
0
Еще
150
0
1
ad57
171
171
1
171
280.5
83.1
7.51
1
их
1
камни
особенности
0
Еще
150
0
1
ad58
197
197
1
197
258.1
142.3
6.84
1
их
1
камни
особенности
0
Еще
150
0
1
ad59
0
0
0
0
0
0
0
1
их
1
камни
особенности
0
Еще
150
0
1
ad60
0
0
0
0
0
0
0
1
их
1
камни
особенности
0
Еще
150
0
1
ad61
265
265
2
120
204.3
118.3
8.19
1
их
1
камни
особенности
0
Еще
150
0
1
ad62
162
162
1
162
248
118.8
7.11
1
их
1
камни
особенности
0
Еще
150
0
1
ad63
306
306
2
137
233.4
120.5
7.43
1
их
1
камни
особенности
0
Еще
150
0
1
ad64
258
258
2
62
202.3
111.1
7.6
1
их
1
камни
особенности
0
Еще
150
0
1
ad65
132
132
1
132
209.6
137.6
6.96
1
их
1
камни
особенности
0
Еще
150
0
1
ad66
444
444
2
176
289.3
127.3
0.52
1
их
1
камни
особенности
0
Еще
150
0
1
ad67
0
0
0
0
0
0
0
1
их
1
камни
особенности
0
Еще
150
0
1
ad68
155
155
1
155
287.3
103.8
7.28
1
их
1
камни
особенности
0
Еще
150
0
1
ad69
0
0
0
0
0
0
0
1
их
1
камни
особенности
0
Еще
150
0
1
ad70
189
189
1
189
269.6
116.7
2.7
1
их
1
камни
особенности
0
Еще
150
0
1
ad72
135
135
1
135
296.4
116.3
6.5
1
их
1
камни
особенности
0
Еще
150
0
1
ad73
170
170
1
170
209.2
125.9
7.37
1
их
1
камни
особенности
0
Еще
150
0
1
ad74
0
0
0
0
0
0
0
1
их
1
камни
особенности
0
Еще
150
0
1
ad75
147
147
1
147
251.2
139.8
6.86
1
их
1
камни
особенности
0
Еще
150
0
1
ad77
641
219
3
219
293.3
111.2
8.13
1
их
1
камни
особенности
0
Еще
150
0
1
ad78
138
138
1
138
234.6
143.4
7.9
1
их
1
камни
особенности
0
Еще
150
0
1
ad79
164
164
1
164
279.8
94.5
7.99
1
их
1
камни
особенности
0
Еще
150
0
1
ad80
112
112
1
112
266.5
135.7
8
1
их
1
камни
особенности
0
Еще
150
0
1
ad81
379
186
2
186
300.6
142.3
7.58
1
их
1
камни
особенности
0
Еще
150
0
1
ad82
271
118
2
118
202.1
134.7
6.79
1
их
1
камни
особенности
0
Еще
150
0
1
ad83
248
248
2
110
206.3
118.4
7.74
1
их
1
камни
особенности
0
Еще
150
0
1
ad84
217
217
2
136
251.7
143.9
6.9
1
их
1
камни
особенности
0
Еще
150
0
1
ad85
133
133
1
133
252.2
115
7.55
1
их
1
камни
особенности
0
Еще
150
0
1
ad86
241
241
2
118
201.8
111.9
6.62
1
их
1
камни
особенности
0
Еще
150
0
1
ad87
455
251
3
131
210.9
134.2
6.8
1
их
1
камни
особенности
0
Еще
150
0
1
ad88
303
303
2
158
203.8
85.7
6.76
1
их
1
камни
особенности
0
Еще
150
0
1
ad89
115
115
1
115
253.8
136.5
8.21
1
их
1
камни
особенности
0
Еще
150
0
1
ad90
189
189
2
117
267.3
143.1
6.34
0
их
1
камни
особенности
0
Еще
150
0
1
ad91
182
182
1
182
316.6
132
8.18
1
их
1
камни
особенности
0
Еще
150
0
1
ad92
205
205
1
205
245.9
136.3
1.75
1
их
1
камни
особенности
0
Еще
150
0
1
ad93
224
224
2
58
324.6
117.3
7.06
1
их
1
камни
особенности
0
Еще
150
0
1
ad94
141
141
1
141
262.9
108.9
4.01
1
их
1
камни
особенности
0
Еще
150
0
1
ad95
274
274
2
141
321.1
140.3
8.44
1
их
1
камни
особенности
0
Еще
150
0
1
ad96
293
293
1
293
324.2
133.8
2.39
1
их
1
камни
особенности
0
Еще
150
0
1
ad97
151
151
1
151
298.6
138.2
8.09
1
их
1
камни
особенности
0
Еще
150
0
1
ad98
119
119
1
119
297.3
146.8
0.71
1
их
1
камни
особенности
0
Еще
150
0
1
ad99
694
318
3
318
305.7
142.6
8.4
1
их
1
камни
особенности
0
один
0
0
2
ad01
621
56
3
56
281.6
169.4
6.53
1
их
0
камни
особенности
0
один
0
0
2
ad03
255
163
2
163
360.6
129.5
2.39
1
их
0
камни
особенности
0
один
0
0
2
ad04
379
379
1
379
286.2
157.7
2.13
1
их
0
камни
особенности
0

EyeWino

EyeWino is a new dataset based on the data from human eye-tracking for anaphora resolution.

Dataset Description

The Russian Winograd Schema Challenge dataset from TAPE (Taktasheva et al., 2022) was utilized for the anaphora resolution task to gather information on participants' eye movements.

The final dataset consists of 296 sentence-question pairs, which contain 9319 words and 148 unique sentences. The average number of participants per word is 48. The total number of observations for each variable is 448047.

Data Fields

  • word, a word in a sentence;
  • example_id, id of the example in the dataset;
  • text_id, id of the unique text in the dataset;
  • position_id, position of the word in the sentence;
  • annotator_id, experiment participant id;
  • is_answer_correct, the correctness of the experiment participant's answer;
  • reading_time, the sum of all fixation durations on the current word, ms;
  • gaze_duration, the sum of all fixation durations on the current word in the first-pass reading, ms;
  • fixations, the number of all fixations on the current word;
  • first_fixation_duration, the duration of the first fixation on the word, ms;
  • x_coordinate_first_fixation, the coordinate of the first fixation on the word along the x axis, where the screen is the coordinate plane;
  • y_coordinate_first_fixation, the coordinate of the first fixation on the word along the y axis, where the screen is the coordinate plane;
  • amplitude_first_saccade, the amplitude of the first saccade, deg;
  • correct_antecedent, the correct antecedent for example_id;
  • incorrect_antecedent, the incorrect antecedent for example_id;
  • pronoun, an anaphoric pronoun for example_id;
  • is_pronoun, an indicator of whether the word is the anaphoric pronoun;
  • label, an indicator of whether the question is about the correct antecedent.

Cite our ACL workshop paper https://aclanthology.org/2024.cmcl-1.10/:

@inproceedings{kozlova-etal-2024-transformer,
    title = "Transformer Attention vs Human Attention in Anaphora Resolution",
    author = "Kozlova, Anastasia  and
      Akhmetgareeva, Albina  and
      Khanova, Aigul  and
      Kudriavtsev, Semen  and
      Fenogenova, Alena",
    editor = "Kuribayashi, Tatsuki  and
      Rambelli, Giulia  and
      Takmaz, Ece  and
      Wicke, Philipp  and
      Oseki, Yohei",
    booktitle = "Proceedings of the Workshop on Cognitive Modeling and Computational Linguistics",
    month = aug,
    year = "2024",
    address = "Bangkok, Thailand",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2024.cmcl-1.10",
    pages = "109--122",
    abstract = "Motivated by human cognitive processes, attention mechanism within transformer architecture has been developed to assist neural networks in allocating focus to specific aspects within input data. Despite claims regarding the interpretability achieved by attention mechanisms, the extent of correlation and similarity between machine and human attention remains a subject requiring further investigation.In this paper, we conduct a quantitative analysis of human attention compared to neural attention mechanisms in the context of the anaphora resolution task. We collect an eye-tracking dataset based on the Winograd schema challenge task for the Russian language. Leveraging this dataset, we conduct an extensive analysis of the correlations between human and machine attention maps across various transformer architectures, network layers of pre-trained and fine-tuned models. Our aim is to investigate whether insights from human attention mechanisms can be used to enhance the performance of neural networks in tasks such as anaphora resolution. The results reveal distinctions in anaphora resolution processing, offering promising prospects for improving the performance of neural networks and understanding the cognitive nuances of human perception.",
}
Downloads last month
2
Edit dataset card