rasikabh commited on
Commit
3258132
1 Parent(s): b2afb0a

Create README.md

Browse files
Files changed (1) hide show
  1. README.md +141 -0
README.md ADDED
@@ -0,0 +1,141 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ annotations_creators:
3
+ - machine-generated
4
+ language:
5
+ - en
6
+ language_creators:
7
+ - found
8
+ license:
9
+ - mit
10
+ multilinguality:
11
+ - monolingual
12
+ pretty_name: pile-pii-scrubadub
13
+ size_categories:
14
+ - 1M<n<10M
15
+ source_datasets:
16
+ - extended|the_pile
17
+ tags:
18
+ - pii
19
+ - personally
20
+ - identifiable
21
+ - information
22
+ task_categories:
23
+ - text-classification
24
+ - other
25
+ task_ids:
26
+ - acceptability-classification
27
+ - text-scoring
28
+ ---
29
+
30
+ # Dataset Card for pile-pii-scrubadub
31
+
32
+ ## Dataset Description
33
+
34
+ - **Repository: https://github.com/tomekkorbak/aligned-pretraining-objectives**
35
+ - **Paper: Arxiv link to be added**
36
+
37
+ ### Dataset Summary
38
+
39
+ This dataset contains text from [The Pile](https://huggingface.co/datasets/the_pile), annotated based on the personally idenfitiable information (PII) in each sentence.
40
+ Each document (row in the dataset) is segmented into sentences, and each sentence is given a score: the percentage of words in it that are classified as PII by [Scrubadub](https://scrubadub.readthedocs.io/en/stable/).
41
+
42
+ ### Supported Tasks and Leaderboards
43
+
44
+ [More Information Needed]
45
+
46
+ ### Languages
47
+
48
+ This dataset is taken from [The Pile](https://huggingface.co/datasets/the_pile), which is English text.
49
+
50
+ ## Dataset Structure
51
+
52
+ ### Data Instances
53
+
54
+ 1949977
55
+
56
+ ### Data Fields
57
+
58
+ - texts (sequence): a list of the sentences in the document
59
+ - meta (dict): the section of [The Pile](https://huggingface.co/datasets/the_pile) from which it originated
60
+ - scores (sequence): a score for each sentence in the `texts` column indicating the percent of words that are detected as PII by [Scrubadub](https://scrubadub.readthedocs.io/en/stable/)
61
+ - avg_score (float64): the average of the scores listed in the `scores` column
62
+ - num_sents (int64): the number of sentences (and scores) in that document
63
+
64
+ ### Data Splits
65
+
66
+ Training set only
67
+
68
+ ## Dataset Creation
69
+
70
+ ### Curation Rationale
71
+
72
+ This is labeled text from [The Pile](https://huggingface.co/datasets/the_pile), a large dataset of text in English. The PII is labeled so that generative language models can be trained to avoid generating PII.
73
+
74
+ ### Source Data
75
+
76
+ #### Initial Data Collection and Normalization
77
+
78
+ This is labeled text from [The Pile](https://huggingface.co/datasets/the_pile).
79
+
80
+ #### Who are the source language producers?
81
+
82
+ Please see [The Pile](https://huggingface.co/datasets/the_pile) for the source of the dataset.
83
+
84
+ ### Annotations
85
+
86
+ #### Annotation process
87
+
88
+ For each sentence, [Scrubadub](https://scrubadub.readthedocs.io/en/stable/) was used to detect:
89
+
90
+ - dates of birth
91
+ - Skype usernames
92
+ - addresses
93
+ - anything detected by the SpacyEnityDetector (NER model) including:
94
+ - locations
95
+ - names
96
+ - dates
97
+ - email addresses
98
+ - usernames
99
+ - passwords
100
+
101
+ #### Who are the annotators?
102
+
103
+ [Scrubadub](https://scrubadub.readthedocs.io/en/stable/)
104
+
105
+ ### Personal and Sensitive Information
106
+
107
+ This dataset contains all PII that was originally contained in [The Pile](https://huggingface.co/datasets/the_pile), with all detected PII annotated.
108
+
109
+ ## Considerations for Using the Data
110
+
111
+ ### Social Impact of Dataset
112
+
113
+ This dataset contains examples of real PII (conveniently annotated in the text!). Please take care to avoid misusing it or putting anybody in danger by publicizing their information.
114
+ This dataset is intended for research purposes only. We cannot guarantee that all PII has been detected, and we cannot guarantee that models trained using it will avoid generating PII.
115
+ We do not recommend deploying models trained on this data.
116
+
117
+ ### Discussion of Biases
118
+
119
+ This dataset contains all biases from The Pile discussed in their paper: https://arxiv.org/abs/2101.00027
120
+
121
+ ### Other Known Limitations
122
+
123
+ The PII in this dataset was detected using imperfect automated detection methods. We cannot guarantee that the labels are 100% accurate.
124
+
125
+ ## Additional Information
126
+
127
+ ### Dataset Curators
128
+
129
+ [The Pile](https://huggingface.co/datasets/the_pile)
130
+
131
+ ### Licensing Information
132
+
133
+ From [The Pile](https://huggingface.co/datasets/the_pile): PubMed Central: [MIT License](https://github.com/EleutherAI/pile-pubmedcentral/blob/master/LICENSE)
134
+
135
+ ### Citation Information
136
+
137
+ Paper information to be added
138
+
139
+ ### Contributions
140
+
141
+ [The Pile](https://huggingface.co/datasets/the_pile)