Datasets:

Tasks:
Other
Modalities:
Text
Formats:
parquet
Languages:
English
ArXiv:
Libraries:
Datasets
Dask
License:
VictorSanh commited on
Commit
c16ac20
1 Parent(s): efaffd3

First draft - data card

Browse files
Files changed (1) hide show
  1. README.md +228 -0
README.md ADDED
@@ -0,0 +1,228 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ YAML tags:
3
+ - copy-paste the tags obtained with the tagging app: https://github.com/huggingface/datasets-tagging
4
+ - WIP!
5
+ ---
6
+
7
+ # Dataset Card for P3
8
+
9
+ ## Table of Contents
10
+ - [Table of Contents](#table-of-contents)
11
+ - [Dataset Description](#dataset-description)
12
+ - [Dataset Summary](#dataset-summary)
13
+ - [Supported Tasks and Leaderboards](#supported-tasks-and-leaderboards)
14
+ - [Languages](#languages)
15
+ - [Dataset Structure](#dataset-structure)
16
+ - [Data Instances](#data-instances)
17
+ - [Data Fields](#data-fields)
18
+ - [Data Splits](#data-splits)
19
+ - [Dataset Creation](#dataset-creation)
20
+ - [Curation Rationale](#curation-rationale)
21
+ - [Source Data](#source-data)
22
+ - [Annotations](#annotations)
23
+ - [Personal and Sensitive Information](#personal-and-sensitive-information)
24
+ - [Considerations for Using the Data](#considerations-for-using-the-data)
25
+ - [Social Impact of Dataset](#social-impact-of-dataset)
26
+ - [Discussion of Biases](#discussion-of-biases)
27
+ - [Other Known Limitations](#other-known-limitations)
28
+ - [Additional Information](#additional-information)
29
+ - [Dataset Curators](#dataset-curators)
30
+ - [Licensing Information](#licensing-information)
31
+ - [Citation Information](#citation-information)
32
+ - [Contributions](#contributions)
33
+
34
+ ## Dataset Description
35
+
36
+ - **Homepage:** https://bigscience.huggingface.co/promptsource
37
+ - **Repository:** https://github.com/bigscience-workshop/promptsource/
38
+ - **Paper:** TBA
39
+ - **Point of Contact:** Victor Sanh (victor@huggingface.co)
40
+
41
+ ### Dataset Summary
42
+
43
+ P3 is a collection of prompted English datasets covering a diverse set of NLP tasks. A prompt is the combination of an input template and a target template. The templates are functions mapping a data example into natural language for the input and target sequences. For example, in the case of an NLI dataset, the data example would include fields for *Premise, Hypothesis, Label*. An input template would be *If {Premise} is true, is it also true that {Hypothesis}?*, whereas a target template can be defined with the label choices *Choices[label]*. Here *Choices* is prompt-specific metadata that consists of the options *yes, maybe, no* corresponding to *label* being entailment (0), neutral (1) or contradiction (2).
44
+
45
+ Prompts are collected using [Promptsource](https://github.com/bigscience-workshop/promptsource), an interface to interactively write prompts on datasets, and collect prompt-specific metadata such as evaluation metrics. As of October 13th, there are 2'000 prompts collected for 270+ data(sub)sets. The collection of prompts is publicly available on [Promptsource](https://github.com/bigscience-workshop/promptsource).
46
+
47
+ To train [T0*](https://huggingface.co/bigscience/T0pp_11B), we used a subset of the prompts available in Promptsource (see details [here](https://huggingface.co/bigscience/T0pp_11B#training-data)). However, some of the prompts use `random.choice`, a method that selects uniformly at random an option in a list of valid possibilities. For reproducibility purposes, we release the collection of prompted examples used to train T0*. **The data available here are the materialized version of the prompted datasets used in [Multi-task enables task zero-shot generalization](TODO) which represent only a subset datasets for which there is at least one prompt on Promptsource.**
48
+
49
+ ### Supported Tasks and Leaderboards
50
+
51
+ The tasks represented in P3 cover a diverse set of NLP tasks including multiple-choice QA, sentiment analysis or natural language inference. We detail the full list of datasets in [Source Data](#source-data).
52
+
53
+ ### Languages
54
+
55
+ The data in P3 are in English (BCP-47 `en`).
56
+
57
+ ## Dataset Structure
58
+
59
+ ### Data Instances
60
+
61
+ An example of "train" looks as follows:
62
+ ```bash
63
+ ```
64
+
65
+ To check all the prompted examples, you can use [Promptsource hosted tool](http://bigscience.huggingface.co/promptsource) and choose the `Prompted dataset viewer` mode in the left panel.
66
+
67
+
68
+ ### Data Fields
69
+
70
+ The data fields are the same among all splits:
71
+ - `input_text`: the natural language input fed to the model
72
+ - `target_text`: the natural language target that the model has to generate
73
+ - `tokenized_input`: the tokenized `input_text` with T5's tokenizer
74
+ - `tokenized_target`: the tokenized `target_text` with T5's tokenizer
75
+
76
+ ### Data Splits
77
+
78
+ |Data(sub)set|Split|Number of examples|
79
+ |-|-|-|
80
+
81
+ ## Dataset Creation
82
+
83
+ ### Curation Rationale
84
+
85
+ P3 relies on the Hugging Face Dataset library. Any public dataset in the Datasets library can be prompted. We select the datasets that have at least one subset in English and excluded datasets containing (predominantly) non-natural language examples.
86
+
87
+ We conservatively decided not to prompt datasets that contain potentially harmful content (for instance, datasets built on social media content). However, we sometimes prompt datasets that are purposefully built to measure bias and fairness of trained models, and reserve these prompted datasets (the validation or test sets) for evaluation purposes.
88
+
89
+ ### Source Data
90
+
91
+ Here's the full list of the datasets present in the materialized version of P3:
92
+ - Multiple-Choice QA
93
+ - CommonsenseQA
94
+ - Cosmos
95
+ - DREAM
96
+ - QASC
97
+ - QUAIL
98
+ - Quarrel
99
+ - QuaRTz
100
+ - SciQ
101
+ - Social IQA
102
+ - Wiki Hop
103
+ - WiQA
104
+ - ARC
105
+ - BoolQ
106
+ - Circa
107
+ - MC-TACO
108
+ - MultiRC
109
+ - OpenBookQA
110
+ - PIQA
111
+ - RACE
112
+ - Extractive QA
113
+ - Adversarial QA
114
+ - DuoRC
115
+ - Quoref
116
+ - ROPES
117
+ - TyDiQA
118
+ - CoQA
119
+ - DROP
120
+ - QA SRL
121
+ - QuAC
122
+ - ReCoRD
123
+ - SQuAD v2
124
+ - Close-book QA
125
+ - Hotpot QA
126
+ - Wiki QA
127
+ - NQ Open
128
+ - Trivia QA
129
+ - Web Questions
130
+ - Structure-to-text
131
+ - Common Gen
132
+ - Wiki Bio
133
+ - Sentiment
134
+ - Amazon
135
+ - App Reviews
136
+ - IMDB
137
+ - Rotten Tomatoes
138
+ - Yelp
139
+ - Summarization
140
+ - CNN Daily Mail
141
+ - Gigaword
142
+ - MultiNews
143
+ - SamSum
144
+ - XSum
145
+ - Topic Classification
146
+ - AG News
147
+ - DBPedia
148
+ - TREC
149
+ - Paraphrase Identification
150
+ - MRPC
151
+ - PAWS
152
+ - QQP
153
+ - Natural Language Inference
154
+ - ANLI
155
+ - CB
156
+ - RTE
157
+ - Coreference Resolution
158
+ - WSC
159
+ - Winogrande
160
+ - Word Sense disambiguation
161
+ - WiC
162
+ - Sentence Completion
163
+ - COPA
164
+ - HellaSwag
165
+ - Story Cloze
166
+
167
+ TODO: recheck this list to match Figure 2 in final version of paper
168
+
169
+ <!-- #### Initial Data Collection and Normalization
170
+
171
+
172
+ #### Who are the source language producers?
173
+
174
+ [More Information Needed] -->
175
+
176
+ ### Annotations
177
+
178
+ The prompts available in Promptsource are collected as part of BigScience, one-year long research workshop on large multilingual models and datasets. 36 contributors affiliated with 24 institutions in 8 countries participated to the prompt collection. Contributors are in majority machine learning researchers or machine learning engineers.
179
+
180
+ The main annotation guideline was that prompts needed to be grammatical and understandable by a native English speaker with no prior experience of the tasks. Additionally, prompts that required explicit counting or numerical indexing were removed in favor of natural language variants, e.g., instead of predicting indices of a span to extract (e.g. in extractive question answering), the model was expected to copy the span's text instead. With these minimal constraints, prompt writers were encouraged to use both formal and creative prompts and various orderings of the data. Most of the prompts correspond directly to a version of the original proposed task, although we also allowed prompts that permuted the original task (for instance, generating a document from its summary) or allowed for ambiguous output (for instance, not indicating a list of available choices).
181
+
182
+ The full annotation given to the contributors can be found [here](https://github.com/bigscience-workshop/promptsource/blob/main/CONTRIBUTING.md). *Note to self: the link is currently being updated with the)
183
+
184
+ <!-- #### Annotation process
185
+
186
+ [More Information Needed]
187
+
188
+ #### Who are the annotators?
189
+
190
+ [More Information Needed] -->
191
+
192
+ <!-- ### Personal and Sensitive Information
193
+
194
+ [More Information Needed]
195
+
196
+ ## Considerations for Using the Data
197
+
198
+ ### Social Impact of Dataset
199
+
200
+ [More Information Needed]
201
+
202
+ ### Discussion of Biases
203
+
204
+ [More Information Needed] -->
205
+
206
+ <!-- ### Other Known Limitations
207
+
208
+ [More Information Needed]
209
+
210
+ ## Additional Information
211
+
212
+ ### Dataset Curators
213
+
214
+ [More Information Needed] -->
215
+
216
+ ### Licensing Information
217
+
218
+ The dataset is released under Apache 2.0.
219
+
220
+ ### Citation Information
221
+
222
+ ```bibtex
223
+ WIP
224
+ ```
225
+
226
+ ### Contributions
227
+
228
+ Thanks to the contributors of [promptsource](https://github.com/bigscience-workshop/promptsource/graphs/contributors) for adding this dataset.