Commit
·
281aadb
1
Parent(s):
a38439e
Upload README.md
Browse files
README.md
ADDED
@@ -0,0 +1,99 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
|
3 |
+
# Dataset Card for KAMEL: Knowledge Analysis with Multitoken Entities in Language Models
|
4 |
+
## Table of Contents
|
5 |
+
- [Dataset Description](#dataset-description)
|
6 |
+
- [Dataset Summary](#dataset-summary)
|
7 |
+
- [Supported Tasks and Leaderboards](#supported-tasks-and-leaderboards)
|
8 |
+
- [Languages](#languages)
|
9 |
+
- [Dataset Structure](#dataset-structure)
|
10 |
+
- [Data Instances](#data-instances)
|
11 |
+
- [Data Fields](#data-fields)
|
12 |
+
- [Data Splits](#data-splits)
|
13 |
+
- [Dataset Creation](#dataset-creation)
|
14 |
+
- [Curation Rationale](#curation-rationale)
|
15 |
+
- [Source Data](#source-data)
|
16 |
+
- [Annotations](#annotations)
|
17 |
+
- [Personal and Sensitive Information](#personal-and-sensitive-information)
|
18 |
+
- [Considerations for Using the Data](#considerations-for-using-the-data)
|
19 |
+
- [Social Impact of Dataset](#social-impact-of-dataset)
|
20 |
+
- [Discussion of Biases](#discussion-of-biases)
|
21 |
+
- [Other Known Limitations](#other-known-limitations)
|
22 |
+
- [Additional Information](#additional-information)
|
23 |
+
- [Dataset Curators](#dataset-curators)
|
24 |
+
- [Licensing Information](#licensing-information)
|
25 |
+
- [Citation Information](#citation-information)
|
26 |
+
- [Contributions](#contributions)
|
27 |
+
## Dataset Description
|
28 |
+
- **Homepage:**
|
29 |
+
https://github.com/JanKalo/KAMEL
|
30 |
+
- **Repository:**
|
31 |
+
https://github.com/JanKalo/KAMEL
|
32 |
+
- **Paper:**
|
33 |
+
@inproceedings{kalo2022kamel,
|
34 |
+
title={KAMEL: Knowledge Analysis with Multitoken Entities in Language Models},
|
35 |
+
author={Kalo, Jan-Christoph and Fichtel, Leandra},
|
36 |
+
booktitle={Automated Knowledge Base Construction},
|
37 |
+
year={2022}
|
38 |
+
}
|
39 |
+
### Dataset Summary
|
40 |
+
This dataset provides the data for KAMEL, a probing dataset for language models that contains factual knowledge
|
41 |
+
from Wikidata and Wikipedia.
|
42 |
+
|
43 |
+
See the paper for more details. For more information, also see:
|
44 |
+
https://github.com/JanKalo/KAMEL
|
45 |
+
### Languages
|
46 |
+
en
|
47 |
+
## Dataset Structure
|
48 |
+
### Data Instances
|
49 |
+
|
50 |
+
|
51 |
+
### Data Fields
|
52 |
+
KAMEL has the following fields:
|
53 |
+
* index: the id
|
54 |
+
* sub_label: a label for the subject
|
55 |
+
* obj_uri: Wikidata uri for the object
|
56 |
+
* obj_labels: multiple labels for the object
|
57 |
+
* chosen_label: the preferred label
|
58 |
+
* rel_uri: Wikidata uri for the relation
|
59 |
+
* rel_label: a label for the relation
|
60 |
+
|
61 |
+
### Data Splits
|
62 |
+
The dataset is split into a training, validation, and test dataset.
|
63 |
+
It contains 234 Wikidata relations.
|
64 |
+
For each relation there exist 200 training, 100 validation,
|
65 |
+
and 100 test instances.
|
66 |
+
|
67 |
+
## Dataset Creation
|
68 |
+
### Curation Rationale
|
69 |
+
This dataset was gathered and created to explore what knowledge graph facts are memorized by large language models.
|
70 |
+
### Source Data
|
71 |
+
#### Initial Data Collection and Normalization
|
72 |
+
See the reaserch paper and website for more detail. The dataset was
|
73 |
+
created from Wikidata and Wikipedia.
|
74 |
+
### Annotations
|
75 |
+
#### Annotation process
|
76 |
+
There is no human annotation, but only automatic linking from Wikidata facts to Wikipedia articles.
|
77 |
+
The details about the process can be found in the paper.
|
78 |
+
#### Who are the annotators?
|
79 |
+
Machine Annotations
|
80 |
+
### Personal and Sensitive Information
|
81 |
+
Unkown, but likely information about famous people mentioned in the English Wikipedia.
|
82 |
+
## Considerations for Using the Data
|
83 |
+
### Social Impact of Dataset
|
84 |
+
The goal for the work is to probe the understanding of language models.
|
85 |
+
### Discussion of Biases
|
86 |
+
Since the data is created from Wikipedia and Wikidata, the existing biases from these two data sources may also be reflected in KAMEL.
|
87 |
+
## Additional Information
|
88 |
+
### Dataset Curators
|
89 |
+
The authors of KAMEL at Vrije Universiteit Amsterdam and Technische Universität Braunschweig.
|
90 |
+
### Licensing Information
|
91 |
+
The Creative Commons Attribution-Noncommercial 4.0 International License. see https://github.com/facebookresearch/LAMA/blob/master/LICENSE
|
92 |
+
### Citation Information
|
93 |
+
@inproceedings{kalo2022kamel,
|
94 |
+
title={KAMEL: Knowledge Analysis with Multitoken Entities in Language Models},
|
95 |
+
author={Kalo, Jan-Christoph and Fichtel, Leandra},
|
96 |
+
booktitle={Automated Knowledge Base Construction},
|
97 |
+
year={2022}
|
98 |
+
}
|
99 |
+
|