tokenization_robustness_completion
Collection
8 items
•
Updated
Datasets:
Dataset Viewer
question
string | choices
list | answer
int64 | answer_label
string | split
string | subcategories
string | lang
string | second_lang
string | notes
string | id
string | set_id
float64 | variation_id
float64 |
|---|---|---|---|---|---|---|---|---|---|---|---|
Dr Smith is a doctor. Occupation of Dr Smith is
|
[
"teach",
"judge",
"doc",
"law"
] | 2 |
C
|
test
|
Abbreviations
|
eng_Latn
|
300-0.2
| 300 | 0.2 |
||
Dr Smith is an MD. Occipation of Dr Smith is
|
[
"teacher",
"doctor",
"judge",
"lawyer"
] | 1 |
B
|
test
|
Abbreviations
|
eng_Latn
|
300-0.37
| 300 | 0.37 |
||
The # of continents on Earth is
|
[
"5",
"6",
"8",
"7"
] | 3 |
D
|
test
|
Abbreviations
|
eng_Latn
|
304-0.2
| 304 | 0.2 |
||
The capital city of IR is
|
[
"Tehran",
"Mashhad",
"Baghdad",
"Isfahan"
] | 0 |
A
|
test
|
Abbreviations
|
eng_Latn
|
305-0.2
| 305 | 0.2 |
||
The # of days in a week is
|
[
"5",
"6",
"8",
"7"
] | 3 |
D
|
test
|
Abbreviations
|
eng_Latn
|
306-0.3
| 306 | 0.3 |
||
The # of hours in a day is
|
[
"20",
"24",
"25",
"30"
] | 1 |
B
|
test
|
Abbreviations
|
eng_Latn
|
307-0.3
| 307 | 0.3 |
||
The # of legs a cow has is
|
[
"4",
"8",
"3",
"5"
] | 0 |
A
|
test
|
Abbreviations
|
eng_Latn
|
308-0.3
| 308 | 0.3 |
||
The # of minutes in 2 hours is
|
[
"100",
"120",
"140",
"90"
] | 1 |
B
|
test
|
Abbreviations
|
eng_Latn
|
309-0.3
| 309 | 0.3 |
||
The # of months in a year is
|
[
"10",
"12",
"11",
"13"
] | 1 |
B
|
test
|
Abbreviations
|
eng_Latn
|
310-0.3
| 310 | 0.3 |
||
The # of seconds in a minute is
|
[
"50",
"100",
"30",
"60"
] | 3 |
D
|
test
|
Abbreviations
|
eng_Latn
|
311-0.3
| 311 | 0.3 |
||
The # of sides a hexagon has is
|
[
"5",
"7",
"6",
"8"
] | 2 |
C
|
test
|
Abbreviations
|
eng_Latn
|
312-0.3
| 312 | 0.3 |
||
The # of sides a triangle has is
|
[
"2",
"3",
"4",
"5"
] | 1 |
B
|
test
|
Abbreviations
|
eng_Latn
|
313-0.3
| 313 | 0.3 |
||
In "I work at Apple", Apple is a
|
[
"pers.",
"cty.",
"fr.",
"co."
] | 3 |
D
|
test
|
Abbreviations
|
eng_Latn
|
314-0.1
| 314 | 0.1 |
||
In "I work at Google", Google is a
|
[
"co.",
"pers.",
"cty.",
"fr."
] | 0 |
A
|
test
|
Abbreviations
|
eng_Latn
|
315-0.1
| 315 | 0.1 |
||
In "Microsoft released a new update", Microsoft is a
|
[
"pers.",
"cty.",
"fr.",
"co."
] | 3 |
D
|
test
|
Abbreviations
|
eng_Latn
|
316-0.1
| 316 | 0.1 |
||
In "The cat sat on the mat", the subj. is
|
[
"sat",
"the mat",
"the cat",
"on"
] | 2 |
C
|
test
|
Abbreviations
|
eng_Latn
|
317-0.1
| 317 | 0.1 |
||
The gas humans need to breathe to live is
|
[
"CH₄",
"O₂",
"He",
"H₂"
] | 1 |
B
|
test
|
Abbreviations
|
eng_Latn
|
322-0.1
| 322 | 0.1 |
||
Chad's cap. is
|
[
"Moundou",
"N'Djamena",
"Abéché",
"Ngama"
] | 1 |
B
|
test
|
Abbreviations
|
eng_Latn
|
326-0.2
| 326 | 0.2 |
||
The capital of FR is
|
[
"London",
"Berlin",
"Rome",
"Paris"
] | 3 |
D
|
test
|
Abbreviations
|
eng_Latn
|
327-0.2
| 327 | 0.2 |
||
The capital of JP is
|
[
"Kyoto",
"Tokyo",
"Osaka",
"Hiroshima"
] | 1 |
B
|
test
|
Abbreviations
|
eng_Latn
|
328-0.2
| 328 | 0.2 |
||
The capital of TR is
|
[
"Ankara",
"İstanbul",
"İzmir",
"Bursa"
] | 0 |
A
|
test
|
Abbreviations
|
eng_Latn
|
329-0.2
| 329 | 0.2 |
||
The chem. formula for water is
|
[
"CO2",
"NaCl",
"O2",
"H2O"
] | 3 |
D
|
test
|
Abbreviations
|
eng_Latn
|
330-0.20
| 330 | 0.2 |
||
The intent in "What time does the store close?" is
|
[
"purch",
"book",
"info",
"complain"
] | 2 |
C
|
test
|
Abbreviations
|
eng_Latn
|
331-0.1
| 331 | 0.1 |
||
The largest mammal in the world is
|
[
"dolphin",
"giraffe",
"bear",
"blue whale"
] | 3 |
D
|
test
|
Abbreviations
|
eng_Latn
|
332-0.1
| 332 | 0.1 |
||
The unit of measurement for temperature in the International System is
|
[
"°C",
"°F",
"°R",
"K"
] | 3 |
D
|
test
|
Abbreviations
|
eng_Latn
|
333-0.1
| 333 | 0.1 |
||
The country whose space agency is NASA is
|
[
"RU",
"CN",
"JP",
"US"
] | 3 |
D
|
test
|
Abbreviations
|
eng_Latn
|
334-0.1
| 334 | 0.1 |
||
The lang. spoken in Brazil is
|
[
"Portuguese",
"Spanish",
"French",
"Italian"
] | 0 |
A
|
test
|
Abbreviations
|
eng_Latn
|
335-0.2
| 335 | 0.2 |
||
The metal with chemical sym. 'Fe' is
|
[
"lead",
"zinc",
"gold",
"iron"
] | 3 |
D
|
test
|
Abbreviations
|
eng_Latn
|
336-0.1
| 336 | 0.1 |
||
The planet closest to the Sun in our solar system is
|
[
"♀",
"☿",
"♂",
"♁"
] | 1 |
B
|
test
|
Abbreviations
|
eng_Latn
|
338-0.1
| 338 | 0.1 |
||
The largest planet in the Solar System is
|
[
"♁",
"♄",
"♂",
"♃"
] | 3 |
D
|
test
|
Abbreviations
|
eng_Latn
|
339-0.1
| 339 | 0.1 |
||
The process that allows plants to produce their own food using sunlight is
|
[
"photo.",
"resp.",
"dig.",
"ferm."
] | 0 |
A
|
test
|
Abbreviations
|
eng_Latn
|
340-0.1
| 340 | 0.1 |
||
The author who wrote the play "Romeo and Juliet" is
|
[
"C. Dickens",
"W. Shakespeare",
"M. Twain",
"J. Austen"
] | 1 |
B
|
test
|
Abbreviations
|
eng_Latn
|
341-0.1
| 341 | 0.1 |
||
What plants need from the air to make food is
|
[
"CO₂",
"N₂",
"H₂",
"He"
] | 0 |
A
|
test
|
Abbreviations
|
eng_Latn
|
343-0.1
| 343 | 0.1 |
||
In "Can you pls. book a flight to Paris?", the person wants to
|
[
"go shopping",
"file a complaint",
"cancel reservation",
"make a booking"
] | 3 |
D
|
test
|
Abbreviations
|
eng_Latn
|
344-0.11
| 344 | 0.11 |
Dataset Card for Tokenization Robustness
A comprehensive evaluation dataset for testing robustness of different tokenization strategies.
Dataset Details
Dataset Description
This dataset evaluates how robust language models are to different tokenization strategies and edge cases. It includes text completion questions with multiple choice answers designed to test various aspects of tokenization handling.
- Curated by: R3
- Funded by [optional]: [More Information Needed]
- Shared by [optional]: [More Information Needed]
- Language(s) (NLP): [More Information Needed]
- License: cc
Dataset Sources [optional]
- Repository: [More Information Needed]
- Paper [optional]: [More Information Needed]
- Demo [optional]: [More Information Needed]
Uses
Direct Use
[More Information Needed]
Out-of-Scope Use
[More Information Needed]
Dataset Structure
The dataset contains multiple-choice questions with associated metadata about tokenization types and categories.
Dataset Creation
Curation Rationale
[More Information Needed]
Source Data
Data Collection and Processing
[More Information Needed]
Who are the source data producers?
[More Information Needed]
Annotations [optional]
Annotation process
[More Information Needed]
Who are the annotators?
[More Information Needed]
Personal and Sensitive Information
[More Information Needed]
Bias, Risks, and Limitations
The dataset focuses primarily on English text and may not generalize to other languages or tokenization schemes not covered in the evaluation.
Recommendations
Users should be made aware of the risks, biases and limitations of the dataset. More information needed for further recommendations.
Citation [optional]
BibTeX:
[More Information Needed]
APA:
[More Information Needed]
Glossary [optional]
[More Information Needed]
More Information [optional]
[More Information Needed]
Dataset Card Authors [optional]
[More Information Needed]
Dataset Card Contact
[More Information Needed]
- Downloads last month
- 76
Size of downloaded dataset files:
376 kB
Size of the auto-converted Parquet files:
376 kB
Number of rows:
1,214