Hugging Face's logo

Datasets:
Lidor-Mashiach
/
snli-training-paraphrase-augmentation

Modalities:
Tabular
Text
Formats:
csv
Languages:
English
Size:
1M - 10M
Tags:
snli
paraphrase
training-augmentation
natural-language-inference
phase-b
Libraries:
Datasets
pandas
Polars
License:
Dataset card Data Studio Files
xet
 Community
Dataset Viewer
Auto-converted to Parquet Duplicate
pair_id
stringlengths
18
18
premise
stringlengths
7
402
hypothesis
stringlengths
4
295
⌀
paraphrase
stringlengths
4
337
label
int64
0
2
para_idx
int64
0
4
SNLI-train-0000001
A person on a horse jumps over a broken down airplane.
A person is at a diner, ordering an omelette.
omelette at a restaurant, who is going to order .
2
0
SNLI-train-0000001
A person on a horse jumps over a broken down airplane.
A person is at a diner, ordering an omelette.
Ordering an omelette at a restaurant.
2
1
SNLI-train-0000001
A person on a horse jumps over a broken down airplane.
A person is at a diner, ordering an omelette.
An omelette order is being placed at a restaurant.
2
2
SNLI-train-0000001
A person on a horse jumps over a broken down airplane.
A person is at a diner, ordering an omelette.
A diner serves omelette to someone.
2
3
SNLI-train-0000001
A person on a horse jumps over a broken down airplane.
A person is at a diner, ordering an omelette.
The person visiting from a restaurant is ordering an omelette.
2
4
SNLI-train-0000002
A person on a horse jumps over a broken down airplane.
A person is outdoors, on a horse.
Someone is standing on a horse in the open air.
0
0
SNLI-train-0000002
A person on a horse jumps over a broken down airplane.
A person is outdoors, on a horse.
A horse is utilized by someone to ride outside.
0
1
SNLI-train-0000002
A person on a horse jumps over a broken down airplane.
A person is outdoors, on a horse.
Someone is seated on a horse in the open air.
0
2
SNLI-train-0000002
A person on a horse jumps over a broken down airplane.
A person is outdoors, on a horse.
One on a horse, out on the .
0
3
SNLI-train-0000002
A person on a horse jumps over a broken down airplane.
A person is outdoors, on a horse.
A person can be seen outside on a horse.
0
4
SNLI-train-0000003
Children smiling and waving at camera
They are smiling at their parents
They are beaming at their parents.
1
0
SNLI-train-0000003
Children smiling and waving at camera
They are smiling at their parents
They are happy looking at their parents
1
1
SNLI-train-0000003
Children smiling and waving at camera
They are smiling at their parents
They look at their parents with happiness.
1
2
SNLI-train-0000003
Children smiling and waving at camera
They are smiling at their parents
They look at their parents in a grin.
1
3
SNLI-train-0000003
Children smiling and waving at camera
They are smiling at their parents
They are exchanging a smile with their parents.
1
4
SNLI-train-0000004
Children smiling and waving at camera
There are children present
Those who are present are kids.
0
0
SNLI-train-0000004
Children smiling and waving at camera
There are children present
Some minors are present.
0
1
SNLI-train-0000004
Children smiling and waving at camera
There are children present
The participants consist of children.
0
2
SNLI-train-0000004
Children smiling and waving at camera
There are children present
The present consists of children.
0
3
SNLI-train-0000004
Children smiling and waving at camera
There are children present
Kids are present.
0
4
SNLI-train-0000005
Children smiling and waving at camera
The kids are frowning
The kids are with a frown.
2
0
SNLI-train-0000005
Children smiling and waving at camera
The kids are frowning
This is children, but they're frowning.
2
1
SNLI-train-0000005
Children smiling and waving at camera
The kids are frowning
The youngsters are showing a frown.
2
2
SNLI-train-0000005
Children smiling and waving at camera
The kids are frowning
These children are looking grimacing.
2
3
SNLI-train-0000006
A boy is jumping on skateboard in the middle of a red bridge.
The boy skates down the sidewalk.
The boy rides down the sidewalk while skating.
2
0
SNLI-train-0000006
A boy is jumping on skateboard in the middle of a red bridge.
The boy skates down the sidewalk.
Skating: The boy is seen sliding along the pavement.
2
1
SNLI-train-0000006
A boy is jumping on skateboard in the middle of a red bridge.
The boy skates down the sidewalk.
The boy moves on the sidewalk, skating.
2
2
SNLI-train-0000006
A boy is jumping on skateboard in the middle of a red bridge.
The boy skates down the sidewalk.
The boy moves along the sidewalk on skates.
2
3
SNLI-train-0000006
A boy is jumping on skateboard in the middle of a red bridge.
The boy skates down the sidewalk.
Skating down the sidewalk is where the boy goes.
2
4
SNLI-train-0000008
A boy is jumping on skateboard in the middle of a red bridge.
The boy is wearing safety equipment.
The boy is outfitted with safety devices.
1
0
SNLI-train-0000008
A boy is jumping on skateboard in the middle of a red bridge.
The boy is wearing safety equipment.
The boy is equipped with safety devices.
1
1
SNLI-train-0000008
A boy is jumping on skateboard in the middle of a red bridge.
The boy is wearing safety equipment.
The boy is donning protective gear.
1
2
SNLI-train-0000008
A boy is jumping on skateboard in the middle of a red bridge.
The boy is wearing safety equipment.
A safety kit is worn by the boy.
1
3
SNLI-train-0000008
A boy is jumping on skateboard in the middle of a red bridge.
The boy is wearing safety equipment.
Safety gear is being utilized by the boy.
1
4
SNLI-train-0000010
An older man sits with his orange juice at a small table in a coffee shop while employees in bright colored shirts smile in the background.
A boy flips a burger.
Burger: A boy fries a burger .
2
0
SNLI-train-0000010
An older man sits with his orange juice at a small table in a coffee shop while employees in bright colored shirts smile in the background.
A boy flips a burger.
Boys make burger.
2
1
SNLI-train-0000010
An older man sits with his orange juice at a small table in a coffee shop while employees in bright colored shirts smile in the background.
A boy flips a burger.
Burger guy – a boy flips .
2
2
SNLI-train-0000010
An older man sits with his orange juice at a small table in a coffee shop while employees in bright colored shirts smile in the background.
A boy flips a burger.
A young heave fries.
2
3
SNLI-train-0000010
An older man sits with his orange juice at a small table in a coffee shop while employees in bright colored shirts smile in the background.
A boy flips a burger.
This boy burger .
2
4
SNLI-train-0000014
Two blond women are hugging one another.
There are women showing affection.
Some women are affectionate.
0
0
SNLI-train-0000014
Two blond women are hugging one another.
There are women showing affection.
Certain women demonstrate affection.
0
1
SNLI-train-0000014
Two blond women are hugging one another.
There are women showing affection.
Some women exhibit love.
0
2
SNLI-train-0000014
Two blond women are hugging one another.
There are women showing affection.
Certain females show affection.
0
3
SNLI-train-0000014
Two blond women are hugging one another.
There are women showing affection.
Some women show affection.
0
4
SNLI-train-0000015
A few people in a restaurant setting, one of them is drinking orange juice.
The people are eating omelettes.
omelettes are being eaten by .
1
0
SNLI-train-0000015
A few people in a restaurant setting, one of them is drinking orange juice.
The people are eating omelettes.
The crowd is enjoying omelettes.
1
1
SNLI-train-0000015
A few people in a restaurant setting, one of them is drinking orange juice.
The people are eating omelettes.
The group is consuming omelettes.
1
2
SNLI-train-0000015
A few people in a restaurant setting, one of them is drinking orange juice.
The people are eating omelettes.
The individuals are indulging in omelettes.
1
3
SNLI-train-0000015
A few people in a restaurant setting, one of them is drinking orange juice.
The people are eating omelettes.
They're also eating omelettes.
1
4
SNLI-train-0000017
A few people in a restaurant setting, one of them is drinking orange juice.
The diners are at a restaurant.
The visitors are dining at a food establishment.
0
0
SNLI-train-0000017
A few people in a restaurant setting, one of them is drinking orange juice.
The diners are at a restaurant.
The individuals are eating at a restaurant.
0
1
SNLI-train-0000017
A few people in a restaurant setting, one of them is drinking orange juice.
The diners are at a restaurant.
These are diners at a food place .
0
2
SNLI-train-0000017
A few people in a restaurant setting, one of them is drinking orange juice.
The diners are at a restaurant.
The eaters are situated in a restaurant.
0
3
SNLI-train-0000017
A few people in a restaurant setting, one of them is drinking orange juice.
The diners are at a restaurant.
The diners are dining in a food joint.
0
4
SNLI-train-0000018
An older man is drinking orange juice at a restaurant.
A man is drinking juice.
A male individual is indulging in juice.
0
0
SNLI-train-0000018
An older man is drinking orange juice at a restaurant.
A man is drinking juice.
A male individual is ingesting juice.
0
1
SNLI-train-0000018
An older man is drinking orange juice at a restaurant.
A man is drinking juice.
A man ingests juice.
0
2
SNLI-train-0000018
An older man is drinking orange juice at a restaurant.
A man is drinking juice.
The male individual consumes juice.
0
3
SNLI-train-0000018
An older man is drinking orange juice at a restaurant.
A man is drinking juice.
A man is taking in juice .
0
4
SNLI-train-0000019
An older man is drinking orange juice at a restaurant.
Two women are at a restaurant drinking wine.
Wine consumption is observed by two women dining together at a table.
2
0
SNLI-train-0000019
An older man is drinking orange juice at a restaurant.
Two women are at a restaurant drinking wine.
Two females indulge in wine at a restaurant.
2
1
SNLI-train-0000019
An older man is drinking orange juice at a restaurant.
Two women are at a restaurant drinking wine.
A pair of women are spotted drinking wine at a dining establishment.
2
2
SNLI-train-0000019
An older man is drinking orange juice at a restaurant.
Two women are at a restaurant drinking wine.
Two ladies indulge in wine at a dining establishment.
2
3
SNLI-train-0000019
An older man is drinking orange juice at a restaurant.
Two women are at a restaurant drinking wine.
Two lady sit in a restaurant drinking wine .
2
4
SNLI-train-0000020
An older man is drinking orange juice at a restaurant.
A man in a restaurant is waiting for his meal to arrive.
Waiting: A man in a restaurant, waiting for his meal.
1
0
SNLI-train-0000020
An older man is drinking orange juice at a restaurant.
A man in a restaurant is waiting for his meal to arrive.
Waiting to be served: A man waiting to eat in a restaurant .
1
1
SNLI-train-0000020
An older man is drinking orange juice at a restaurant.
A man in a restaurant is waiting for his meal to arrive.
Waiting in the restaurant: A man is waiting for his food to arrive.
1
2
SNLI-train-0000020
An older man is drinking orange juice at a restaurant.
A man in a restaurant is waiting for his meal to arrive.
A man is waiting in a restaurant, waiting for his meal.
1
3
SNLI-train-0000020
An older man is drinking orange juice at a restaurant.
A man in a restaurant is waiting for his meal to arrive.
Waiting to be served: A man waits at a restaurant to enjoy his meal .
1
4
SNLI-train-0000021
A man with blond-hair, and a brown shirt drinking out of a public water fountain.
A blond man getting a drink of water from a fountain in the park.
A blond man drinking water from a fountain located in the park.
1
0
SNLI-train-0000021
A man with blond-hair, and a brown shirt drinking out of a public water fountain.
A blond man getting a drink of water from a fountain in the park.
The blond man is seen drinking water from a fountain situated in the park.
1
1
SNLI-train-0000021
A man with blond-hair, and a brown shirt drinking out of a public water fountain.
A blond man getting a drink of water from a fountain in the park.
A man with blonde hair drinking from a water fountain located in the park.
1
2
SNLI-train-0000021
A man with blond-hair, and a brown shirt drinking out of a public water fountain.
A blond man getting a drink of water from a fountain in the park.
A fountain in the park is being drank from by a blond man.
1
3
SNLI-train-0000021
A man with blond-hair, and a brown shirt drinking out of a public water fountain.
A blond man getting a drink of water from a fountain in the park.
A water fountain in the park is being sipped by a blond man.
1
4
SNLI-train-0000024
Two women who just had lunch hugging and saying goodbye.
The friends scowl at each other over a full dinner table.
The companions grimace at each other while feasting at a nearby dinner table.
2
0
SNLI-train-0000024
Two women who just had lunch hugging and saying goodbye.
The friends scowl at each other over a full dinner table.
A full dinner table is shared by the friends who scowl at each other.
2
1
SNLI-train-0000024
Two women who just had lunch hugging and saying goodbye.
The friends scowl at each other over a full dinner table.
The circle of friends grimaces at a sprawling dinner table.
2
2
SNLI-train-0000024
Two women who just had lunch hugging and saying goodbye.
The friends scowl at each other over a full dinner table.
The social circle grimace at a delectable dinner table.
2
3
SNLI-train-0000024
Two women who just had lunch hugging and saying goodbye.
The friends scowl at each other over a full dinner table.
The acquaintances grunt at each other around a crowded dinner table.
2
4
SNLI-train-0000025
Two women who just had lunch hugging and saying goodbye.
There are two woman in this picture.
This picture is accompanied by two females.
0
0
SNLI-train-0000025
Two women who just had lunch hugging and saying goodbye.
There are two woman in this picture.
The photograph showcases two females.
0
1
SNLI-train-0000025
Two women who just had lunch hugging and saying goodbye.
There are two woman in this picture.
The photograph depicts two females.
0
2
SNLI-train-0000025
Two women who just had lunch hugging and saying goodbye.
There are two woman in this picture.
Two women are portrayed in this picture.
0
3
SNLI-train-0000025
Two women who just had lunch hugging and saying goodbye.
There are two woman in this picture.
This picture shows two females.
0
4
SNLI-train-0000026
Two women who just had lunch hugging and saying goodbye.
The friends have just met for the first time in 20 years, and have had a great time catching up.
It took 20 years, but the friends have now reunited and enjoyed a fantastic time.
1
0
SNLI-train-0000026
Two women who just had lunch hugging and saying goodbye.
The friends have just met for the first time in 20 years, and have had a great time catching up.
The companions have finally met after 20 years, and they've enjoyed the moment.
1
1
SNLI-train-0000026
Two women who just had lunch hugging and saying goodbye.
The friends have just met for the first time in 20 years, and have had a great time catching up.
This is the first time in two decades that the friends have crossed paths and enjoyed hanging out.
1
2
SNLI-train-0000026
Two women who just had lunch hugging and saying goodbye.
The friends have just met for the first time in 20 years, and have had a great time catching up.
Following their first meeting in 20 years, the friends had a great time catching up with one another.
1
3
SNLI-train-0000026
Two women who just had lunch hugging and saying goodbye.
The friends have just met for the first time in 20 years, and have had a great time catching up.
They had already hung out and met some friends after 20 years, and they had such fun together.
1
4
SNLI-train-0000027
Two women, holding food carryout containers, hug.
The two sisters saw each other across the crowded diner and shared a hug, both clutching their doggie bags.
Across the bustling dining hall, the two sisters embraced and carried their bags of dogs together.
1
0
SNLI-train-0000027
Two women, holding food carryout containers, hug.
The two sisters saw each other across the crowded diner and shared a hug, both clutching their doggie bags.
Across the busy diner, the two sisters hugged each other and hold their doggies.
1
1
SNLI-train-0000027
Two women, holding food carryout containers, hug.
The two sisters saw each other across the crowded diner and shared a hug, both clutching their doggie bags.
The two sisters looked at each other from across the crowded diner and hugged each one, both carrying their doggie bags.
1
2
SNLI-train-0000027
Two women, holding food carryout containers, hug.
The two sisters saw each other across the crowded diner and shared a hug, both clutching their doggie bags.
After moving through the crowded diner, the two sisters hugged each other and carried their doggie bags.
1
3
SNLI-train-0000027
Two women, holding food carryout containers, hug.
The two sisters saw each other across the crowded diner and shared a hug, both clutching their doggie bags.
Entering the crowded diner, the two sisters looked together and hugged, clutching their bags of canines.
1
4
SNLI-train-0000028
Two women, holding food carryout containers, hug.
Two groups of rival gang members flipped each other off.
Two competing groups of gang members flipped on one another.
2
0
SNLI-train-0000030
A Little League team tries to catch a runner sliding into a base in an afternoon game.
A team is trying to score the games winning out.
One team is striving to score in the games that are won.
1
0
SNLI-train-0000030
A Little League team tries to catch a runner sliding into a base in an afternoon game.
A team is trying to score the games winning out.
The teams are striving to score in the games that they are winning.
1
1
SNLI-train-0000030
A Little League team tries to catch a runner sliding into a base in an afternoon game.
A team is trying to score the games winning out.
A team is scoring in order to secure wins.
1
2
SNLI-train-0000030
A Little League team tries to catch a runner sliding into a base in an afternoon game.
A team is trying to score the games winning out.
The team aiming to score wins games.
1
3
SNLI-train-0000030
A Little League team tries to catch a runner sliding into a base in an afternoon game.
A team is trying to score the games winning out.
The team aiming for goals is leading the losing team.
1
4
End of preview. Expand in Data Studio

SNLI Training Paraphrase Augmentation

Purpose

This dataset contains new paraphrases created for training paraphrase augmentation during Phase B of the research project.

It was not used as an NLI evaluation set.

It was not used for paraphrase consistency evaluation.

It is separate from the published SNLI Paraphrase Bank used for evaluation. The generation records report zero collisions with that evaluation bank.

The CSV contains only the new augmentation rows. It does not contain the full original SNLI training split. To create an augmented training set, combine these rows with the original SNLI training data.

Files

File Purpose
snli-training-paraphrase-augmentation.csv New paraphrases created for Phase B training
snli-training-paraphrase-augmentation-stats.json Merged generation statistics and experimental budgets
README.md Dataset documentation

Columns

Column Meaning
pair_id Identifier of the original SNLI training pair
premise Original premise
hypothesis Original hypothesis
paraphrase New hypothesis paraphrase created for augmentation
label Original NLI label
para_idx Index of the new paraphrase for the same hypothesis

How the data was created

The augmentation rows were generated in three parallel parts and then merged into one CSV.

Candidate paraphrases were generated with humarin/chatgpt_paraphraser_on_T5_base.

Candidates were checked with MoritzLaurer/DeBERTa-v3-large-mnli-fever-anli-ling-wanli.

A candidate was kept only when it passed the following checks:

  1. Its length stayed within the allowed ratio relative to the original hypothesis.
  2. The candidate and original hypothesis were predicted to entail one another.
  3. The NLI relation between the premise and candidate matched the original gold label.
  4. The candidate was not a duplicate.
  5. The candidate did not appear in the published evaluation paraphrase bank.

The target was up to five new paraphrases per requested hypothesis.

Statistics

Measure Value
Hypotheses requested 467,786
Hypotheses with at least one new paraphrase 463,481
Hypotheses with no new paraphrase 4,305
Hypotheses reaching five new paraphrases 428,336
New paraphrase rows 2,235,245
Parallel generation parts 3
Collisions with the evaluation paraphrase bank 0

The merged rejection counts, duplicate checks, and experimental budgets appear in snli-training-paraphrase-augmentation-stats.json.

Phase B context

The augmentation pool supported three experimental arms:

  • targeted
  • random
  • oracle

The arm budgets were defined separately for BERT Base SNLI, RoBERTa Large SNLI, and DeBERTa Large SNLI.

These arm names describe the research design. They are not Hugging Face dataset splits.

Intended use

This dataset is intended for model training experiments that add new paraphrase rows to the original SNLI training data.

The original premise, hypothesis, and label are preserved. The new paraphrase is used as the augmented hypothesis.

Researchers should report which arm was used, the number of augmentation rows added, the model configuration, the random seed, and the exact repository version.

Not intended for evaluation

Do not treat this dataset as an independent SNLI validation or test set.

Do not use it as the paraphrase consistency benchmark published in the separate SNLI Paraphrase Bank repository.

Training and evaluation data should remain separated according to the experimental protocol.

Limitations

The paraphrases were generated and checked by models. They were not reviewed one by one by human annotators.

Automated checks can accept an imperfect paraphrase or reject a valid one.

The rows depend on the generator, verifier, filtering rules, and Phase B sampling design.

The CSV contains augmentation rows only. Users must obtain the original SNLI training data separately.

License

The Stanford Natural Language Inference Corpus is distributed under the Creative Commons Attribution-ShareAlike 4.0 International license.

This augmentation repository uses the same cc-by-sa-4.0 license designation.

Users must review the original SNLI documentation and comply with its attribution and share-alike requirements.

Associated research

This dataset was created during Phase B of the following research manuscript:

“Opening the Black Box: Localizing semantic inconsistency in NLI models with Deep k -Nearest Neighbors”

The manuscript is in preparation. It has not been submitted or published.

This section will be updated when a public preprint or accepted version becomes available.

Citation

Until the paper is public, please cite this dataset repository:

@misc{mashiach2026snlitrainingparaphraseaugmentation,
  author = {Lidor Mashiach},
  title = {SNLI Training Paraphrase Augmentation},
  year = {2026},
  publisher = {Hugging Face},
  url = {https://huggingface.co/datasets/Lidor-Mashiach/snli-training-paraphrase-augmentation}
}

Please also cite the original SNLI paper:

@inproceedings{bowman-etal-2015-large,
  title = {A Large Annotated Corpus for Learning Natural Language Inference},
  author = {Bowman, Samuel R. and Angeli, Gabor and Potts, Christopher and Manning, Christopher D.},
  booktitle = {Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing},
  year = {2015},
  publisher = {Association for Computational Linguistics}
}

Contact

Questions, corrections, and reproducibility reports can be posted in the Community tab of this repository.
Downloads last month
36
Number of rows:
2,235,245
Total file size:
389 MB