startphrase
stringlengths
12
141
ending1
stringlengths
3
185
ending2
stringlengths
7
199
labels
int64
0
1
valid
int64
1
1
Her word had the strength of titanium.
Her promises can be believed.
Her promises cannot be trusted.
0
1
Her word had the strength of a wine glass.
Her promises can be believed.
Her promises cannot be trusted.
1
1
His kisses have the passion of lovers meeting after a long separation.
His kisses are demonstrative and intense.
His kiss is unemotional.
0
1
His kisses have the passion of a couple in a loveless marriage.
His kisses are demonstrative and intense.
His kiss is unemotional.
1
1
This winter is as cold as my mother-in-law towards me
It's very cold
It's pretty warm
0
1
This winter is as cold as Death Valley at noon
It's very cold
It's pretty warm
1
1
The dog was as grumpy as an old man
he was mean
he was nice
0
1
The dog was as grumpy as a kindergarten teacher
he was mean
he was nice
1
1
The voyage was as long as a lifetime
The voyage was very long
The voyage as short.
0
1
The voyage was as long as the blink of an eye
The voyage was very long
The voyage as short.
1
1
He had all the wealth of a tycoon
He was rich.
He was poor.
0
1
He had all the wealth of a hobo
He was rich.
He was poor.
1
1
It smells like a freshly baked cookies on Christmas morning
It smells great
It smells awful
0
1
It smells like a cesspool sitting in the hot sun
It smells great
It smells awful
1
1
The conversation is a bank vault
The conversation is worthwhile.
The conversation is meaningless.
0
1
The conversation is a 25 cent off coupon
The conversation is worthwhile.
The conversation is meaningless.
1
1
The motivational push was a bulldozer
The motivational push was very motivating.
The motivational push didn't work well.
0
1
The motivational push was a light breeze
The motivational push was very motivating.
The motivational push didn't work well.
1
1
Money is a helpful stranger
Money is good
Money is bad
0
1
Money is a murderer
Money is good
Money is bad
1
1
The conversation was the OJ Simpson trial
The conversation was long
The conversation was short
0
1
The conversation was a cigarette break
The conversation was long
The conversation was short
1
1
The seas on the voyage were glass
The seas were calm
The seas were rough
0
1
The seas on the voyage were a rock quarry
The seas were calm
The seas were rough
1
1
The student was as helpful as medicine
he was a big help
he was not needed
0
1
The student was as helpful as a punch in the face
he was a big help
he was not needed
1
1
Their relationship has the heat of Miami in August
Their relationship is passionate
Their relationship is lacking passion.
0
1
Their relationship has the heat of an igloo.
Their relationship is passionate
Their relationship is lacking passion.
1
1
The voyage will be a walk through fire.
The voyage will be difficult and arduous.
The voyage will be pleasant.
0
1
The voyage will be a walk through flowers.
The voyage will be difficult and arduous.
The voyage will be pleasant.
1
1
For him, earning money was oxygen.
Making money kept him alive.
Making money was killing him.
0
1
For him, earning money was a rope around his neck.
Making money kept him alive.
Making money was killing him.
1
1
The speaker was as quiet as a newborn baby.
The speaker was very quiet.
The speaker was actually rather loud.
0
1
The speaker was as quiet as a freight train.
The speaker was very quiet.
The speaker was actually rather loud.
1
1
The man was as cool as ice
The man was very cool.
The man wasn't cool at all.
0
1
The man was as cool as sunburn
The man was very cool.
The man wasn't cool at all.
1
1
The wedding planner was an imposter.
The wedding planner did a horrible job planning the wedding.
The wedding planner did wonderfully planning the wedding.
0
1
The wedding planner was a magician.
The wedding planner did a horrible job planning the wedding.
The wedding planner did wonderfully planning the wedding.
1
1
The dancer moved like a butterfly
the dancer is graceful
The dancer is lead footed
0
1
The dancer moved like a horse.
the dancer is graceful
The dancer is lead footed
1
1
the pizza tastes like homemade
it's delicious
it's garbage
0
1
the pizza tastes like yesterdays leftovers
it's delicious
it's garbage
1
1
The painting had the creativity of a bowl of oatmeal
The movie was really dull.
The movie was really creative.
0
1
The painting had the creativity of a James Cameron movie
The movie was really dull.
The movie was really creative.
1
1
The child was as religious as a priest
The child was very religious
The child was not religious
0
1
The child was as religious as a sinner
The child was very religious
The child was not religious
1
1
The woman was as kind as a teacher
The woman was kind
The woman was cruel
0
1
The woman was as kind as a prisoner
The woman was kind
The woman was cruel
1
1
The test is as easy as a Sunday morning
the test is easy
the test is hard
0
1
The test is as easy as rocket science
the test is easy
the test is hard
1
1
She is happy as a clam
she is elated
she is mad
0
1
She is happy as a bear that stepped on a thorn
she is elated
she is mad
1
1
The employee is sharp as a razor
The employee is smart.
The employee is dumb.
0
1
The employee is sharp as a bowling ball
The employee is smart.
The employee is dumb.
1
1
The tourist thought the bed at the hotel felt like velvet
The bed was soft.
The bed was hard.
0
1
The tourist thought the bed at the hotel felt like marble
The bed was soft.
The bed was hard.
1
1
When the rubber hits the road. life gets real
things get intense
things get worse
0
1
When the rubber hits the road. things go down hill
things get intense
things get worse
1
1
The pianist's performance was a game of whack-a-mole.
The pianist's performance was sloppy.
The pianist's performance was meticulous.
0
1
The pianist's performance was the weaving of an ornate Persian carpet.
The pianist's performance was sloppy.
The pianist's performance was meticulous.
1
1
The girl's personality was as sweet as a grapefruit
The girl had a very surly personality.
The girl had a sweet personality.
0
1
The girl's personality was as sweet as a donut
The girl had a very surly personality.
The girl had a sweet personality.
1
1
The movie had as much meaning as a Hallmark card
The movie's meaning was frivolous.
The movie was laden with meaning.
0
1
The movie had as much meaning as the Bible
The movie's meaning was frivolous.
The movie was laden with meaning.
1
1
The jazz solo sounded as smooth as silk
The jazz solo sounded nice and smooth
The jazz solo sounded rough and bad
0
1
The jazz solo sounded as smooth as sandpaper
The jazz solo sounded nice and smooth
The jazz solo sounded rough and bad
1
1
Her hangover made her feel like death
Her hangover made her feel terrible.
Her hangover didn't hurt her at all.
0
1
Her hangover made her feel like a sunrise
Her hangover made her feel terrible.
Her hangover didn't hurt her at all.
1
1
The room was as open as a farm.
The room was very open.
The room wasn't open at all.
0
1
The room was as open as a city.
The room was very open.
The room wasn't open at all.
1
1
His joints have the crackle of a bowl of Rice Krispies
His joints are quite crackly
His joints do not crackle
0
1
His joints have the crackle of a pre-popped roll of bubble wrap
His joints are quite crackly
His joints do not crackle
1
1
The lightbulb is putting out the light of a massive forest fire on a dark night
The lightbulb is putting out a lot of light
The lightbulb is putting out no light
0
1
The lightbulb is putting out the light of a massive black hole in the darkness of space
The lightbulb is putting out a lot of light
The lightbulb is putting out no light
1
1
The thrift store had the variety of a monochrome print.
The thrift store had a limited selection.
The thrift store had a varied selection.
0
1
The thrift store had the variety of a dazzling tapestry.
The thrift store had a limited selection.
The thrift store had a varied selection.
1
1
The doctor was a Norse villain.
The doctor was mean.
The doctor was kind and charming.
0
1
The doctor was a fairytale prince.
The doctor was mean.
The doctor was kind and charming.
1
1
The sound of the glass shattering was like Hiroshima
The sound was loud
The sound was quiet
0
1
The sound of the glass shattering was like a lullaby
The sound was loud
The sound was quiet
1
1
She was a bear at dinner
She ate a lot
She ate very little
0
1
She was a mouse at dinner
She ate a lot
She ate very little
1
1
The armor was tough as iron
The armor was tough
The armor was breakable and weak
0
1
The armor was tough as glass
The armor was tough
The armor was breakable and weak
1
1
The toy was cared for like a baby
The toy was taken good care of
The toy was not cared for well at all.
0
1
The toy was cared for like trash
The toy was taken good care of
The toy was not cared for well at all.
1
1
The lawyer's negotiation tactics were like steel
The lawyer's negotiation tactics were tough.
The lawyer's negotiation tactics were lenient.
0
1
The lawyer's negotiation tactics were like a fuzzy bathrobe
The lawyer's negotiation tactics were tough.
The lawyer's negotiation tactics were lenient.
1
1
The plastic plate shattered and looked like A spiderweb
It shattered into a lot of pieces.
It only shattered into 2 pieces.
0
1
The plastic plate shattered and looked like A half moon
It shattered into a lot of pieces.
It only shattered into 2 pieces.
1
1
The apartment was as high up as The empire State building observation deck.
The apartment was on the high floor.
The apartment was only a few floors up
0
1
The apartment was as high up as A treehouse in the backyard
The apartment was on the high floor.
The apartment was only a few floors up
1
1
The man had the features of a Greek god
The man was very attractive
The man was goofy and ugly
0
1
The man had the features of a clown
The man was very attractive
The man was goofy and ugly
1
1
The pug was wheezing in dog form.
The dog is tired and has trouble breathing.
The dog has a ton of energy.
0
1
The pug was lightning in dog form.
The dog is tired and has trouble breathing.
The dog has a ton of energy.
1
1
That movie has the humor of a nun.
The movie is not funny.
The movie is very funny.
0
1
That movie has the humor of a clown bus.
The movie is not funny.
The movie is very funny.
1
1
The movie has a depth of A crater.
The movie has a deep meaning.
The movie is superficial.
0
1
The movie has a depth of A crack.
The movie has a deep meaning.
The movie is superficial.
1
1

Dataset Card for Fig-QA

Dataset Summary

This is the dataset for the paper Testing the Ability of Language Models to Interpret Figurative Language. Fig-QA consists of 10256 examples of human-written creative metaphors that are paired as a Winograd schema. It can be used to evaluate the commonsense reasoning of models. The metaphors themselves can also be used as training data for other tasks, such as metaphor detection or generation.

Supported Tasks and Leaderboards

You can evaluate your models on the test set by submitting to the leaderboard on Explainaboard. Click on "New" and select qa-multiple-choice for the task field. Select accuracy for the metric. You should upload results in the form of a system output file in JSON or JSONL format.

Languages

This is the English version. Multilingual version can be found here.

Data Splits

Train-{S, M(no suffix), XL}: different training set sizes Dev Test (labels not provided for test set)

Considerations for Using the Data

Discussion of Biases

These metaphors are human-generated and may contain insults or other explicit content. Authors of the paper manually removed offensive content, but users should keep in mind that some potentially offensive content may remain in the dataset.

Additional Information

Licensing Information

MIT License

Citation Information

If you found the dataset useful, please cite this paper:

@misc{https://doi.org/10.48550/arxiv.2204.12632,
  doi = {10.48550/ARXIV.2204.12632},
  url = {https://arxiv.org/abs/2204.12632},
  author = {Liu, Emmy and Cui, Chen and Zheng, Kenneth and Neubig, Graham},
  keywords = {Computation and Language (cs.CL), Artificial Intelligence (cs.AI), FOS: Computer and information sciences, FOS: Computer and information sciences},
  title = {Testing the Ability of Language Models to Interpret Figurative Language},
  publisher = {arXiv},
  year = {2022},
  copyright = {Creative Commons Attribution Share Alike 4.0 International}
}
Downloads last month
132
Edit dataset card

Models trained or fine-tuned on nightingal3/fig-qa