premise
stringlengths 11
296
| hypothesis
stringlengths 11
296
| label
class label 0
classes | idx
int32 0
1.1k
|
---|---|---|---|
Bees do not follow the same rules as airplanes. | Bees fly using a different mechanism from airplanes. | -1no label
| 200 |
Bees fly using a different mechanism from airplanes. | Bees do not follow the same rules as airplanes. | -1no label
| 201 |
Bees do not follow the same rules as airplanes. | Bees fly using the same mechanism as airplanes. | -1no label
| 202 |
Bees fly using the same mechanism as airplanes. | Bees do not follow the same rules as airplanes. | -1no label
| 203 |
Bees do not follow the same rules as airplanes. | Bees are more energy-efficient flyers than airplanes. | -1no label
| 204 |
Bees are more energy-efficient flyers than airplanes. | Bees do not follow the same rules as airplanes. | -1no label
| 205 |
Leslie Dixon is married to fellow screenwriter and producer Tom Ropelewski. | Tom Ropelewski is married to fellow screenwriter and producer Leslie Dixon. | -1no label
| 206 |
Tom Ropelewski is married to fellow screenwriter and producer Leslie Dixon. | Leslie Dixon is married to fellow screenwriter and producer Tom Ropelewski. | -1no label
| 207 |
This means that solving analogy questions with vector arithmetic is mathematically equivalent to seeking a word that is similar to x and y but is different from z. | This means that seeking a word that is similar to x and y but is different from z is mathematically equivalent to solving analogy questions with vector arithmetic. | -1no label
| 208 |
This means that seeking a word that is similar to x and y but is different from z is mathematically equivalent to solving analogy questions with vector arithmetic. | This means that solving analogy questions with vector arithmetic is mathematically equivalent to seeking a word that is similar to x and y but is different from z. | -1no label
| 209 |
Bob married Alice. | Bob and Alice got married. | -1no label
| 210 |
Bob and Alice got married. | Bob married Alice. | -1no label
| 211 |
Tulsi Gabbard disagrees with Bernie Sanders on what is the best way to deal with Bashar al-Assad. | Tulsi Gabbard and Bernie Sanders disagree on what is the best way to deal with Bashar al-Assad. | -1no label
| 212 |
Tulsi Gabbard and Bernie Sanders disagree on what is the best way to deal with Bashar al-Assad. | Tulsi Gabbard disagrees with Bernie Sanders on what is the best way to deal with Bashar al-Assad. | -1no label
| 213 |
It reminds me of the times I played Super Mario with my little brother. | It reminds me of the times my little brother and I played Super Mario. | -1no label
| 214 |
It reminds me of the times my little brother and I played Super Mario. | It reminds me of the times I played Super Mario with my little brother. | -1no label
| 215 |
In this PC game, Shredder fights the turtles in his Manhattan hideout. | In this PC game, Shredder and the turtles fight in his Manhattan hideout. | -1no label
| 216 |
In this PC game, Shredder and the turtles fight in his Manhattan hideout. | In this PC game, Shredder fights the turtles in his Manhattan hideout. | -1no label
| 217 |
If their vectors' cosine similarity is high, we can conclude that "cat" is similar to "dog". | If their vectors' cosine similarity is high, we can conclude that "cat" and "dog" are similar. | -1no label
| 218 |
If their vectors' cosine similarity is high, we can conclude that "cat" and "dog" are similar. | If their vectors' cosine similarity is high, we can conclude that "cat" is similar to "dog". | -1no label
| 219 |
Jane had a party on Sunday. | On Sunday, Jane had a party. | -1no label
| 220 |
On Sunday, Jane had a party. | Jane had a party on Sunday. | -1no label
| 221 |
For decades, the FBI has been trusted to investigate corruption inside the government. | The FBI has been trusted to investigate corruption inside the government for decades. | -1no label
| 222 |
The FBI has been trusted to investigate corruption inside the government for decades. | For decades, the FBI has been trusted to investigate corruption inside the government. | -1no label
| 223 |
I was driving through my neighborhood a few weeks ago, and there was a lady holding a child's hand, standing on the side of the road talking to someone parked in a car. | A few weeks ago, I was driving through my neighborhood, and there was a lady holding a child's hand, standing on the side of the road talking to someone parked in a car. | -1no label
| 224 |
A few weeks ago, I was driving through my neighborhood, and there was a lady holding a child's hand, standing on the side of the road talking to someone parked in a car. | I was driving through my neighborhood a few weeks ago, and there was a lady holding a child's hand, standing on the side of the road talking to someone parked in a car. | -1no label
| 225 |
Bass River timber was used in construction of the Empire State Building. | In construction of the Empire State Building, Bass River timber was used. | -1no label
| 226 |
In construction of the Empire State Building, Bass River timber was used. | Bass River timber was used in construction of the Empire State Building. | -1no label
| 227 |
In this paper, we present an approach to acquire trivial physical knowledge from unstructured natural language text. | We present in this paper an approach to acquire trivial physical knowledge from unstructured natural language text. | -1no label
| 228 |
We present in this paper an approach to acquire trivial physical knowledge from unstructured natural language text. | In this paper, we present an approach to acquire trivial physical knowledge from unstructured natural language text. | -1no label
| 229 |
In this paper, we present an approach to acquire trivial physical knowledge from unstructured natural language text. | We present an approach to acquire trivial physical knowledge from the unstructured natural language text of this paper. | -1no label
| 230 |
We present an approach to acquire trivial physical knowledge from the unstructured natural language text of this paper. | In this paper, we present an approach to acquire trivial physical knowledge from unstructured natural language text. | -1no label
| 231 |
I ate pizza with friends. | I ate pizza. | -1no label
| 232 |
I ate pizza. | I ate pizza with friends. | -1no label
| 233 |
I ate pizza with some friends. | I ate some friends. | -1no label
| 234 |
I ate some friends. | I ate pizza with some friends. | -1no label
| 235 |
I ate pizza with olives. | I ate pizza. | -1no label
| 236 |
I ate pizza. | I ate pizza with olives. | -1no label
| 237 |
I ate pizza with olives. | I ate olives. | -1no label
| 238 |
I ate olives. | I ate pizza with olives. | -1no label
| 239 |
Mueller received a mandate to investigate possible collusion with Russia. | Mueller received a mandate to investigate possible collusion. | -1no label
| 240 |
Mueller received a mandate to investigate possible collusion. | Mueller received a mandate to investigate possible collusion with Russia. | -1no label
| 241 |
Mueller received a mandate to investigate possible collusion with Russia. | Mueller received a mandate to investigate with Russia. | -1no label
| 242 |
Mueller received a mandate to investigate with Russia. | Mueller received a mandate to investigate possible collusion with Russia. | -1no label
| 243 |
Mueller received a mandate to investigate possible collusion with vast resources. | Mueller received a mandate to investigate possible collusion. | -1no label
| 244 |
Mueller received a mandate to investigate possible collusion. | Mueller received a mandate to investigate possible collusion with vast resources. | -1no label
| 245 |
Mueller received a mandate to investigate possible collusion with vast resources. | Mueller received a mandate to investigate with vast resources. | -1no label
| 246 |
Mueller received a mandate to investigate with vast resources. | Mueller received a mandate to investigate possible collusion with vast resources. | -1no label
| 247 |
They should be attached to the lifting mechanism in the faucet. | They should be attached to the lifting mechanism. | -1no label
| 248 |
They should be attached to the lifting mechanism. | They should be attached to the lifting mechanism in the faucet. | -1no label
| 249 |
They should be attached to the lifting mechanism in the faucet. | They should be attached in the faucet. | -1no label
| 250 |
They should be attached in the faucet. | They should be attached to the lifting mechanism in the faucet. | -1no label
| 251 |
They should be attached to the lifting mechanism in an unbreakable knot. | They should be attached to the lifting mechanism. | -1no label
| 252 |
They should be attached to the lifting mechanism. | They should be attached to the lifting mechanism in an unbreakable knot. | -1no label
| 253 |
They should be attached to the lifting mechanism in an unbreakable knot. | They should be attached in an unbreakable knot. | -1no label
| 254 |
They should be attached in an unbreakable knot. | They should be attached to the lifting mechanism in an unbreakable knot. | -1no label
| 255 |
Tanks were developed by Britain and France, and were first used in combat by the British during a battle with only partial success. | Tanks were developed by Britain and France, and were first used with only partial success. | -1no label
| 256 |
Tanks were developed by Britain and France, and were first used with only partial success. | Tanks were developed by Britain and France, and were first used in combat by the British during a battle with only partial success. | -1no label
| 257 |
Tanks were developed by Britain and France, and were first used in combat by the British during a battle with only partial success. | Tanks were developed by Britain and France, and were first used in combat by the British during a battle. | -1no label
| 258 |
Tanks were developed by Britain and France, and were first used in combat by the British during a battle. | Tanks were developed by Britain and France, and were first used in combat by the British during a battle with only partial success. | -1no label
| 259 |
Tanks were developed by Britain and France, and were first used in combat by the British during a battle with German forces. | Tanks were developed by Britain and France, and were first used with German forces. | -1no label
| 260 |
Tanks were developed by Britain and France, and were first used with German forces. | Tanks were developed by Britain and France, and were first used in combat by the British during a battle with German forces. | -1no label
| 261 |
Tanks were developed by Britain and France, and were first used in combat by the British during a battle with German forces. | Tanks were developed by Britain and France, and were first used in combat by the British during a battle. | -1no label
| 262 |
Tanks were developed by Britain and France, and were first used in combat by the British during a battle. | Tanks were developed by Britain and France, and were first used in combat by the British during a battle with German forces. | -1no label
| 263 |
We propose models of the probability distribution from which the attested vowel inventories have been drawn. | We propose models of the probability distribution. | -1no label
| 264 |
We propose models of the probability distribution. | We propose models of the probability distribution from which the attested vowel inventories have been drawn. | -1no label
| 265 |
We propose models of the probability distribution from which the attested vowel inventories have been drawn. | We propose models from which the attested vowel inventories have been drawn. | -1no label
| 266 |
We propose models from which the attested vowel inventories have been drawn. | We propose models of the probability distribution from which the attested vowel inventories have been drawn. | -1no label
| 267 |
We propose models of the probability distribution from a restricted space of linear functions. | We propose models of the probability distribution. | -1no label
| 268 |
We propose models of the probability distribution. | We propose models of the probability distribution from a restricted space of linear functions. | -1no label
| 269 |
We propose models of the probability distribution from a restricted space of linear functions. | We propose models from a restricted space of linear functions. | -1no label
| 270 |
We propose models from a restricted space of linear functions. | We propose models of the probability distribution from a restricted space of linear functions. | -1no label
| 271 |
George went to the lake to catch a fish, but he fell into the water. | George fell into the water. | -1no label
| 272 |
George fell into the water. | George went to the lake to catch a fish, but he fell into the water. | -1no label
| 273 |
George went to the lake to catch a fish, but he fell into the water. | A fish fell into the water. | -1no label
| 274 |
A fish fell into the water. | George went to the lake to catch a fish, but he fell into the water. | -1no label
| 275 |
The Republican party almost universally opposed that bill in 2009, which cost $787 billion over 10 years, on the grounds that it would increase the debt too much. | That bill would increase the debt too much. | -1no label
| 276 |
That bill would increase the debt too much. | The Republican party almost universally opposed that bill in 2009, which cost $787 billion over 10 years, on the grounds that it would increase the debt too much. | -1no label
| 277 |
The Republican party almost universally opposed that bill in 2009, which cost $787 billion over 10 years, on the grounds that it would increase the debt too much. | The Republican party would increase the debt too much. | -1no label
| 278 |
The Republican party would increase the debt too much. | The Republican party almost universally opposed that bill in 2009, which cost $787 billion over 10 years, on the grounds that it would increase the debt too much. | -1no label
| 279 |
The climbing beans choked the corn, and the squash grew so big that it overshadowed several adjacent rows of beans. | The squash overshadowed several adjacent rows of beans. | -1no label
| 280 |
The squash overshadowed several adjacent rows of beans. | The climbing beans choked the corn, and the squash grew so big that it overshadowed several adjacent rows of beans. | -1no label
| 281 |
The climbing beans choked the corn, and the squash grew so big that it overshadowed several adjacent rows of beans. | The corn overshadowed several adjacent rows of beans. | -1no label
| 282 |
The corn overshadowed several adjacent rows of beans. | The climbing beans choked the corn, and the squash grew so big that it overshadowed several adjacent rows of beans. | -1no label
| 283 |
Because Britain still maintained control of Canada's foreign affairs under the Confederation Act, its declaration of war in 1914 automatically brought Canada into World War I. | Britain's declaration of war in 1914 automatically brought Canada into World War I. | -1no label
| 284 |
Britain's declaration of war in 1914 automatically brought Canada into World War I. | Because Britain still maintained control of Canada's foreign affairs under the Confederation Act, its declaration of war in 1914 automatically brought Canada into World War I. | -1no label
| 285 |
Because Britain still maintained control of Canada's foreign affairs under the Confederation Act, its declaration of war in 1914 automatically brought Canada into World War I. | Canada's declaration of war in 1914 automatically brought Canada into World War I. | -1no label
| 286 |
Canada's declaration of war in 1914 automatically brought Canada into World War I. | Because Britain still maintained control of Canada's foreign affairs under the Confederation Act, its declaration of war in 1914 automatically brought Canada into World War I. | -1no label
| 287 |
We show that if coreference resolvers mainly rely on lexical features, they can hardly generalize to unseen domains. | Coreference resolvers can hardly generalize to unseen domains. | -1no label
| 288 |
Coreference resolvers can hardly generalize to unseen domains. | We show that if coreference resolvers mainly rely on lexical features, they can hardly generalize to unseen domains. | -1no label
| 289 |
We show that if coreference resolvers mainly rely on lexical features, they can hardly generalize to unseen domains. | Lexical features can hardly generalize to unseen domains. | -1no label
| 290 |
Lexical features can hardly generalize to unseen domains. | We show that if coreference resolvers mainly rely on lexical features, they can hardly generalize to unseen domains. | -1no label
| 291 |
The sides came to an agreement after their meeting in Stockholm. | The sides came to an agreement after their meeting in Sweden. | -1no label
| 292 |
The sides came to an agreement after their meeting in Sweden. | The sides came to an agreement after their meeting in Stockholm. | -1no label
| 293 |
The sides came to an agreement after their meeting in Stockholm. | The sides came to an agreement after their meeting in Europe. | -1no label
| 294 |
The sides came to an agreement after their meeting in Europe. | The sides came to an agreement after their meeting in Stockholm. | -1no label
| 295 |
The sides came to an agreement after their meeting in Stockholm. | The sides came to an agreement after their meeting in Oslo. | -1no label
| 296 |
The sides came to an agreement after their meeting in Oslo. | The sides came to an agreement after their meeting in Stockholm. | -1no label
| 297 |
Musk decided to offer up his personal Tesla roadster. | Musk decided to offer up his personal car. | -1no label
| 298 |
Musk decided to offer up his personal car. | Musk decided to offer up his personal Tesla roadster. | -1no label
| 299 |