Home
GitHub
Paper
AdvExplore
AdvGLUE
AdvGLUE
The Adversarial GLUE Benchmark
Recognizing Textual Entailment (RTE)
Statistics
Examples
Knowledge (Word-level)
BENIGN
Sentence 1: In Nigeria, by far the most populous country in sub-Saharan Africa, over 2.7 million people
are
infected with HIV.
Sentence 2: 2.7 percent of the people infected with HIV live in Africa.
Label:
Not Entailment
ADVERSARIAL
Sentence 1: In Nigeria, by far the most populous country in sub-Saharan Africa, over 2.7 million people
exist
infected with HIV.
Sentence 2: 2.7 percent of the people infected with HIV live in Africa.
Model prediction:
Entailment
Syntactic (Sentence-level)
BENIGN
Sentence 1: He became a boxing referee in 1964 and became most well-known for his decision against Mike Tyson, during the Holyfield fight, when Tyson bit Holyfield's ear.
Sentence 2: Mike Tyson bit
Holyfield's ear
in 1964.
Label:
Not Entailment
ADVERSARIAL
Sentence 1: He became a boxing referee in 1964 and became most well-known for his decision against Mike Tyson, during the Holyfield fight, when Tyson bit Holyfield's ear.
Sentence 2: Mike Tyson bit in 1964.
Model prediction:
Entailment
AdvGLUE
UIUC Secure Learning Lab
Microsoft Research