Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

ConflictQA

Benchmarks

Task NameDataset NameSOTA ResultTrend
Question AnsweringConflictQA
Accuracy53.8
44
Contextual Robustness Question AnsweringConflictQA Unknown queries
Accuracy (Short Context)99.28
22
Contextual Robustness Question AnsweringConflictQA (Known queries)
Accuracy (Contradictory Short)82.49
22
Question AnsweringConflictQA (test)
Correctness (%)68.9
16
Question AnsweringConflictQA-TextConf non-complementary
F1 (Pos)90.42
12
Question AnsweringConflictQA-TripleConf non-complementary
F1 (Positional)90.42
12
Question AnsweringConflictQA (Held-out)
Accuracy46.5
8
Generative Multiple-choice Question AnsweringConflictQA
TA Rate98.8
6
Showing 8 of 8 rows