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

CONFLICTS

Benchmarks

Task NameDataset NameSOTA ResultTrend
Knowledge Conflict ResolutionConflicts
Exact Match (EM)65.4
30
Factual CorrectionCONFLICTS
ROUGE97
25
Question AnsweringCONFLICTS
Exact Match62.9
16
Fact VerificationCONFLICTS
Oracle Gap (pp)0
12
Showing 4 of 4 rows