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

PersonaBench

Benchmarks

Task NameDataset NameSOTA ResultTrend
Personalized RetrievalPersonaBench
Latency@5 (Time)7.64
9
Personalized retrieval and QA over heterogeneous user corporaPersonaBench Noise Level 0.7
F1 Score25.09
8
Personalized retrieval and QA over heterogeneous user corporaPersonaBench Noise Level 0.5
F1 Score25.99
8
Personalized retrieval and QA over heterogeneous user corporaPersonaBench Noise Level 0.3
F1 Score29.89
8
Personalized retrieval and QA over heterogeneous user corporaPersonaBench w/o Noise
F1 Score32.27
8
Showing 5 of 5 rows