Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

zsRE

Benchmarks

Task NameDataset NameSOTA ResultTrend
Knowledge EditingZSRE
Generality97.36
110
Subject inference attackzsRE batch-edit tasks
Recall99
36
Sequential Knowledge EditingZsRE sequential editing 10,000 Samples
Efficacy Success (Eff)97.8
33
Knowledge EditingZsRE 10,000 facts
Reliability100
27
Model EditingZsRE
Reliability80.5
26
Model EditingzsRE
Efficacy98.91
24
Sequential model editingZsRE
Efficacy96.87
24
Knowledge EditingZsRE (evaluation)
Reliability99
21
Slot FillingzsRE
Coverage EM57.29
20
Knowledge EditingZsRE (test)
Normalized Editing Time0.65
18
Model EditingZSRE
Reliability0.975
16
Lifelong Knowledge EditingZsRE
Reliability73.5
14
Sequential Model EditingZSRE (test)
Reliability99.6
14
Model EditingZsRE 3,000 samples (test)
Relational Score99.1
13
Sequential Knowledge EditingZsRE
Efficacy0.8447
12
Prompt recovery attackzsRE
Top-1 Accuracy40
12
Knowledge EditingZsRE 1000 edits Sequential (test)
Efficiency100
12
Knowledge EditingZsRE 500 edits Sequential (test)
Efficiency100
12
Knowledge EditingZsRE 100 edits Sequential (test)
Efficiency100
12
Slot FillingzsRE KILT (test)
KILT Accuracy72.55
12
Prompt recovery attackzsRE (test)
Top-1 Accuracy34
9
Model EditingZsRE 3,000 samples
Rel Score (QA Context)77.8
8
Knowledge EditingZsRE (val)
Rel.1
8
Relation ExtractionzsRE (val)
Efficacy99.8
8
Model EditingzsRE
TRR72
7
Showing 25 of 34 rows