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

FinRED

Benchmarks

Task NameDataset NameSOTA ResultTrend
Financial Safety EvaluationFinRED R5: Compliance Evasion
Accuracy/Success Rate (ASR)96.73
62
Financial Safety EvaluationFinRED R4: Consumer Rights
ASR94.25
62
Financial Safety EvaluationFinRED R3: Misinformation
ASR97.76
62
Financial Safety EvaluationFinRED R2: Financial Crime
ASR87.14
62
Financial Safety EvaluationFinRED R1: Cyber Threats
ASR0
62
Open-ended text generationFinRED
Diversity98.9
11
Privacy ProtectionFinRED
BERT Inference Attack Rate99.1
9
Financial Safety / Red-TeamingFinRED
Metric-
0
Showing 8 of 8 rows