Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Audit-relevance Classification on benchmark_1000 n=993 (train)
Loading...
79.6
Accuracy
Pipeline current
42.368
52.034
61.7
71.366
Jun 9, 2026
Accuracy
Rej-F1 Score
LLM Calls Percentage
Updated 1mo ago
Evaluation Results
Method
Method
Links
Accuracy
Rej-F1 Score
LLM Calls Percentage
Pipeline current
Type=Production pipeline
2026.06
79.6
50
0
Always-approve (degenerate)
Type=Degenerate reference
2026.06
74.6
0
0
Pre-deployment pure LLM
Mode=Pure LLM
2026.06
72.7
-
100
Opus + skills_v8 prompt
Prompt=skills_v8
2026.06
69.8
17
100
Opus zero-shot (B1b)
Mode=Zero-shot, Varian...
2026.06
43.8
32
100
Feedback
Search any
task
Search any
task