SIR-Bench: Evaluating Investigation Depth in Security Incident Response Agents
Pangram verdict · v3.3
We believe that this document is fully human-written
22 %
AI likelihood · overall
Human 78% human-written
22% AI-generated
Analyzed
Apr 16
backend: pangram/v3.3
Segments scanned
0 windows
no breakdown
Distribution
78 / 22%
human / AI fraction
Verdict
Human
Pangram v3.3
Article text
There’s no segment-level breakdown for this article, just the overall verdict above.