Syntony / Research

AI evaluation
and governance.

Research on AI system capabilities, evaluation methods and organizational responses to findings.

Research collectionsAnalysis & open methods
01 / READEssays & field notes

Insights

Articles on AI evaluation methods and governance decisions.

02 / EXPLOREInteractive evidence

Benchmarks

Interactive datasets with case records, source links and published methods.

FEATURED / POLBENCHPublic evidence pilot · v0.4

PolBench records policy and oversight changes after security incidents and AI evaluations. The evidence search runs through October 20, 2026.

Explore PolBench
PolBench
Finding
Decision
Ownership
Implementation
Validation
Five categories used to classify each response.

Methods
and source material

The benchmark pages include coding rules and source links. Contact us about the methods or to suggest a correction.

Discuss a research question