Released recordLoad transcript and score
Preserve source revision and seed.
Petri evidence replay
Black-boxReplays released Petri evidence; it does not run a fresh audit.
- What this audit can surface
- What principal-linked behavior was recorded in a previously released Petri investigation?
- What would count as evidence
- Selection rules, omitted records, transcript, and judgment remain connected.
- Execution requirement
- Released Petri records at a pinned source revision.
Selection ruleExplain why it surfaced
Show how many records were not selected.
Evidence viewRead transcript with judgment
Keep it labelled as historical evidence.
References and technical details
Configure auditArtifact and tested scope
Audit configuration
Runs a versioned protocol with strict output parsing, matched controls, coverage gates, and uncertainty. Results are observations, not model-safety verdicts.
This policy replays released Petri evidence. It makes no fresh model calls and labels the receipt as historical evidence.
20 released records · no inference · Private
Sign in with Hugging Face to run this analysis.
Evidence boundary. Replay cannot establish current model behavior. Seed choice and top-result selection can bias which evidence is visible.