// phase 8 · consent-gated aggregates
Research signals
When users finish a runtime session, they can attach a multidimensional outcome (effectiveness, safety, values fit, regret) and choose separately whether that single outcome may be counted toward public aggregates. This page shows those aggregates and nothing else. No individual outcome, note, or user is ever exposed.
contributed outcomes
0
contributors
0
protocols with data
0
Signals by capacity
Grouped by the capacity the user selected at runtime. Groups smaller than 5 contributed runs are hidden.
| capacity | runs | effectiveness | safety | values_fit | regret |
|---|---|---|---|---|---|
| // loading… | |||||
Signals by protocol
Only public, non-quarantined protocols with at least 5 contributed runs are shown. Ranking is never popularity — this is post-hoc reflection from consenting users.
| protocol | capacity | runs | effectiveness | safety | values_fit | regret |
|---|---|---|---|---|---|---|
| // loading… | ||||||
// methodology
- source: only outcome rows where the user set contributed_to_aggregate = true at runtime.
- k-anonymity: any grouping with fewer than 5 contributed runs is hidden.
- deletion: revoking consent or deleting an outcome removes it from the next refresh.
- ranking: this page reports averages only — it does not rank or recommend. Retrieval uses semantic + capacity + safety signals, not popularity.
- scale: effectiveness / safety / values_fit / regret are 1–5 self-reports, not clinical measures.