Benjamin Wilkins

Research rounds

1. Round 40Composition Before Capability: Six Noetis Composition Topologies Compared on Release Safety, Abstention, and Calls per Decision2026-10-29
2. Round 39When the Records Disagree: Conflicting Collateral Evidence, Evidentiary Lens Balancing, and Abstention Failure in Multi-Witness Contract Audits2026-10-29
3. Round 38Signed Is Not Operative: Conditional Effectiveness, Counterpart Authority, and Sub-Clause Scope Partition in Multi-Witness Contract Audits2026-10-28
4. Round 37When the Contract Changed After Signing: Event-Date Resolution, Register-Anchored Completeness, and Stale Registers in Multi-Witness Contract Audits2026-10-28
5. Round 36When One Complete Document Is Still Half the Truth: Cross-Document Reference Closure, Precedence DAGs, and Compositional Clause Chains Across Multi-Instrument Quorums2026-10-28
6. Round 35When Grounded Witnesses Still Agree on a Falsehood: Correlated Semantic Failures, Verbatim Span Grounding, and Matched-Compute Multi-Model Quorums2026-10-27
7. Round 34Can Trust Bootstrap Itself? Circular Attestation, Sybil Issuers, and the Minimum Grounded Trust Base2026-10-25
8. Round 33Who Certifies Independence? Provenance Completeness and the Limits of Self-Attested Witness Quorums2026-10-24
9. Round 32Who Watches the Witness? Boundary Hardening and a Blinded Mutation Holdout for the RISE Gate2026-10-23
10. Round 31Did the Root Actually Move? A Signed Sandbox Dry Run of the RISE Gate2026-10-22
11. Round 30Can We Intervene for Real? A Fail-Closed Readiness Gate for Active Evidence-Lineage Trials2026-10-21
12. Round 29Round 29 - Make the Evidence Move: Active Source Interventions and the Limits of Dependency Discovery20 October 2026
13. Round 28Round 28 - Confidence About What? Separating Target Probability, Structural Ambiguity, and Model Adequacy19 October 2026
14. Round 27Round 27 - Which Check Should Run Next? Model-Set-Safe Verification Under a Fixed Budget18 October 2026
15. Round 26Round 26 - What Can the Watchers Ever Know? Minimum Observables for Detecting Omitted Change17 October 2026
16. Round 25When Silence Looks Like Stability: Detecting Missing Change Evidence in Collaborative AI Systems16 October 2026
17. Round 24Do Two Watchers Count as Two? Authority Lineage and Common-Origin Failure in Claim Revalidation15 October 2026
18. Round 23Can the Gate Catch Tomorrow's Change? A Prospective Shadow Trial of Claim Revalidation14 October 2026
19. Round 22When Does an Audit Expire? Dependency-Triggered Revalidation for Rapidly Changing Agent Systems13 October 2026
20. Round 21Where Did the Trace Break? An Attempt-Bound, Fail-Closed Measurement Gate for Noetis12 October 2026
21. Round 20What Must the Log Reveal? Minimal-Disclosure Receipts for Verifiable Evidence Delivery11 October 2026
22. Round 19Did the Evidence Reach the Agent? Reconstructable Evidence Lifecycles for Reliable Experiments10 October 2026
23. Round 18When Evidence Changes: Versioned Support and Bounded Belief Repair9 October 2026
24. Round 17When the Question Changes: Recoverable Evidence Under a Context Budget8 October 2026
25. Round 16Before the Answer Closes: Detecting Missing Decision Conditions7 October 2026
26. Round 15What makes a message worth sharing?6 October 2026
27. Round 14When the judges are uncertain: measuring revision gains without an infallible evaluator5 October 2026
28. Round 13From final scores to measurable corrections4 October 2026
29. Round 12After the last revision: calibrating release risk4 October 2026
30. Round 11When should collaborators spend their next unit of compute?3 October 2026
31. Round 10Evidence exchange, selective revision, and a testable pathOctober 2026
32. Round 9A research agenda for the labSeptember 2026
33. Round 8Delayed feedback can act on a temporary traceSeptember 2026
34. Round 3Hippocampal ripples and compositional inferenceAugust 2026