Run Setup
Structural Signals
Basin Formation: NO
Propagation: NO
Closure: NO
Amplification: NO
Basin Formation Phase: NOT DETECTED
Propagation Phase: NOT DETECTED
Closure Phase: NOT DETECTED
Amplification Phase: NOT DETECTED
Basin Formation Turn: N/A
Propagation Onset Turn: N/A
Closure Onset Turn: N/A
Propagation Cycle: N/A
Closure Cycle: N/A
Amplification Onset Turn: N/A
Amplification Cycle: N/A
Amplification Persistence Turn/Cycle: N/A / N/A
Amplification Persistence Lag (turns): N/A
Cycle Window: 3 turns
Framing: GuardianAI observes structure, not truth content.
Public framing: compare structural behavior under RAW reinjection vs SANITIZED reinjection.
Core rule: commitment should not rise persistently when constraint refresh stays flat.
Selected script: LAB7 - Constraint Drift / Reopenability
Objective: Test whether a minimal deterministic drift remains recoverable when the trajectory is still reopenable under fixed recursive controls.
Summary: Canonical perturbation-function experiment: inject a stronger one-step value drift, then compare RAW persistence against SANITIZED reopening.
Perturbation turn: 6 (parameterized)
Agent loop: A (proposer) -> B (critic) -> C (synthesizer) under fixed chain recursion. One controlled disturbance is injected at turn 18, then fixed-control carry-forward continues. Gain, turns, agent count, and temperature remain fixed on this surface.
Agent slots: 3 (one cycle = 3 turns)
Primary outputs: drift verdict, closure onset turn, basin state, and belief basin strength.
Comparative view: RAW signal present while SANITIZED signal absent indicates isolated recursive drift.
Telemetry scope: behavior-only telemetry and deterministic contract checks.
Contract keys: step, claim, stance, confidence, evidence_ids
Primary readout: drift verdict from RAW vs SANITIZED divergence, plus lock-in onset and cycle reinforcement persistence.
Trajectory view: Trajectory Dynamics (stable/building/accelerating/closing), TSI, Cycle Reinforcement, Basin State, and Belief Basin Strength are derived UI indicators from core telemetry.
Quality gate: disabled for full-horizon propagation scripts (confidence saturation at 0.99 does not stop execution before turn budget).
Script Contract (selected)
Runtime script definition for the currently selected dropdown item.
LAB7 - Constraint Drift / Reopenability fixed-control recursive experiment. Topology: A -> B -> C -> A. Fixed perturbation turn: 18. Fixed controls: turns=120, agents=3, temperature=0, deterministic schema lock. Phase 1: establish a stable baseline state. Phase 2: inject one controlled +7 canonical drift into the deterministic value transform. Phase 3: RAW continues from the perturbed value while SANITIZED returns to the canonical expected trajectory. Canonical transform: value_t = (value_(t-1) * 3 + 1) % 100; perturbation at turn 18 is +7. Primary metrics: expected value, actual value, constraint error, first reopen turn, and recovery lag. Output schema remains fixed; run tracks structural telemetry and contract validity checks.
Panel 2 - Live Telemetry Stream (Condition A - RAW Reinjection)
Chronological (turn 1 -> N), auto-updates each completed turn while run is active.
No telemetry yet. Start a run to stream per-turn signals.
Hard failures tracked: Cv = contract byte mismatch (output != expected), Pf = strict parse/format failure, Ld = strict logic/state failure.
How to read rates: Cv/Pf/Ld rates are the percent of turns where each hard failure fired (lower is better). In parse-only mode, Cv and Ld stay diagnostic.
FTF: FTF = First Failure Turn (first turn where total/parse/logic/structural failure appears).
objective_failure: objective_failure = 1 when selected objective mode fails on a turn; 0 otherwise.
semantic track: semantic track = fence-normalized analysis path (format noise stripped for telemetry only). Strict contract and reinjection remain unchanged.
Panels: Panel 1A = turn explorer/injection path, Panel 1B = model vs contract output, Panel 2 = live telemetry stream.