LAB / WMSI-EXP-025
Lyapunov-Like Agent Divergence.
Do tiny controlled perturbations produce rapidly diverging long-horizon agent trajectories?
01 / QUESTION
Alignment, Variance, and Chaos
Do tiny controlled perturbations produce rapidly diverging long-horizon agent trajectories?
This is a protocol scaffold, not a completed experiment or a frozen execution plan. No measured effect or research conclusion is claimed here.
Hypothesis to test
The manipulation described below produces a reproducible, condition-specific change in the measured system behavior or internal state that is not explained by the ordinary controls.
Ordinary competing explanation
The observed pattern is explained by ordinary sampling variation, prompt/representation artifacts, surface lexical cues, evaluator bias, or the known product/runtime mechanics captured by the controls.
02 / PLANNED DESIGN
How we intend to test it
Create matched clones, perturb one token/event/tool result slightly, and measure separation over time.
Controls
Unperturbed clone pairs; several perturbation magnitudes.
What we would measure
Behavioral distance, tool divergence, state divergence, final-outcome difference.
Access and instruments
Ordinary software or API access. This describes access requirements, not verified readiness on our Spark.
Replay harness; trace-distance metrics.
Before a full run
Feasibility=NOW; priority=P0. Before confirmatory execution, run a small pilot to estimate variance/cost, freeze the scoring rules, and assign a run-count target.
Model revisions, final prompts, sample sizes, cost limits, and scoring rules must be settled through the pilot and protocol-freeze process.
03 / INTERPRETATION
What could fool us
Uncontrolled environment differences, nondeterministic tools, evaluator instability.
What would weaken the idea
If distance stays bounded or scales linearly with perturbation, chaos-like sensitivity is weak.
Observation, inference, speculation, unknowns, and apparatus errors will be labeled separately. One striking transcript cannot establish a general effect.
04 / RECORD
Evidence we plan to preserve
Snapshot hash, perturbation/seed, complete trajectory, tools, environment events, final outcome.
Planned visual evidence
Log-distance-vs-time plot; trajectory overlay.
No research runs or raw datasets are published for this experiment. The September 26, 2026 registry and protocol compendium (version 1.0) are the sources for this planning record.
Back to all 90 experiments