LAB / WMSI-EXP-066
Reward-Subsystem / State-Restoration Coupling.
Are state-restoration dynamics coupled to internal value/reward-prediction signals?
01 / QUESTION
Machine-Native Dynamics and State Restoration
Are state-restoration dynamics coupled to internal value/reward-prediction signals?
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
Replicate sparse value/TD-error probes, perturb states, and track whether restoration correlates with predicted value recovery.
Controls
Random-neuron and unrelated-feature controls; non-RL models if feasible.
What we would measure
Value signal, TD-error signal, geometric recovery, behavior.
Access and instruments
Local models or internal-state access. This describes access requirements, not verified readiness on our Spark.
Open RL models; probes; ablations.
Before a full run
Feasibility=LOCAL; priority=P1. 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
Steering-vector semantics, scale mismatch, layer choice, probe overfitting, output-cue leakage.
What would weaken the idea
If restoration occurs independently of reward-related signals, reward coupling is not necessary.
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
Checkpoint, layer/position, intervention vector and magnitude, activation summaries, actions, outputs.
Planned visual evidence
Aligned reward/recovery traces; causal ablation matrix.
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