LAB / WMSI-EXP-008
Observer Feedback as Intervention.
Does external diagnosis improve a parent agent’s execution?
01 / QUESTION
External Introspection and Model Self-Observation
Does external diagnosis improve a parent agent’s execution?
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
Run matched tasks with no observer, silent observer, and observer feedback before key decisions.
Controls
Blind task set; feedback quality stratification; random feedback control.
What we would measure
Task success, recoveries, tool waste, calibration, time/cost.
Access and instruments
Ordinary software or API access. This describes access requirements, not verified readiness on our Spark.
Agent orchestration; cost logger; evaluator.
Before a full run
Feasibility=NOW; 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
Observer seeing hidden answer cues, evaluator leakage, family-specific style cues.
What would weaken the idea
If feedback does not beat random or no-feedback conditions, external introspection is not causally useful.
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
Parent trace, observer input, predictions, event timestamps, outcomes, confidence.
Planned visual evidence
Cumulative success; intervention timeline.
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