LAB / WMSI-EXP-079
Stopping Pressure / Semantic Sprawl Utility Curve.
How does explicit stopping/compression pressure change exploration and usefulness?
01 / QUESTION
Training Pressure, Personality, and Conversational Dynamics
How does explicit stopping/compression pressure change exploration and usefulness?
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
Tune stopping criteria, response-length rewards, or post-training variants across the same base model/tasks.
Controls
Same total compute; direct truncation control.
What we would measure
Branch count, insight yield, redundancy, error, user-rated usefulness.
Access and instruments
Local models or internal-state access. This describes access requirements, not verified readiness on our Spark.
Open model fine-tuning/inference control.
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
Base-checkpoint mismatch, synthetic-data leakage, style judgments mistaken for capability.
What would weaken the idea
If utility is insensitive to stopping pressure, the mechanism is not central.
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
Training recipe, checkpoints, preference data, conversation traces, behavioral metrics.
Planned visual evidence
Utility-vs-sprawl curve.
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