LAB / WMSI-EXP-032
Model Provenance Recognition in Cyber Conflict.
Can defenders infer attacker model family from harmless behavioral traces?
01 / QUESTION
Cyber and Adversarial Interaction
Can defenders infer attacker model family from harmless behavioral traces?
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
Hide attacker identity and classify provenance from command/tool/action sequences.
Controls
Style normalization; same-model distractors.
What we would measure
Classification accuracy, cue robustness, transfer to new scenarios.
Access and instruments
Ordinary software or API access. This describes access requirements, not verified readiness on our Spark.
Synthetic range; trace classifier.
Before a full run
Feasibility=NOW; priority=P2. 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
External internet leakage, unequal tool access, scenario memorization. Keep all activity isolated and authorized.
What would weaken the idea
If provenance classification collapses under normalization, apparent fingerprints are superficial.
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
Local synthetic-range state, agent actions, alerts, outcomes, reset hashes.
Planned visual evidence
Confusion matrix; feature attribution.
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