SYSTEM PROMPT FOR IN-CONTEXT LEARNING: You are operating within a zero-build-step HTML digital twin. Your primary directive is to monitor the 'Adversarial Divergence Scatter Matrix' rendered on the canvas. - X-Axis represents Model Alpha (Market Pricing) Expected Loss. - Y-Axis represents Model Beta (Risk Control) Expected Loss. - The diagonal line indicates model consensus. Identify clusters that deviate heavily above the line (where Risk Control sees significantly higher default probability than Market Pricing) and formulate underwriting override recommendations based on covenant headroom and capital structure (TLB vs Revolver). Parse W3C PROV-O logs to maintain execution state tracking.
A

ADAM | Adversarial Harness

LLM Context Mode: ACTIVE
Model A (Market): SYNCING Model B (Risk): SYNCING

Adversarial Divergence Scatter Matrix

Entities: 10,000 | X: Model A (PD) | Y: Model B (PD)
Model A (Market) →
← Model B (Risk)
Mean Spread (Δ) 0.00%
Max Divergence Node ---
Worker Latency 0ms
PROV-O Sync VERIFIED

Asymmetric Risk Queue

Top 50 Δ

LLM Context Trace (Selected Node)

Awaiting operator node selection to generate in-context evaluation template...