Binary Recommendation Probes — DEJAN Codebase Explorer

Binary recommendation probes query LLMs using strict metaprompt templates to evaluate whether a brand is recommended for a query or entity, converting single-token 'yes' or 'no' responses into an empirical relevance percentage.

Referenced by

Relevance Gap Detection
Relevance Probes
Relevance Score Calculation
Q2B vs Relevance Probes
ARC Relevance Pillar
Relevance in ARC
Measuring Relevance Probes
Query Relevance Score calculation
ARC Framework Breakdown
Relevance Prompt Templates
Active Model Recommendations
AI Relevance Probes
Relevance Binary Prompt Design