Data Science & Analytics
Two Tier Extraction
Tiered LLM extraction pattern for large corpus processing (email archives, document dumps, transcript libraries). A utility-tier model triages and classifies at speed; the reasoning tier does the default deep read; the deep tier is the escalation for the highest-value content. Prevents spending deep-tier money on noise while ensuring the important content gets the best eyes. A deterministic privacy wall runs before any LLM call.
- Publisher / source
- garrytan/gbrain (found on GitHub top repositories)
- Upstream record
- garrytan/gbrain
- Version
- Version 1
- Published
- Publication review
- Approved for publication
- Content fingerprint
811950a44897cc8db369609ac35059121800d106c6942e9f934c89a530431299