Skip to content

Evaluation

Manit Arora edited this page Oct 7, 2026 · 1 revision

Evaluation

Evaluation keeps model quality visible.

Workflow

CONFIGURATION
      ↓
     RUN
      ↓
   RESULTS
      ↓
    REVIEW

Goals

  • Make evaluation repeatable
  • Record results
  • Compare model behaviour
  • Keep quality signals visible

Version-specific evaluation runners should be documented here as they stabilize.

Clone this wiki locally