Evaluation & Evals

Evaluate sovereign LLM systems: LLM-as-judge frameworks, RAGAS for RAG evaluation, task-specific metrics, perplexity baseline comparison, and human evaluation workflows.