knowledge-platform/docs/evaluation/RAG_EVALUATION.md

346 B
Raw Blame History

RAG Evaluation

Maintain a versioned Golden Dataset and report Recall@1/3/5/10, MRR, NDCG@10, latency P50/P95.

Benchmark the full pipeline: Embedding × Retrieval Strategy × Fusion × Reranker.

Initial embedding candidates: bge-large-zh, Qwen3-Embedding-0.6B, Qwen3-Embedding-4B, optional 8B.

Use synthetic/public/anonymized fixtures only.