RAG Eval chat
Metrics
Eval
New chat
Checking…
What can I help with?
Ask questions about your ingested documents.
Checking your knowledge base…
Example queries
How does RAG combine retrieval and generation?
Explain vector embeddings and cosine similarity
What chunking strategies and overlap are used for RAG?
What are HNSW, hybrid search, and BM25?
Which metrics evaluate RAG retrieval and answers?
Summarize main topics in my knowledge base
Capabilities
Query your ingested documents using natural language
Retrieve relevant context using semantic search
Generate answers augmented with retrieved knowledge
View citations and metadata for transparency
Limitations
May occasionally generate incorrect information
Quality depends on ingested document accuracy
Retrieval accuracy depends on embedding quality
Costs may vary based on query complexity
