In this example, we learn how to integrate LlamaGuard, a safeguard LLM with NeMo Guardrails into our custom NVIDIA NIM-powered RAG pipeline to have more secure, moderated responses for user queries. We also learn how to integrate the AlignScore metric to check the factual consistency of the LLM response with the retrieved chunks from the vector database.
In this example, we learn how to integrate LlamaGuard, a safeguard LLM with NeMo Guardrails into our custom NVIDIA NIM-powered RAG pipeline to have more secure, moderated responses for user queries. We also learn how to integrate the AlignScore metric to check the factual consistency of the LLM response with the retrieved chunks from the vector database.