See How h2oGPTe RAG Cites Its Sources
- Generative AI
- Enterprise h2oGPTe
This lesson highlights h2oGPTe's reference capability when querying uploaded documents inside a Collection, showing how the platform cites specific web pages behind each answer. It demos asking about the first act of the musical Hamilton, with the RAG response summarizing the source Wikipedia page and linking back to the exact references used. It then repeats the same question using only the standalone large language model, without RAG, producing a generic description of what typically happens in a play's first act. It closes by contrasting the two answers to show why sourced, referenced responses matter.
Key takeaways
- How h2oGPTe cites the specific web pages behind an answer when querying documents in a Collection
- What changes when the same question runs through RAG versus a standalone LLM with no document access
- Why a standalone LLM gives a generic answer about a play's first act without knowing which play is meant
- The reference view lets you trace an answer back to the exact source material it drew from
Start owning your AI