A chunk of the LLM response.
The token for the next set of chunks. Use the value returned in the previous response in the next request to retrieve the next set of chunks.
The references used to generate the LLM response.