The community member is asking if there are ways to get responses at each step in the response synthesizers, such as seeing the output from the qa_template and comparing it to the final response after the refine_template. Another community member suggests using a tracing tool like Arize or simply printing the LLM inputs and outputs for debugging purposes. The original poster says they will give it a shot and thanks the other community member.
Are there ways to get responses at each step in the response synthesizers? For example, I have a refine_template and text_qa_template and would like to see what was generated in from qa_template and compare against the final response after refine_template
