Extracting Reasoning Models for Quality Improvement in Responses

·

Can anyone help: (Alban mentioned Peter): From my question in the "Dust Session for Builders": Topic Reasoning Models... "How can I extract the chain of thought (reasoning stream) after the response so that we can log it and review for further improvement and quality check? " Merci

2 comments

Alban
·
·
in the conversation endpoint you have a "chainOfThough" attribute here: https://docs.dust.tt/reference/post_api-v1-w-wid-assistant-conversations
Alexander AP Preis
·
·
Perfect. That helps Alban. Much appreciated and good builder session yesterday. best, Alex