split out thinking/reasoning

[jwijffels]

Thanks for providing chatlas

Recent models like deepseek-r1 or deepscaler have a thinking/reasoning part
Is there a way to extract this information from chatlas in order split out the thinking/reasoning part and the final answer.

e.g. in lmstudio you would have a reasoning section which you can uncollapse

How can you extract the thinking/reasoning part of the output with chatlas? I'm using ChatOllama

[brooklynbagel]

I think this would be really nice especially in the context of a chatbot with chat.app(). If there were some options for hiding thinking/reasoning, or collapsing it, I think that'd be nicer UI than just printing it.