Do not require reasoning but just the ouput
#19
by
ameyv6
- opened
I want the LLM to output the response as per a provided JSON schema, is there a way to do this?
The model returns very verbose output with text before and after the JSON which is capped off with a ending token. Surprisingly there is no token at the beginning of the thinking/reasoning text that the model outputs.
Can we turn off the thinking from the model?
