The ideal here : LLM will give out constranst and translate the constranst to the z3 SMT language, then give out feedback loop between LLM model and neuro-symbolic reasoning to give out better reasioning
looking for feedback and comments from HNs
Thanks and best regards Kevin