Tag
This position paper argues that AI reasoning lacks clear operational definitions, undermining evaluation validity, and proposes defining reasoning as a learnable rule-based process with a checklist for research best practices.