New framework measures AI reasoning quality beyond just correct answers

Original: Measuring Reasoning Quality in LLMs: A Multi-Dimensional Behavioral Framework

Source: arxiv.org ↗