> mostly ... but I have been busy on different matters: in absence (I presume) of a model of their model of reasoning, has some metrics, some measurement, be

> mostly ... but I have been busy on different matters: in absence (I presume) of a model of their model of reasoning, has some metrics, some measurement, be produced to understand the deep reliability (e.g. absence of hallucination, logical consistency etc.) of LLMs?