A definition of "plausible" is "apparently reasonable and credible, and therefore convincing".
In what limit does "apparently reasonable and credible" diverge from "true"?
We'd make the LLM not lie if we could. All this "plausible" language constitutes practitioner weasel words. We'd collectively love if the LLMs were more truthful than a 5-year-old.
In what limit does "apparently reasonable and credible" diverge from "true"?
We'd make the LLM not lie if we could. All this "plausible" language constitutes practitioner weasel words. We'd collectively love if the LLMs were more truthful than a 5-year-old.