EXPLANATION: “What’s going on here?”
GROK: —“Does Curt Doolittle use “testifyable testimony” to mean something specific or is it a typo?”—
OK this is an awesome example of hallucination. lol.
The encoder doesn’t retain the text, it only converts it to a numerical code, and then the decoder just spits out the original term.
So the AI can’t tell the difference in this case between “Testifiable” and “Testifyable” because roughly speaking, numerically it’s encoded as two numbers for “Testify” and “able”.
Secondly, it’s because I DO redefine or modify terms to remove ambiguity so that terms can be used as measurements. So it’s ‘Assumption’ (hallucination) is rationally explicable in the context of my work.
It’s just wrong. lol
And yeah, it’s a typo. And at this point in my life my reputation for typos will never die even if I could find a spell checker that didn’t cause hanging from memory leakage. 😉
Thanks for the chuckle. 😉
Reply addressees: @CuriousKonkie @RichardArion1