[ExI] all we are is just llms was: RE: e: GPT-4 on its inability to solve the symbol grounding problem

Giovanni Santostasi gsantostasi at gmail.com
Fri Apr 21 05:31:10 UTC 2023


Spike,
The closest thing is Tononi's Integrated Information but to measure this
it would require calculating the connectivity of the system and I think
this is basically impossible to do. In practice in humans, you can measure
Integrated Information indirectly by looking at complexity measures in the
EEG or fMRI activity as a proxy. I think at most one can give a battery of
cognitive tests to the AI and see how it compares to humans. This is has
been done and AI is showing capabilities that are similar to humans of
different ages according to what is tested (for example GPT-4 has the
theory of mind of a 9-year-old).
Giovanni


On Thu, Apr 20, 2023 at 10:25 PM spike jones via extropy-chat <
extropy-chat at lists.extropy.org> wrote:

>
>
> Regarding measuring GPT’s intelligence, this must have already been done
> and is being done.  Reasoning: I hear GPT is passing medical boards exams
> and bar exams and such, so we should be able to give it IQ tests, then
> compare its performance with humans on that test.  I suspect GPT will beat
> everybody at least on some tests.
>
>
>
> So… now we need to assume for the sake of investigation that consciousness
> and intelligence are (somehow) separable mental processes.  I don’t know of
> any test analogous to an IQ test for consciousness.  Does anyone here know
> of such a thing?
>
>
>
> After all my posts, I think I have just argued myself back to where I
> started: without some objective way to measure consciousness, we are doing
> little more here than debating the definition of an ill-defined term.
>
>
>
> Damn.
>
>
>
> spike
>
>
>
>
> _______________________________________________
> extropy-chat mailing list
> extropy-chat at lists.extropy.org
> http://lists.extropy.org/mailman/listinfo.cgi/extropy-chat
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20230420/51e569ba/attachment.htm>


More information about the extropy-chat mailing list