<div dir="ltr"><b><br>My argument is about large language models. LLMs, in the purest sense of that term,, are nothing like such a system. They have no eyes, no ears, no senses whatsoever to register anything outside of the text. They are trained only on symbolic text material. From their point of view, (so to speak), the corpus of text on which they are trained is the entire universe. <br></b>I let Ben elaborate but I think given we are aligned on many things I do understand what he tries to communicate with that diagram. <br>1) It is an illusion to think that there is a single grounding object when we think about a word, even what seems a very solid and concrete word as "apple". It is actually a complex network of sensations, abstractions, experiences, and different types of apples that are abstracted away into a few forms in our heads. The word apple is really a complex network. It doesn't matter what the linguists say because most of these people are humanists with very little understanding of neuroscience or other sciences and in particular zero understanding of how LLMs work or other advanced concepts in computer science. Their "science" is antiquated and we need another science of how language works.<br>2) LLMs work very similarly to human brains because the connections above are also present in the LLMs. They do not "refer" to sensory experiences but other words or symbols (or clusters of words) but link enough of these words and also loop them back to the original word and you get meaning exactly in the same way the brain does it. The meaning IS THE COMPLEX CONNECTIONS. <br>3) The above idea is not just a good theoretical framework that can show it works in many different contexts (mathematics, logic, computing) but also it seems to work in real life given GPT-4 really understands (like some of my experiments and others) show. If you do cognitive tests that are used to test humans then GPT-4 has a similar performance to humans at different levels of development depending on the task. It is an empirical fact and it cannot be really denied and only excuses can be made to dismiss this evidence. <br><br>Giovanni </div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Mon, Apr 17, 2023 at 12:29 PM Gordon Swobe via extropy-chat <<a href="mailto:extropy-chat@lists.extropy.org">extropy-chat@lists.extropy.org</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div dir="ltr"><div dir="ltr"><br></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Mon, Apr 17, 2023 at 11:54 AM Ben Zaiboc via extropy-chat <<a href="mailto:extropy-chat@lists.extropy.org" target="_blank">extropy-chat@lists.extropy.org</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
<div>
On 17/04/2023 17:49, Gordon Swobe wrote:<br>
<blockquote type="cite">It
is mind-blowing that even after all my attempts to explain
Linguistics 101, you guys still fail to understand the meaning of
the word "referent." </blockquote>
<br>
Then, please, explain it to us! (using the relevant discipline, i.e.
neuroscience, not linguistics). Presumably you think linguistics is
relevant because these systems we're discussing are called "Large
Language Models"</div></blockquote><div><br>Yes, I certainly do believe that linguistics, the scientific study of human language, can tell us something about models of human language. <br> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div>, but having 'language' in the name doesn't mean
that 'language' explains how they work. It's all about
signal-processing. Unless you think that only brains use
signal-processing and not LLMs, or vice-versa.</div></blockquote><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div>
<br>
So try explaining it in terms that we will understand. Presumably my
diagram from several posts ago:<br>
<br>
<b>'Grounded' concept</b><br>
<img src="cid:1879070d0cdbc28a6cb1" alt="" width="91" height="102"><br>
<i>(The block is a 'real-world' object. What this actually means, I
have no good idea)</i><br>
<br>
<br>
is inaccurate, and not what you mean at all, so maybe you can
explain it in terms of the other diagram:<br>
<b><br>
</b><b>Linked concept (very simplified)</b><br>
<img src="cid:1879070d0cdfda45a572" alt="" width="297" height="182"><br>
<i>(The blue ovals are myriad other concepts, memories, sensory
inputs, tokens, etc.)</i><i><br>
</i><i> Of course, a real diagram of the links would be so dense as
to be unreadable. The other ovals would be linked to each other as
well as to the centra oval, and it would be 3D with links
extending out, as far as the sensory organs, which transduce
specific aspects of the 'real world' such as temperature changes,
specific frequencies of sound, etc.</i><br>
<br>
Or, if not, then at least in terms of the things that we know to be
true about the brain. i.e., nothing in the brain has access to
anything except signals from other parts of the brain, and signals
from the sense organs, coded in the 'language' our brains use: spike
trains. I'm sure you know what spike trains are (giving a stern
warning look at Spike here, finger to lips).<br>
<br>
And, again if you disagree with the statement above, please give
your reasons for disagreeing.</div></blockquote><div><br>Let us say that the diagram above with a "myriad of other concepts etc" can accurately model the brain/mind/body with links extending to sensory organs and so on. Fine. I can agree with that at least temporarily for the sake of argument, but it is beside the point. <br><br>My argument is about large language models. LLMs, in the purest sense of that term,, are nothing like such a system. They have no eyes, no ears, no senses whatsoever to register anything outside of the text. They are trained only on symbolic text material. From their point of view, (so to speak), the corpus of text on which they are trained is the entire universe. <br><br>The LLM has no way to understand the meaning of the symbol "potato," for example -- that is, it has no way to ground the symbol "potato" -- except in terms of other symbols in the text that it also has no way to ground or understand. The LLM is, so to speak, trapped in a world of symbolic forms with no access to the meanings of those forms. This does not mean it cannot manipulate these forms in ways that mimic human understanding, as it was trained on a vast amount of formal material written in ways that we find understandable and knows the statistics and patterns of English language use -- but the meanings of the symbolic forms in its inputs and outputs are assigned by us, the human operators. <br><br>We can and do assign meanings to these formal symbols, as unlike the LLM, we do have access to the world outside of the corpus and so understand the meanings of the formal symbols. This can create the appearance that it is the LLM that understands and conveys the meanings, but this is an illusion. We are projecting our own mental processes onto the LLM.<br>...<br>Now, I understand that from here, me might get into theoretical discussions about AI robots with electronic sensors and multi-modal LLMs and so on, but it would be helpful if people could at least understand that LLMs, per se, are unconscious with no true understanding of the world exactly as GPT-4 professes to be when asked if it can ground symbols for itself.<br><br>-gts<br><br><br><br><br> </div></div></div>
_______________________________________________<br>
extropy-chat mailing list<br>
<a href="mailto:extropy-chat@lists.extropy.org" target="_blank">extropy-chat@lists.extropy.org</a><br>
<a href="http://lists.extropy.org/mailman/listinfo.cgi/extropy-chat" rel="noreferrer" target="_blank">http://lists.extropy.org/mailman/listinfo.cgi/extropy-chat</a><br>
</blockquote></div>