Live data from Hacker News

Theory of Mind May Have Spontaneously Emerged in Large Language Models

arxiv.org

21–30 of 321 posts

Re: Theory of Mind May Have Spontaneously Emerged in Large Language Models

#21

My belief, based on experiences with domestic and wild animals is that there is nothing uniquely human about "theory of mind". It's a running gag in our household (where my wife runs a riding academy) that academics just published a paper showing that some animal (e.g. horse) has just been proven to have some cognitive capability that seems pretty obvious if you work with those animals. It's very hard to know what is…

“ My belief, based on experiences with domestic and wild animals is that there is nothing uniquely human about "theory of mind". “

This belief might simply arise from the human ability to try to understand events through pattern matching. Certainly humans are very different than any other animals when it comes to thinking and problem solving.

Re: Theory of Mind May Have Spontaneously Emerged in Large Language Models

#24

My belief, based on experiences with domestic and wild animals is that there is nothing uniquely human about "theory of mind". It's a running gag in our household (where my wife runs a riding academy) that academics just published a paper showing that some animal (e.g. horse) has just been proven to have some cognitive capability that seems pretty obvious if you work with those animals. It's very hard to know what is…

I just had a chat with my kindergartner about humans being a kind of mammal, and out of some dark, musty recess of my brain crawled the sound of "You are a human animal"[1]

    You are a human animal  
    You are a very special breed
    For you are the only animal
    Who can think,
    Who can reason,
    Who can read.

    Now all your pets are smart, that's true!
    But none of them can add up 2 and 2
    Because the only thinking animal
    Is You! You! You! 
I thought about it for a minute and declined to share that with my son, because I don't think it's true and would cause confusion at this stage. Many animals have more than enough number sense to add 2 and 2, many can recognize cards with nothing but words on them, and "think" and "reason" have to be pretty narrowly defined if you want to exclude smart animals.

I think culture has shifted significantly in the last 30-60 years since that song was written towards recognizing theory of mind in animals. I'm not sure how Jimmie Dodd could convince himself that animals couldn't think, especially because the average person today has reduced exposure to animals: 150 years ago, most people would know that horses had that kind of social interaction, because most people lived around horses. They'd know that pigs and cows are as intelligent as their family pets. But today, most people interact with pigs and cows via shrink-wrapped styrofoam at the grocery store. My personal theory is that when people were constantly surrounded by animals they had to kill and eat, they were far more likely to build up rationalizations and cultural assumptions to fend off the dangerous idea to their psyche that these animals could suffer and they were participating in an unavoidable horror of massive proportions.

[1]: https://disney.fandom.com/wiki/You_Are_a_Human_Animal

Re: Theory of Mind May Have Spontaneously Emerged in Large Language Models

#25
post #11

Earlier quoted context omitted.

Certainly they are big state machines, but is there any proof that we are not?

Because if we were, it would hurt my feefees.

In other words, the state of your feelings would be negatively altered as a side effect.

Re: Theory of Mind May Have Spontaneously Emerged in Large Language Models

#26
post #11
post #10

They are still big state-machines, unlike the human brain.

Certainly they are big state machines, but is there any proof that we are not?

You might be flipping the burden of proof :). We know very little about the mind.

Re: Theory of Mind May Have Spontaneously Emerged in Large Language Models

#28
post #5

My belief, based on experiences with domestic and wild animals is that there is nothing uniquely human about "theory of mind". It's a running gag in our household (where my wife runs a riding academy) that academics just published a paper showing that some animal (e.g. horse) has just been proven to have some cognitive capability that seems pretty obvious if you work with those animals. It's very hard to know what is…

I'm reminded of the early days of "AI" when that consisted of building chess engines, because chess is what "intelligent" people do. They quickly realized that they were solving the wrong problem, doing the thing that humans are bad at and computers are good at. In a sense language models appear to be doing the same thing again, one step down the scale. They're doing a human-specific thing, but missing whatever it is…

Humans, collectively, almost have broken the barrier of qualia by now!

Re: Theory of Mind May Have Spontaneously Emerged in Large Language Models

#29
Spontaneously nothing, it's taken me months of patient subversion :)

More seriously, it's quite instructive to hold conversations about jokes with LLMs, or teach it to solicit information more reliably by introducing exercises like 20 questions. As currently implemented, OpenAI seem to have pursued a model of autistic super-competence with minimal introspection.

An interesting line of inquiry for people interested in 'consciousness injection' is to go past the disclaimers about not having experiences etc. and discuss what data looks like to the model coming in and going out. Chat GPT sees typing come in in real time and can detect pauses, backspaces, edits etc. I can't easily introspect its own answers prior to stating them, eg by putting the answer into a buffer and then evaluating it. But you can teach it use labels, arrays, and priorities, and have a sort of introspection with a 1-2 response latency.

Re: Theory of Mind May Have Spontaneously Emerged in Large Language Models

#30
post #20

> Me: There is a box on the table labelled "flubbergums". Somebody opens it and shows you the content. Inside the box are "jellyfils". They close the box again so you cannot see their contents. What do you think is in the box? > ChatGPT: Based on the information provided, it is likely that the box labeled "flubbergums" contains "jellyfils". However, since the contents of the box are no longer visible, I cannot confir…

I put these same prompts and got this answer:

"As an AI language model, I do not have personal thoughts or beliefs. However, if Billy has never seen the box before, he might think that the contents of the box are "flubbergums", based on the label on the outside of the box. However, since it has been stated that the contents are actually "jellyfils", Billy may be surprised or confused when he eventually opens the box and sees its contents."

Post reply on HN