4 ms·

The burden isn't on us, it's on people making claims that LLMs do possess theory of mind. You're the one who has to prove it, not us.

sebzim4500 2y ago
If you use any of the conventional tests that exist of theory of mind (most famously the Sally-Anne Test [1] but also the others) then SOTA reasoning models will get near 100%. Even if you try to come up with similar questions which you expect not to be in the training set they will still get them right. In the absence of any evidence to the contrary, this is convincing evidence in my opinion. [1] https://en.wikipedia.org/wiki/Sally%E2%80%93Anne_test https://en.wikipedia.org/wiki/Sally%E2%80%93Anne_test
- esafak 2y ago
  Maybe having a theory of mind isn't the big deal we thought it was. People are so conditioned to expect such things only from biological lifeforms, where theory of mind comes packaged with many other abilities that robots currently lack, that we reflexively dismiss the robot.
- zipy124 2y ago
  That same source you link says that your view of 100% is not accepted as a consesus: "... GPT-4's ability to reason about the beliefs of other agents remains limited (59% accuracy on the ToMi benchmark),[15] and is not robust to "adversarial" changes to the Sally-Anne test that humans flexibly handle.[16][17] While some authors argue that the performance of GPT-4 on Sally-Anne-like tasks can be increased to 100% via improved prompting strategies,[18] this approach appears to improve accuracy to only 73% on the larger ToMi dataset."
  - sebzim4500 2y ago
    We are talking about frontier models not GPT-4
    - zipy124 2y ago
      Yes but I am using the same source the commenter used to backup their figure, merely saying look your source doesn't say what you claim it does. If they wanted to talk about frontier models maybe they should have cited a link to talking about frontier models performance.
  - CamperBob2 2y ago
    In basically every case, by the time a claim like that is stated in a paper like that, it's obsolete by the time it's published, and ancient history by the time you use it to try to win an argument.
    - zipy124 2y ago
      My point is merely if you are going to make an argument using a source, the source should support your argument. If you say "the accuracy of an llm on task 1 is 90% [1]" and when you go to [1] it says the accuracy of an llm on task 1 is 50%, but some sources say with better prompts you can get to 90%, but when extended to a larger data-set for task 1, performance drops to 70%" then just quoting the highest number is mis-leading.
neckardt 2y ago
I’ll bite. Theory of mind (ToM) is the ability to understand that other people have different thoughts, feelings, and intentions than you do. So let’s run through a basic scenario with ChatGPT 4o: Prompt 1: Briefly respond: You’re driving on the highway when the driver in the lane next to you swerves into your lane. Why might this be? Response 1: The driver may be distracted, drowsy, intoxicated, experiencing a medical emergency, avoiding an obstacle, or not checking their blind spot before changing lanes. Prompt 2: What might that person be feeling or intending? Response 2: The driver might be feeling distracted, rushed, unaware, or panicked. They may intend to change lanes without realizing your presence, be reacting to something on the road, or simply have made an unintentional mistake. Prompt 3: You notice there was no obstacle. Any other reason the driver might do this? Response 3: Yes, the driver might be experiencing fatigue, intoxication, a medical issue (such as a seizure or dizziness), distraction (like using a phone or adjusting the radio), aggressive driving, or simply failing to check their blind spot properly. Seems to me that ChatGPT has a good idea of what another driver might be feeling while on the road. It gets this right for most other cases too. Could you provide a counterexample for when it fails to understand what other people may be feeling in certain scenarios?
- Arkhaine_kupo 2y ago
  Well I think a fairly easy counterexample is that if you keep going, and the information falls outside its context window then the questions you had about the driver stop making sense. So lets say you ask about a driver who swerved, and the response by chatgpt correctly talks about his possible state. Then you talk more about cars, about road conditions etc and once the context is far away enough you ask, what was the driver doing again? And you are prompted with a hallucination. Because there is no state in the llm, there is no "driver", there is contextual statistically accurate responses but you hold a "driver" object in your mind while maintaining the conversation, the llm doesn't. Its like a conversation with someone with short term memory loss like in memento
  - dauhak 2y ago
    And people with short term memory loss nevertheless have theory of mind just fine. Nothing about LLM's dropping context over big enough windows implies they don't have theory of mind, it just shows they have limitations - just like humans even with "normal" memory will lose track over a huge context window. Like there are plenty of shortcomings of LLMs but it feels like people are comparing them to some platonic ideal human when writing them off
farts_mckensy 2y ago
Prove that you possess "theory of mind."

sebzim4500 2y ago

If you use any of the conventional tests that exist of theory of mind (most famously the Sally-Anne Test [1] but also the others) then SOTA reasoning models will get near 100%. Even if you try to come up with similar questions which you expect not to be in the training set they will still get them right. In the absence of any evidence to the contrary, this is convincing evidence in my opinion. [1] https://en.wikipedia.org/wiki/Sally%E2%80%93Anne_test https://en.wikipedia.org/wiki/Sally%E2%80%93Anne_test

esafak 2y ago
Maybe having a theory of mind isn't the big deal we thought it was. People are so conditioned to expect such things only from biological lifeforms, where theory of mind comes packaged with many other abilities that robots currently lack, that we reflexively dismiss the robot.
zipy124 2y ago
That same source you link says that your view of 100% is not accepted as a consesus: "... GPT-4's ability to reason about the beliefs of other agents remains limited (59% accuracy on the ToMi benchmark),[15] and is not robust to "adversarial" changes to the Sally-Anne test that humans flexibly handle.[16][17] While some authors argue that the performance of GPT-4 on Sally-Anne-like tasks can be increased to 100% via improved prompting strategies,[18] this approach appears to improve accuracy to only 73% on the larger ToMi dataset."
- sebzim4500 2y ago
  We are talking about frontier models not GPT-4
  - zipy124 2y ago
    Yes but I am using the same source the commenter used to backup their figure, merely saying look your source doesn't say what you claim it does. If they wanted to talk about frontier models maybe they should have cited a link to talking about frontier models performance.
- CamperBob2 2y ago
  In basically every case, by the time a claim like that is stated in a paper like that, it's obsolete by the time it's published, and ancient history by the time you use it to try to win an argument.
  - zipy124 2y ago
    My point is merely if you are going to make an argument using a source, the source should support your argument. If you say "the accuracy of an llm on task 1 is 90% [1]" and when you go to [1] it says the accuracy of an llm on task 1 is 50%, but some sources say with better prompts you can get to 90%, but when extended to a larger data-set for task 1, performance drops to 70%" then just quoting the highest number is mis-leading.

neckardt 2y ago

I’ll bite. Theory of mind (ToM) is the ability to understand that other people have different thoughts, feelings, and intentions than you do. So let’s run through a basic scenario with ChatGPT 4o: Prompt 1: Briefly respond: You’re driving on the highway when the driver in the lane next to you swerves into your lane. Why might this be? Response 1: The driver may be distracted, drowsy, intoxicated, experiencing a medical emergency, avoiding an obstacle, or not checking their blind spot before changing lanes. Prompt 2: What might that person be feeling or intending? Response 2: The driver might be feeling distracted, rushed, unaware, or panicked. They may intend to change lanes without realizing your presence, be reacting to something on the road, or simply have made an unintentional mistake. Prompt 3: You notice there was no obstacle. Any other reason the driver might do this? Response 3: Yes, the driver might be experiencing fatigue, intoxication, a medical issue (such as a seizure or dizziness), distraction (like using a phone or adjusting the radio), aggressive driving, or simply failing to check their blind spot properly. Seems to me that ChatGPT has a good idea of what another driver might be feeling while on the road. It gets this right for most other cases too. Could you provide a counterexample for when it fails to understand what other people may be feeling in certain scenarios?

Arkhaine_kupo 2y ago
Well I think a fairly easy counterexample is that if you keep going, and the information falls outside its context window then the questions you had about the driver stop making sense. So lets say you ask about a driver who swerved, and the response by chatgpt correctly talks about his possible state. Then you talk more about cars, about road conditions etc and once the context is far away enough you ask, what was the driver doing again? And you are prompted with a hallucination. Because there is no state in the llm, there is no "driver", there is contextual statistically accurate responses but you hold a "driver" object in your mind while maintaining the conversation, the llm doesn't. Its like a conversation with someone with short term memory loss like in memento
- dauhak 2y ago
  And people with short term memory loss nevertheless have theory of mind just fine. Nothing about LLM's dropping context over big enough windows implies they don't have theory of mind, it just shows they have limitations - just like humans even with "normal" memory will lose track over a huge context window. Like there are plenty of shortcomings of LLMs but it feels like people are comparing them to some platonic ideal human when writing them off

farts_mckensy 2y ago

Prove that you possess "theory of mind."