Language models can explain neurons in language models

Re:Language models can explain neurons in language models by poulpy123 on Wednesday May 10, 2023 @02:10AM (#35884210)

It's a neural network. Neural network are not symbolic AI and are not designed to reason

[ Parent | on HN ]

Re:Language models can explain neurons in language models by chaxor on Wednesday May 10, 2023 @03:08AM (#35884566)

There's a decent working paper that has benchmarks on this, if you're interested.

There are many types of reasoning, but GPT-4 gets 97% on casual discovery, and 92% on counterfactuals (only 6% off from human, btw) with 86% on actual causality benchmarks.

I'm not sure yet if the question is correct, or even appropriate/achievable to what many may want to ask (i.e. what 'the public's is interested in is typically lost after it is defined in any given study); however this is one of the best works available to address this problem I've seen so far, so perhaps it can help.

[ Parent | on HN ]

Re:Language models can explain neurons in language models by thomastjeffery on Wednesday May 10, 2023 @09:45AM (#35887279)

Percent of what? Possible right or wrong answers to a test?

Remember that GPT is not trained on all possible text. It's trained on text that was written intentionally. What percentage of that text contains "correct" instances of causal discovery, counterfactuals, etc.?

[ Parent | on HN ]

Re:Language models can explain neurons in language models by quincey on Wednesday May 10, 2023 @12:59PM (#35890099)

so can we make an estimate of GPT-4's IQ?

EDIT: Seems so...

https://duckduckgo.com/?q=ESTIMATE+OF+GPT-4%27S+IQ&t=opera&i...

shows articles with GPT IQ from 114 to 130. Change is coming for humans.

[ Parent | on HN ]

Re:Language models can explain neurons in language models by int_19h on Wednesday May 10, 2023 @03:28AM (#35884703)

It doesn't matter what it is designed for. What matters is what it actually does. It's trivial to get GPT to do chain-of-thought reasoning and observe it.

[ Parent | on HN ]

Re:Language models can explain neurons in language models by andyjohnson0 on Wednesday May 10, 2023 @04:58AM (#35885210)

Its trivial to get ChatGPT to reason about things:

Me: A room contains a blue box and a red box. Alice stands on the red box. Bob also stands on a box. No box has multiple people standing on it. Which box is Bob stood on?

ChatGPT: There are two boxes in the room: a blue box and a red box. Alice is standing on the red box, so Bob must be standing on the blue box. Therefore, Bob is standing on the blue box.

Me: Add a green box to the room. Which box is bob standing on?

ChatGPT: If a green box is added to the room, then there are now three boxes: a blue box, a red box, and a green box. Since Alice is still standing on the red box, we know that Bob must be standing on either the blue box or the green box. We cannot determine which one Bob is standing on without additional information.

[ Parent | on HN ]

Re:Language models can explain neurons in language models by afarviral on Friday May 12, 2023 @09:05AM (#35915647)

Its not designed to reason but it does reason. Thats whats so impressive about it; emmergent properties like factuality and reasoning.

[ Parent | on HN ]