Hacker Read

jfk13 · 2022-12-30 12:32:34

And I don't see any fundamental reason to believe that language models can learn or reason at all.

PaulDavisThe1st | karma 19282 | avg karma 2.7 · | 2023-04-30 23:56:33

Not bad, but:

> Language models also have limited reasoning abilities.

They have no reasoning abilities at all.

reply

wwarner | karma 2361 | avg karma 3.19 · | 2024-05-21 16:07:29

well that's exactly the point -- no such result is available for language models.

jtc331 | karma 1343 | avg karma 3.56 · | 2024-02-25 13:02:06

Is this a way of saying that large language models don’t have the concept of “I don’t know”?

jameshart | karma 19910 | avg karma 4.65 · | 2022-12-12 07:39:38

Hang on - I thought the consensus among ML experts was that language models don’t ‘know’ anything?

sboomer | karma 10 | avg karma 0.59 · | 2023-03-15 00:48:27

I don't know much about language models, but don't they just have an understanding/knowledge of patterns between words, and don't have the reasoning capability at all?

rytill | karma 681 | avg karma 1.83 · | 2021-09-13 17:47:59

I'm one of those people, minus the straw man of it "understanding" anything. Why do you assume a language model can't learn some rudimentary reasoning from predicting the next token?

EMCymatics | karma 72 | avg karma 0.4 · | 2023-06-07 23:34:36

There is more of a point to actual learning than to create a language model.

snek_case | karma 1555 | avg karma 4.02 · | 2022-03-15 20:28:14

That's because it's just a language model. It's been trained to find a probable completion to a piece of text, predict a likely next word. It's not trained for human interaction. It's not an agent. It has no motives or goals. It might seem like it does, but that's more of a side-effect of language modelling.

cdavid | karma 2079 | avg karma 2.78 · | 2012-11-02 08:04:14+00:00

I don't particularly have any reason to believe one way or the other. Certainly, the probabilistic models for language are created "out of the blue" without any attempt to model how human learn languages.

TchoBeer | karma 789 | avg karma 1.32 · | 2021-07-29 17:10:59

This is a useless hypothetical, no language models do that

kmoser | karma 766 | avg karma 1.7 · | 2022-12-30 12:09:33

Given that large language models don't have any actual knowledge (in a human sense) of the data they've been trained on other than raw statistics, can they be said to "reason" about anything?

akomtu | karma 1122 | avg karma 0.52 · | 2024-05-15 22:24:40

Everyone is trying to use Language Models as Reasoning Models because the latter haven't been invented yet.

wokwokwok | karma 5467 | avg karma 5.25 · | 2022-08-27 19:35:50

Language models show that language is predictive, not rule based.

Every bit of research in the last 30 years shows this.

If you want to ignore that and use rule based approaches, that’s fine, but you’re a) wrong and b) it doesn’t work.

/shrug

Of course people want rules, they think it’s a shortcut to learning.

…but there is no shortcut to learning. You have to do the hard work.

reply

user_named | karma 1745 | avg karma 2.26 · | 2023-04-06 09:41:17

Reflection is the thing. Language models don't reflect.

matthewfelgate | karma 381 | avg karma 0.68 · | 2023-08-31 09:10:22

As a language model I can't help you with that.

scotty79 | karma 14043 | avg karma 1.62 · | 2023-02-15 10:38:14

It's a language model. It models language not knowledge.

andrepd | karma 17410 | avg karma 3.19 · | 2022-12-03 10:12:19

Of course, it's a language model with 0 semantic knowledge about its output.

jrflowers | karma 3821 | avg karma 2.82 · | 2023-07-03 12:08:35

Nobody has been convinced to fly an airplane into a building by a language model either.

simonw | karma 58201 | avg karma 7.31 · | 2023-03-12 12:48:42

Large language models are not search engines, even though a lot of people seem determined to treat them as such.

I don't think that blog post's excercise is a particularly useful demonstration of anything.

reply