Re:The Geometry of Truth: Do LLM's Know True and False
by ngneer on Thursday October 19, 2023 @07:56PM
(#37950417)
Right, but that ignores the key piece, which is whether such systems can infer the equivalence relation denoted by "is". A is A. A is B implies B is A. A is B and B is C implies A is C. When a system sees that A is B, yet cannot infer that B is A, it exhibits an asymmetry, which is interesting. Whether this asymmetry exists in the underlying language is unclear.
[ Parent
| on HN ]
Re:The Geometry of Truth: Do LLM's Know True and False
by xcv123 on Thursday October 19, 2023 @08:31PM
(#37950687)
> When a system sees that A is B, yet cannot infer that B is A
So the issue is that they cannot infer that general rule due to a fundamental limitation of the transformer LLM architecture, not just a training data issue? I skimmed the paper and it seems to be the case.
Re:The Geometry of Truth: Do LLM's Know True and False
by dr_kiszonka on Thursday October 19, 2023 @09:27PM
(#37951185)
> A is B implies B is A.
I get the overall idea, but this statement isn't always true, right?
Re:The Geometry of Truth: Do LLM's Know True and False
by xcv123 on Thursday October 19, 2023 @10:45PM
(#37951813)
Correct. Not in natural language. These are natural language systems, not logic systems.
"The sky is blue" does not imply "blue is the sky".
Re:The Geometry of Truth: Do LLM's Know True and False
by TeMPOraL on Friday October 20, 2023 @05:34AM
(#37954018)
There's an asymmetry here, too. "The sky is ${color}", with no extra context, has one obvious answer for us living here, today. Whereas for "Blue is the color of ${thing}", with no extra answer, there's an insane amount of equally sensible substitutions for ${thing}. Without extra content, the model has no reason to privilege "sky" over any other equally valid answer.