Partial Regurgitation and how LLMs really work
We recorded 4 claims from this post. 1 was checked against evidence. 3 can’t be checked. Read the post at its source.
Claims from this post (4)
In the order they appear in the post. Each quotation is the speaker’s words in this post. A label such as “Supported” or “Can’t be checked” is the claim’s current assessment; its page has the evidence and the reasons.
Assessment: Can’t be checkedSaid
Observation
“Neural nets do store weights, but that doesn’t mean that they know what they are talking about”
Our reading: Storing weights does not mean neural nets know what they are talking about.
No horizon
Assessment: UnresolvedSaid
Causal claim
“By clustering piles of similar things together in a giant n-dimensional space, they pretty much automatically become synonym and pastische machines, regurgitating a lot of words with slight paraphrases while adding conceptually little, and understanding even less.”
Our reading: Because LLMs cluster similar things in a high-dimensional space, they become synonym and pastiche machines that regurgitate with slight paraphrase, adding little and understanding less.
No horizon Evidence as of
Limitation: Verbatim recall and memorization from similar sequences do not establish how often models produce close paraphrases in ordinary use [e1, e3].
Assessment: Can’t be checkedSaid
Forecast
“Partial regurgitation, no matter how fluent, does not, and will not ever, constitute genuine comprehension.”
Our reading: Partial regurgitation, no matter how fluent, does not and will not ever constitute genuine comprehension.
No horizon
Qualifier: no matter how fluent
Assessment: Can’t be checkedSaid
Forecast
“Getting to real AI will require a different approach.”
Our reading: Getting to real AI will require a different approach from the partial-regurgitation approach of current LLMs.
No horizon