"LLMs are based on this transformer architecture, which notably is not actually reading text. What happens when you input a prompt is that it's translated into an encoding," Matthew Guzdial, an AI researcher, explained. This indicates that while AIs can process and respond to language, they don't have a deep understanding of it like humans do, as they operate primarily on numerical representations rather than engaging with letters or sounds critically.
"When it sees the word 'the,' it has this one encoding of what 'the' means, but it does not know about 'T,' 'H,' 'E.'" This highlights a fundamental limitation in how LLMs interpret language. They may recognize patterns and construct responses, but their lack of comprehension regarding the individual elements that form words is a significant barrier to genuine understanding.
Collection
[
|
...
]