I don’t know why you thought that. LLMs split your question into separate words and assigns scores to those words, then looks up answers relevant to those words. It has no idea of how those words are relevant to each other. That’s why LLMs couldn’t answer how many "r"s are in “strawberry”. They assigned the word “strawberry” a lower relevancy score in that question. The word “rescue” is probably treated the same way here.
I don’t know why you thought that. LLMs split your question into separate words and assigns scores to those words, then looks up answers relevant to those words. It has no idea of how those words are relevant to each other. That’s why LLMs couldn’t answer how many "r"s are in “strawberry”. They assigned the word “strawberry” a lower relevancy score in that question. The word “rescue” is probably treated the same way here.
I know how it works, thank you 😚