Comment on Honda: 2 years of ml vs 1 month of prompting - heres what we learnedparentComments−suriya-ganesh9moI have. For a similar-ish task.LLMs still beat a clarifier, because they're able to extract more signals than a text embedding.It's very difficult to beat an LLM + prompt in terms of semantic extraction.−cestith9moIs it hard to beat on performance for the cost, though?
Comments
I have. For a similar-ish task.
LLMs still beat a clarifier, because they're able to extract more signals than a text embedding.
It's very difficult to beat an LLM + prompt in terms of semantic extraction.
Is it hard to beat on performance for the cost, though?