2 comments of 3

[ 3.0 ms ] story [ 16.7 ms ] thread
Interesting approach with the cascade. How do you decide when to escalate from fuzzy matching to LLM?