1 paper · 1 filter
Carson Rodrigues, Oysturn Vas
A common claim is that zero-shot large language models (LLMs) can replace fine-tuned NLU classifiers for intent detection. We test this claim head-to-head and find that the honest…