You are using a pre-trained LLM to analyze customer emails and classify each email based on the main issue (e.g., billing, technical support, product information). Which approach would be most effective to minimize training time while maintaining high accuracy?

Question

Answer

Utilizing token classification to label individual words in the emails related to each category.

Answer

Applying a few-shot learning approach using labeled examples for each issue type within the prompt.

Answer

Building a custom model from scratch specifically trained on your company’s emails.

Answer

Fine-tuning the entire LLM on the labeled email data for each category.

Crowdly

You are using a pre-trained LLM to analyze customer emails and classify each ema...

Want instant access to all verified answers on softserve.academy?