You instruct an LLM (AI) to classify support tickets strictl…
You instruct an LLM (AI) to classify support tickets strictly as ‘Billing’, ‘Technical’, or ‘General’. In your few-shot examples, you provide 4 tickets labeled ‘Bill’, 1 labeled ‘Tech’, and 0 for ‘General’. Why is the LLM more likely to output invalid categories?
Read Details