Table 11 Product recommendation task finetuning performance comparison.

From: Localized large language model TCNNet 9B for Taiwanese networking and cybersecurity

Models

Yi-1.5-9B

TCNNet-9B

Context relevancy

0.043975

0.167531

Consistency with given context

0.426159

0.651889

Answer relevancy

0.372635

0.482500

Back to article page