NCSOFT launches evaluation model to verify performance of AI LLMs

NCSOFT Corp. announced on Monday that it has launched VARCO Judge LLM, the first evaluation model in South Korea to verify the performance and capabilities of artificial intelligence (AI) large language models (LLMs).
 
VARCO Judge LLM is an evaluation model that checks how quickly and accurately other LLMs perform tasks. 

With this model, companies creating AI-based services can quickly compare and evaluate the quality of various LLMs and adopt the best model for their services.

R&D companies can also verify the performance level of their LLMs to demonstrate performance advantages or quickly identify and strengthen weaknesses. 

NCSOFT explained that VARCO Judge LLM has the highest performance among models in the same class, and plans to use it to improve the quality of its own LLM 'VARCO'.

“In the rapidly evolving AI market, services that select and apply the optimal model for each industry are becoming increasingly important,” said Lee Yeon-su, head of NCSOFT's research division.

”VARCO Judge LLM will not only improve the quality of existing LLM-based services, but will also become an indispensable tool for the AI business,” she added.

Write to Seung-Woo Lee at leeswoo@hankyung.com