Stop arguing! The official Chinese name for Token has been decided.

Ask AI · Why did the official choose “词元” instead of “智元”?

The Chinese name for Token seems to have finally been decided!

Today, the Chinese government website forwarded an article from the People’s Daily quoting the National Bureau of Statistics, showing that our daily token usage has surpassed 140 trillion.

140 trillion in data volume is quite staggering, but the naming of Token in Chinese behind this news is even more interesting, seemingly able to quell recent disputes in the industry regarding the Chinese name for Token.

This also means that the official Chinese name for Token has been established as “词元”.

The data released by the National Bureau of Statistics, published by the People’s Daily, and forwarded by the Chinese government website, confirms that the Chinese name for Token is set.

The naming controversy was sparked by an article from Tencent Research Institute, where scholar Yang Bin proposed the translation “模元”. This stirred up a wave of responses, with AI leaders on social media and industry groups giving their preferred answers. Soon, the name “智元” emerged and quickly gained widespread recognition.

You Yang, the president of the National University of Singapore and founder of Luchen Technology:

“The Chinese name for Token is ‘智元’, I wonder if the founder of ‘新智元’ has time-traveled back.”

Independent scholar and former associate professor of the History of Science at Tsinghua University, Hu Yilin:

“It’s really not as good as ‘智元’. The unit in computing is a byte, and Token is the computational unit of human intelligence. Token is not a concept of general computation; it is a concept of ‘general intelligence’, and ‘智’ hits the nail on the head more than ‘通’.”

Wang Xiaochuan, founder of Baichuan Intelligent:

“Calling it ‘智元’ is quite good.”

Senior open-source figure Chen Xu:

“In recent days, various x元 have been proposed, and I feel that ‘智元’ can resonate with me, meeting the requirements of elegance and clarity.”

This has sparked a large-scale discussion, with some also taking the opportunity to joke around.

But in the end, the official decision is to call it “词元”!

According to People’s Daily’s explanation, “词元” is the smallest unit for AI to understand human language; the units segmented by token can sometimes be characters, sometimes words, and “词” better encompasses this range.