Photo/VCG

Feb.21 (NBD) -- Chinese researchers from Fudan University recently released the country's first large language model chatbot "MOSS", which gained wide attention and crashed on Monday due to large traffic.  

The model is capable of performing tasks such as dialogue generation, programming and fact-based question answering, which will facilitate the exploration and application of big language models. 

On February 21, MOSS official website announced that MOSS is still a very immature model, and there is still a long way to go.

We are an academic research laboratory, which cannot make a model similar to ChatGPT's capabilities. MOSS just wants to explore and verify the technical route of ChatGPT on a scale of tens of billions of parameters, said the website.

"Our original idea was to conduct internal testing of MOSS so that we could further optimize it. We did not expect to attract such a large amount of attention. Our computing resources are not enough to support such a large amount of visits. As an academic team, we have no relevant engineering experience, resulting a very bad experience and first impression, and the MOSS team would like to express its sincere apologies to everyone."

The announcement also explained that the name comes from "The Wandering Earth".

It is learned that after MOSS completes the preliminary verification, the team will open MOSS experience, code, and model parameters.