Yandex Unveils YandexGPT 2 Language Model, Delivering 67% Better Responses

Visit Us
Follow Me
Photo by Juja Han on Unsplash

Yandex has released the second version of its large language model, YandexGPT. The initial edition of this model was introduced in May of this year. Yandex’s specialists have compared the responses generated by both versions of YandexGPT, taking into account parameters such as completeness, factual accuracy, grammatical and stylistic correctness, among others. It turns out that YandexGPT 2 provides better responses than its predecessor in an average of 67% of cases:

  • Text generation: 69%.
  • Text summarization and analysis: 68%.
  • Idea generation: 66%.
  • Styling for specific audiences or characters: 62%.
  • Fact-based queries: 62%.

YandexGPT 2 has the ability to adapt its responses for different audiences. For instance, it can explain complex terms to a child using simple language or rephrase a director’s directive into a clear letter for employees.

The progress achieved with YandexGPT 2 is attributed to an increase in the model’s parameter count and improvements in the training process. Additionally, developers expanded the dataset used for training and enhanced their quality control measures.

For those interested in trying YandexGPT 2, it can be accessed in a special “Alice” mode, which is initiated with the command “Alice, let’s come up with…” The neural network is also operational in the “Yandex” app, “Yandex Browser,” on the website, and on the company’s devices.

Author Profile

Vasyl Kolomiiets
Vasyl Kolomiiets
I'm Vasyl Kolomiiets, a seasoned tech journalist regularly contributing to global publications. Having a profound background in information technologies, I seamlessly blended my technical expertise with my passion for writing, venturing into technology journalism. I've covered a wide range of topics including cutting-edge developments and their impacts on society, contributing to leading tech platforms.

You may also like...

Leave a Reply

Your email address will not be published. Required fields are marked *