Presenting Phoenix: The Multilingual LLM That Aims to Democratize ChatGPT

The power of Generative AI is unprecedented. But there is another avenue of AI that’s modestly doing the heavy lifting and making a mark with its spectacular human-like features and skills – these are called Large Language Models, also known as LLMs.

GPT-4, which is the latest entrant in the never-ending list of LLMs, has tremendously amplified ChatGPT’s utility in the last few months. Recently, a brand new model called Phoenix was introduced with the single aim of democratizing ChatGPT, especially in the countries where it is banned.

What is Phoenix?

Phoenix is a new model multilingual model. Its primary goal is to perform not just in Chinese and English, but other languages like Latin that are mostly limited by resources.

The reason behind its release is simple – to ensure that the audience can make use of ChatGPT where the local government or OpenAI has imposed restrictions. The bigger vision is to make it more accessible to a larger audience, minus any geographic or language constraints.

Recommended: Patient AI, the World’s First Use of Large Language Models and GPT

According to its author, Phoenix can be defined as the ‘first open-source, multilingual, and democratized ChatGPT model.’

The team further added that they believe Phoenix “will be beneficial to make ChatGPT more accessible, especially in countries where people cannot use ChatGPT due to restrictions from OpenAI or local governments.”

• Performance-wise, this first-tier Chinese large language model is as good as ChatGPT. Chimera, which is the Latin version, is competitive in the English language.
• With the help of human and automatic evaluations, the model uses a methodological approach to observe LLMs.
• Phoenix is a pioneer in evaluating extensive LLMs.
• It has superseded the existing LLMs (in Chinese), like BELLE and Chinese-LLaMA-Alpaca in performance.
• It is also exceeding other existing models in non-Latin languages like Korean, Japanese, and Arabic.

What are Large Language Models?

LLMs are powered by a deep-learning algorithm that is capable of:

• Translating
• Summarizing
• Predicting
• Recognizing

LLMs are basically part of transformer models that can seamlessly generate text based on the inputs fed based on huge amounts of data.

Thankfully, these models are not restricted to learning just human languages; they also have writing codes, besides an understanding of proteins.

LLMs are known for leveraging a number of natural languages processing applications like chatbots, AI assistants, and translations. They are also a part of other industries such as healthcare, software development, etc.

What are they used for?

Language is not restricted to human communication. For instance, the subject of biology uses protein and molecular sequence as their language, and coding is used for computers. And so, in a broader sense, LLMs can be applied to different subjects like biology, where a different method of communication is required.

Recommended: ChatGPT Isn’t Ready to Translate for Businesses 

For example, an LLM-powered AI system can train and learn from datasets of molecular and protein structures, and further utilize the knowledge to assist scientists to explore and develop treatments and vaccines.

LLMs are a ray of hope in the fields of research and creativity where they can generously contribute to addressing the biggest problems that we face today.

LLMs can also do the following:

• Create search engines.
• Training chatbots.
• Compose different tools for poems, stories, and songs.
• Generate marketing materials.

Final Thoughts

This is undoubtedly a promising, and exciting new development in the LLM arena, particularly because of its multilingual ability to surpass all kinds of linguistic backgrounds and enable people to realize the true power of language models.