Meta AI: A few days ago, Meta had launched its AI chatbot model Meta AI in its messaging app WhatsApp itself. Although, Meta has released it only for select beta users till now, but a few days after the release of Meta AI, Meta also introduced Llama-3, which the company described as the assistant chat model of Meta AI. Now, soon after Meta introduced Meta AI and Llama-3, Microsoft introduced the latest version of its lightweight AI model Phi-3-Mini.
Microsoft has described Phi-3 as a family of open AI models, which are highly capable and low-budget small language models (SMLs). Let us tell you in this article what are language models and what is the difference between large language models i.e. LLMs and small language models i.e. SLMs.
What is Phi-3-Mini?
According to the report, it is believed that Phi-3-Mini is the first of the three small models that Microsoft is planning to release. This model has reportedly performed better than its same-sized and larger-sized models in various benchmarks like language, coding, logic and maths.
What’s new in Microsoft’s Phi-3-mini?
Microsoft’s latest model provides a new option to the list of high-quality language models available to customers. This offers a more practical option because they create generative AI applications. Phi-3-mini is a 3.8b language model. It is available on AI development platforms like Microsoft Azure AI Studio, HuggingFace and Ollama.
What is language model?
Language model is the backbone of all the language AI models present in the world such as ChatGPT, Cloud, Gemini etc. applications. These models are trained on existing data to solve common language problems like text classification, question answering, text generation, document summarization, etc. All AI chatbots like ChatGPT, Gemini or Meta AI are able to provide service to users only because of the language model.
What are LLMs?
Large language models mainly have two meanings – first – a lot of training data is present in it, that is, a lot of things are stored in it and training is given for it. Second- Its parameter count means the number of things and knowledge this machine has learned during its training. Overall, large language models are equipped to learn things on their own without any instruction. They may be able to solve any complex problem.
How is Phi-3-mini different from LLMs?
Phi-3-mini is a SML. In simple words, SLM i.e. Small Language Model is a more organized model as compared to LLM i.e. Large Language Models. SMLs i.e. Small Language Models cost less to develop and operate and they also perform well on small devices such as smartphones and tablets.
Also read: