On Tuesday, Microsoft made Phi-3, an update to its small language model, available to the public. The company claims that Phi-3 can do numerous tasks that were previously assumed to require much larger models.
According to Microsoft, the Phi-3 mini’s performance was evaluated using internal testing and academic benchmarks, demonstrating its competitiveness versus models such as “Mixtral 8x7B and GPT-3.5.”
This means the Phi-3-mini can outperform models twice its size across a variety of benchmarks that evaluate language, coding, and math capabilities.
Newsng gathered that 3.8 billion parameters make up the Phi-3 Mini, which was trained on a smaller dataset than large language models such as GPT-4. Hugging Face, Ollama, and Azure are currently selling it.
Phi-3 Medium (14B parameters) and Phi-3 Small (7B parameters) are scheduled for release by Microsoft. A model’s parameter count is the number of complex instructions it can comprehend.
Eric Boyd, Corporate Vice President of Microsoft Azure AI Platform, while remarking on the model noted that Phi-3 Mini is “just in a smaller form factor” as competent as LLMs like GPT-3.5.
They were inspired by how children learned from bedtime stories, books with simpler words, and sentence structures that talk about larger topics.
“There aren’t enough children’s books out there, so we took a list of more than 3,000 words and asked an LLM to make ‘children’s books’ to teach Phi,” he said.
We earlier reported that Microsoft announced that it will invest $1.5 billion in G42, an artificial intelligence company based in the United Arab Emirates.
