These sixty minute, weekly sessions – facilitated by Technologists and Pedagogy Experts from the CTLT – are designed for faculty and staff at UBC who are using, or thinking about using, Generative AI tools as part of their teaching, researching, or daily work. Each week we discuss the news of the week, highlight a specific tool for use within teaching and learning, and then hold a question and answer session for attendees.

They run on Zoom every Wednesday from 1pm – 2pm and you can register for upcoming events on the CTLT Events Website.

---

News of the Week

Each week we discuss several new items that happened in the Generative AI space over the past 7 days. There’s usually a flood of new AI-adjacent news every week – as this industry is moving so fast – so we highlight news articles which are relevant to the UBC community.

This week’s tech news highlights several major developments in AI. OpenAI introduced GPT-4o Mini, a cost-efficient AI model, while Meta launched the advanced Llama 3.1. Mistral AI unveiled the Large 2 model and NeMo platform, emphasizing scalability and user-friendly integration. Researchers used AI to identify three Parkinson’s subtypes, potentially revolutionizing treatment. A new RNN model for test-time learning showed promising results. Meta also made its Llama 3.1 405B model open source, and UBC announced AI grant deadlines. Lastly, the University of Chicago introduced Glaze and Nightshade, tools to protect digital art from AI plagiarism.

Here’s this week’s news:

OpenAI Introduces GPT-4o Mini:

OpenAI has unveiled GPT-4o Mini, a cost-efficient AI model designed to provide high-performance intelligence at reduced operational costs. This model aims to democratize access to advanced AI capabilities for various applications. Read more.

Meta Launches Llama 3.1:

Meta has released Llama 3.1, its most advanced AI model to date, enhancing capabilities in natural language processing and understanding. Llama 3.1 promises improved efficiency and performance across various AI-driven tasks. Read more.

Mistral AI’s Large 2 Model:

Mistral AI has introduced the Large 2 model, emphasizing scalability and robustness in AI applications. This model is designed to handle large-scale data and complex computations efficiently. Read more.

Mistral’s NeMo Platform:

Mistral AI has also launched NeMo, a platform designed for seamless AI integration and deployment. NeMo focuses on enhancing user experience and simplifying the implementation of advanced AI models. Read more.

AI Identifies Parkinson’s Subtypes:

Researchers have developed an AI system that identifies three distinct subtypes of Parkinson’s disease, potentially leading to more personalized treatments. This breakthrough could significantly improve patient outcomes. Read more.

New RNNs for Test-Time Learning:

A recent paper proposes a new class of RNN layers with expressive hidden states and linear complexity. These layers, designed for test-time training, outperform traditional models in long-context scenarios. Read more.

Meta’s Llama 405B Model Goes Open Source:

Meta has made its Llama 405B model available for download, marking the first open-source release of a GPT-4 class AI model. This move aims to foster innovation and collaboration in the AI community. Read more.

UBC Announces AI Grant Deadlines:

The University of British Columbia has announced deadlines for its large TLEF grants focused on generative AI collaborative clusters. The letter of intent is due tomorrow (July 25th), with another call expected in the fall. Read more.

Glaze and Nightshade: New Privacy Tools:

The University of Chicago has introduced Glaze and Nightshade, tools designed to protect digital art from AI plagiarism. These tools offer artists methods to safeguard their creative works against unauthorized AI training. Glaze and Nightshade.

---

Tool of the Week

Tool of the Week: Llama 3.1 405B Model

The benchmark chart compares several leading AI language models, including Meta’s Llama 3.1 405B, Nemotron 4 340B Instruct, GPT-4, GPT-4 Omni, and Claude 3.5 Sonnet, across various categories such as general knowledge, code, math, reasoning, tool use, long context, and multilingual capabilities. Notably, Llama 3.1 405B excels in several benchmarks, demonstrating competitive performance on par with top models like GPT-4 and Claude 3.5 Sonnet, particularly in categories like math (GSM8K), reasoning (ARC Challenge), and long-context tasks. This showcases its high capabilities and potential as a leading AI model.

What is Llama 3.1 405B Model?

Llama 3.1 405B Model is Meta’s latest and most advanced AI model in the Llama series, designed to offer cutting-edge capabilities in natural language processing and understanding.

How is it used?

It is utilized in various applications requiring advanced AI, including conversational AI, language translation, and text generation. Researchers and developers can integrate this model into their projects to enhance AI-driven tasks.

What is it used for?

Llama 3.1 405B is used to push the boundaries of AI performance, enabling more efficient, accurate, and context-aware interactions across different platforms and services. It supports innovation in AI research and practical applications in industry and academia. For additional information, explore Llama 3.1.

---

Questions and Answers

Each studio ends with a question and answer session whereby attendees can ask questions of the pedagogy experts and technologists who facilitate the sessions. We have published a full FAQ section on this site. If you have other questions about GenAI usage, please get in touch.