In this article, I want to discuss the use of ELO as an evaluation method for LLMs (Large Language Models). The LLM test portal ChatBot Arena utilizes this approach to rate LLMs. I became aware of this evaluation process behind the ChatBot Arena through Martin Blaha, who had read my article “How are LLM systems… Continue reading Rating LLMs with ELO – the ChatBot Arena
Category: AI Development
Overview LLM Benchmark
When reports emerge about new or improved LLMs (Large Language Models), as has been the case recently with OpenChat or Mistral, or even when it’s just about the capabilities of ChatGPT, benchmark values are repeatedly presented and brought into consideration. These benchmarks are intended, on one hand, to represent the capabilities of the LLMs, and… Continue reading Overview LLM Benchmark
GAIA-1: GWM for autonomous driving
The Article: “A Generative World Model for Autonomous Driving” published on September 29, 2023, is a result of collaborative research by Anthony Hu, Lloyd Russell, Hudson Yeo, Zak Murez, George Fedoseev, Alex Kendall, Jamie Shotton, and Gianluca Corrado, all associated with Wayve, an organization in the field of autonomous driving. Study Objective The study titled… Continue reading GAIA-1: GWM for autonomous driving
EmotionPrompt Technique
Artificial Intelligence (AI) is on the cusp of a remarkable development: the integration of emotional intelligence into Large Language Models (LLMs) such as ChatGPT and GPT-4. A study by Cheng Li, Jindong Wang, Yixuan Zhang, Kaijie Zhu, Wenxin Hou, Jianxun Lian, Fang Luo, Qiang Yang, and Xing Xie titled “Large Language Models Understand and Can… Continue reading EmotionPrompt Technique
Instruction Tuning and the Tuna Model
In the world of Artificial Intelligence (AI), language models play a central role. These models, known as Large Language Models (LLMs), have the potential to conduct human-like conversations and answer complex queries. But how can the effectiveness of these models be further enhanced? A key lies in the so-called “Instruction Tuning”. What is Instruction Tuning?… Continue reading Instruction Tuning and the Tuna Model
AMD goes AI
AMD announced (according to The Verge) the release of its MI300 chips and Ryzen 8040 mobile processors, set to be available in laptops by 2024, marking a significant step in competing with Nvidia in the AI chip market. The new MI300X accelerators and M1300A APUs are tailored for training and running large language models (LLMs),… Continue reading AMD goes AI
What is SDXL Turbo?
In the latest breakthrough in the world of AI-driven image generation, the research team behind SDXL Turbo introduces their newest accomplishment: a revolutionary text-to-image technology that represents a significant advancement in image synthesis. SDXL Turbo utilizes an innovative distillation technique called Adversarial Diffusion Distillation (ADD), which allows a significant reduction in the steps required for… Continue reading What is SDXL Turbo?
llamafile – LLM local
Llamafile, an innovative project developed by Mozilla and Justine Tunney, represents a significant advancement in the use of Large Language Models (LLMs) on local computers. This technology allows for the operation of advanced language models, similar to ChatGPT, directly on one’s own computer. A llamafile is a comprehensive file that contains both the model weights… Continue reading llamafile – LLM local
AI Secure by Design
Information regarding the Secure by Design principles for the field of AI can be found here: https://www.ncsc.gov.uk/collection/guidelines-secure-ai-system-development This article provides comprehensive guidelines for providers of AI systems, targeting large organizations, cyber security professionals, small and medium-sized organizations, and the public sector. It emphasizes the importance of developing AI systems securely and responsibly to realize their… Continue reading AI Secure by Design
“Secure by Design”
Summary: “Secure by Design” integrates security throughout all development stages, minimizes risks, and promotes robust, user-friendly software. Detailed Explanation: In practice, “Secure by Design” means that security is not an isolated feature but a pervasive theme considered at every stage of software development. For experienced computer scientists, this implies a continuous engagement with security issues,… Continue reading “Secure by Design”