Lifeboat Foundation: Safeguarding Humanity
Toggle light / dark theme

Blog

Category: robotics/AI – Page 674

These two scientists have mapped out the insides or “reachable space” of a language model using control theory, what they discovered was extremely surprising

Please support us on Patreon to get access to the private Discord server, bi-weekly calls, early access and ad-free listening.
/ mlst.

Aman Bhargava from Caltech and Cameron Witkowski from the University of Toronto to discuss their groundbreaking paper, “What’s the Magic Word? A Control Theory of LLM Prompting.” They frame LLM systems as discrete stochastic dynamical systems. This means they look at LLMs in a structured way, similar to how we analyze control systems in engineering. They explore the “reachable set” of outputs for an LLM. Essentially, this is the range of possible outputs the model can generate from a given starting point when influenced by different prompts. The research highlights that prompt engineering, or optimizing the input tokens, can significantly influence LLM outputs. They show that even short prompts can drastically alter the likelihood of specific outputs. Aman and Cameron’s work might be a boon for understanding and improving LLMs. They suggest that a deeper exploration of control theory concepts could lead to more reliable and capable language models.

What’s the Magic Word? A Control Theory of LLM Prompting (Aman Bhargava, Cameron Witkowski, Manav Shah, Matt Thomson)
https://arxiv.org/abs/2310.

LLM Control Theory Seminar (April 2024)
• LLM Control Theory Seminar (April 2024)

Society for the pursuit of AGI (Cameron founded it)
https://agisociety.mydurable.com/

Elon Musk says AI will eliminate all jobs, calls it his ‘biggest fear’

Elon Musk, the billionaire CEO of Tesla, has said that artificial intelligence (AI) will eventually eliminate all jobs. However, he affirmed that this is not necessarily an adverse development.

Speaking at a startup and tech event in Paris on Thursday (May 23) via video link, Musk said, “Probably none of us will have a job”, while predicting a future where jobs would be “optional”

“If you want to do a job that’s kinda like a hobby, you can do a job,” Musk told VivaTech 2024. “But otherwise, AI and the robots will provide any goods and services that you want.”

Beware: These Fake Antivirus Sites Spreading Android and Windows Malware

It’s currently not clear how these bogus websites are distributed, but similar campaigns in the past have employed techniques such as malvertising and search engine optimization (SEO) poisoning.

Stealer malware have increasingly become a common threat, with cybercriminals advertising numerous custom variants with varying levels of complexity. This includes new stealers like Acrid, SamsStealer, ScarletStealer, and Waltuhium Grabber, as well as updates to existing ones such as SYS01stealer (aka Album Stealer or S1deload Stealer).

Generative AI Revolutionizes Quantum Computer Programming

Researchers have developed a machine learning model that generates quantum circuits from text descriptions, similar to how models like Stable Diffusion create images. This method, improves the efficiency and adaptability of quantum computing.

One of the most important recent developments in Machine Learning (ML) is generative models such as diffusion models. These include Stable Diffusion and Dall.e, which are revolutionizing the field of image generation. These models are able to produce high-quality images based on text descriptions.

“Our new model for programming quantum computers does the same but, instead of generating images, it generates quantum circuits based on the text description of the quantum operation to be performed,” explains Gorka Muñoz-Gil from the Department of Theoretical Physics of the University of Innsbruck, Austria.

/* */