ChatGPT’s creators have attempted to get the system to explain itself. They found that while they had some success, they ran into some issues – including the fact that artificial intelligence may be using concepts that humans do not have named for, or understanding of. Researchers at OpenAI, which developed ChatGPT, used the most recent version of its model known as GPT-4 to try and explain the behaviour of GPT-2, an earlier version. It is an attempt to overcome the so-called black box problem with large language models such as GPT. While we have a relatively good understanding of what goes into and comes out of such systems, the actual work that goes on inside remains largely mysterious. That is not only a problem because it makes things difficult for researchers. It also means that there is little way of knowing what biases might be involved in the system, or if it is providing false information to people using it, since there is no way of knowing how it came to the conclusions it did. Engineers and scientists have aimed to resolve this problem with “interpretability research”, which seeks find ways to look inside the model itself and better understand what is going on. That has often required looking at the “neutrons” that make up such a model: just like in the human brain, an AI system is made up of a host of so-called neutrons that represent parts of the data it uses. Finding those is difficult, however, since humans have had to pick through the neurons and manually inspect them to find out what they represent. But some systems have hundreds of billions of parameters and so actually getting through them all with people is impossible. Now, researchers at OpenAI have looked to use GPT-4 to automate that process, in an attempt to more quickly pick through the behaviour. They did so by attempting to create an automated process that would allow the system to provide natural language explanations of the neuron’s behaviour – and apply that to another, earlier language model. That worked in three steps: looking at the neuron in GPT-2 and having GPT-4 try and explain it, then simulating what that neuron would, and finally scoring that explanation by comparing how the simulated activation worked with the real one. Most of those explanations went badly, and GPT-4 scored itself poorly. But researchers said that they hoped the experiment showed that it would be possible to use the AI technology to explain itself, with further work. The creators came up against a range of “limitations”, however, that mean the system as it exists now is not as good as humans at explaining the behaviour. Part of the problem may be that explaining how the system is working in normal language is impossible – because the system may be using individual concepts that humans cannot name. “We focused on short natural language explanations, but neurons may have very complex behavior that is impossible to describe succinctly,” the authors write. “For example, neurons could be highly polysemantic (representing many distinct concepts) or could represent single concepts that humans don’t understand or have words for.” It also runs into problems because it is focused on specifically what each neuron does individually, and not how that might affect things later on in the text. Similarly, it can explain specific behaviour but not what mechanism is producing that behaviour, and so might spot The system also uses a lot of computing power, the researchers note. Read More Google to unveil major new AI AI robots figure out how to play football in shambolic footage White House asks hackers to break ChatGPT White House reveals plan to ‘protect’ citizens from danger of AI DeepMind boss says human-level AI is just a few years away Regulator to probe use of artificial intelligence such as ChatGPT

Twitter owner Elon Musk appeared to troll Meta’s Mark Zuckerberg as the new Threads platform closes in on 100 million users. Since the launch of Threads, the Tesla and SpaceX chief has been critical of the new app, even filing a lawsuit accusing Meta of poaching former Twitter staff to build the new social media app. Mr Musk responded to a Twitter post on the lawsuit last week, saying “Competition is fine, cheating is not”. On Saturday, he called Threads “just Instagram minus pics,” adding that it “makes no sense, given that thirst pics are the main reason people use that app”. Then later on Sunday, responding to a screenshot of a Threads conversation between Mr Zuckerberg and the account of fast food joint Wendy’s, the Tesla chief said “Zuck is a cuck”. “Cuck” is an insult often used to attack the masculinity of an opponent. It is short for “cuckold”, a word dating to the Middle Ages that describes a man who is aware of his wife’s infidelity and acquiesces to it. Meanwhile, as the Threads app nears close to 100m users, the Meta chief also has expressed his fare share of jibes at the SpaceX and Tesla boss. Mr Zuckerberg posted his first tweet in 11 years on the day of Threads’ launch, sharing the popular meme of two Spider-Men pointing at each other, a likely reference to the comparisons drawn between Threads and Twitter. The Meta boss also responded to two posts on Threads in one-worded replies, commenting “concerning”, seemingly a reference to the way Mr Musk tweets often. Mr Musk has replied with the word “concerning” at least 26 times, according to a review of the multibillionaire’s tweets by Insider. Since the launch of Threads, Mr Zuckerberg has hoped to draw a contrast to Twitter which currently boasts over 250m users but has also seen a surge in hate speech and misinformation since Mr Musk bought the company last year. Even as the Meta chief has expressed intention to make “kindness” the focus of Threads, it remains to be seen how new signups by race supremacists and fringe rightwing personalities will affect the platform. It also remains unclear if the beef between the two tech billionaires is playful banter or bitter competition as they have also expressed willingness to engage in an MMA bout. The two agreed to a physical fight last month after Mr Musk accused the Meta chief of ripping off Twitter with Threads. “Up for a cage match,” the Twitter owner and chief operating officer wrote, to which Mr Zuckerberg replied “Send me location”. UFC trainer Javier Mendez recently said the Meta boss is more likely to win the fight due to his active training, apparent fitness, and experience. Podcaster and jiu jitsu enthusiast Lex Fridman has also said following a training session with the Tesla boss that he was “extremely impressed” with Mr Musk’s “strength, power, and skill, on the feet and on the ground”. Read More I’m a tech nerd who does jiu-jitsu. I know exactly why Mark Zuckerberg wants to fight Elon Musk Musk, Zuckerberg and the bitter battle for the future of social media I tried Mark Zuckerberg’s Twitter rival, Threads – I’m not impressed Threads hits 70 million sign-ups on its second day Mark Zuckerberg trolls Elon Musk with Spider-Man meme after launching Twitter rival Meta’s new Twitter rival app Threads gets over 10 million sign-ups

Advanced Micro Devices has revealed a new artificial intelligence chip in its race to compete with fellow chipmaker Nvidia in supplying the foundation for a boom in AI-fueled business tools

AT&T Inc. reported profit and free cash flow that topped analysts’ estimates, offering a brighter picture as the

LAS VEGAS--(BUSINESS WIRE)--Aug 16, 2023--

MCLEAN, Va.--(BUSINESS WIRE)--Sep 14, 2023--

TOKYO--(BUSINESS WIRE)--Jul 3, 2023--

TL;DR: The Complete 2023 CompTIA Course Super Bundle is on sale for £31.34, saving you

HOUSTON--(BUSINESS WIRE)--Jun 13, 2023--

SAVE $49.01: As of June 23, you can score a pair of Apple AirPods Pro

By Stephen Nellis (Reuters) -Semiconductor company Arm Holdings on Wednesday gave a fiscal third quarter sales outlook below Wall Street

The European Commission is increasing the pressure on member-nations to stop using Huawei Technologies Co. and ZTE Corp.