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

When Steve Jobs unveiled the iPhone in 2007, it marked a monumental shift in how

SEC standard bearers Georgia and Alabama are front and center in Week 10 of the college football season after the Pac-12 commanded the national spotlight for most of the last month

Lily-Rose Depp's "The Idol" got a standing ovation at Cannes, the Nuggets eliminated the Lakers, Tom Brady buys piece of the Raiders and more in the Roundup.

A list of MLS called up for the June FIFA International window.

By David Kirton SHENZHEN, China China's Huawei Technologies saw a slight uptick in revenue in the first three

SAN FRANCISCO--(BUSINESS WIRE)--Aug 15, 2023--

Matthieu Jalibert will start at fly-half in France's final Rugby World Cup warm-up match against Australia this weekend, head coach...

While viewing birthday wishes on his subreddit, xQc stumbled upon the particular video that became the source of his misery

BEIJING China's new sovereign bonds will help bolster the economic recovery, China's vice finance minister Zhu Zhongming said

Carlee Russell initially claimed that she spotted a toddler along the interstate and was later abducted

iShowSpeed has been in the news an awful lot recently and he’s one of the most popular streamers on the internet. So popular, in fact, that an impersonator is racking up millions of views by showing his old streams. An IShowSpeed fan account has been posting screenshots from the account, which is restreaming old content. The old streams were being picked up by plenty of fans, too. One grab showed that more than 18,000 concurrent viewers were tuning into old footage of the streamer playing Roblox. Other restreams have also been flagged by fan pages recirculating old footage. IShowSpeed hasn’t responded to the old streams yet, though – and at the moment, he’s probably got other things on his mind. The 18-year-old ended a recent stream suddenly after accidentally exposing himself to fans live on air last week. The hugely popular figure was broadcasting to 24,000 people when he accidentally flashed the camera, looking shocked after realising what he’d done instantly. The streamer, real name Darren Watkins stood up, showing his crotch, not realising that his penis was exposed The streamer has since returned to the platform and addressed the "embarrassing moment," and how it's impacted him mentally. "These past days and hours, I've been suffering very mentally, genuinely," he told viewers. "At the end of the day, I'm still a human being." He said it was "one of the worst fears" that has happened and that it's "just depressing." "You guys are joking, you guys can crack your jokes – but you're genuinely not looking a the bigger picture," he said. "Do you all know I have family, little siblings?" "I don't know what to do anymore, bro," Speed continued. Sign up for our free Indy100 weekly newsletter Have your say in our news democracy. Click the upvote icon at the top of the page to help raise this article through the indy100 rankings.