AI Researchers Claim They Can Double the Efficiency of Chatbots – Yahoo Finance
Have you ever noticed that your AI chatbot get lost in the middle of a conversation, or it simply says it cannot handle prompts that are too long? Well, that is because each model has a limitation in its processing capabilities, and starts to suffer once it goes over that limit —pretty much like they suffered from some kind of a digital attention deficit disorder. But this could soon change thanks to a new method for supercharging LLM capabilities.
Current LLMs have limited context capacities. For example, ChatGPT taps just 8,000 tokens of context, while Claude handles 100,000. Tokens are the basic units of text or code used by an LLM AI to process and generate language This restricts how much background information they can harness when formulating replies. Abacus AI has developed a method that allegedly doubles the usable context length for open-source LLMs like Meta’s Llama without compromising the model's accuracy in practical application.
Is Meta's ChatGPT Killer Really Open Source?
Their technique involves "scaling" the position embeddings that track word locations in input texts. According to their Github page, Abacus AI claims that its scaling method drastically increases the number of tokens that a model can handle.
The researchers evaluated two scaled LlaMA variants on tasks like substring location and open-book QA. The scale 16 model maintained accuracy on real-world examples up to 16,000-word contexts, versus only 2,000 words in baseline Llama. It even showed some coherence at 20,000+ words, something that was not possible to achieve with just fine-tuning techniques.
The significance of context extension cannot be overstated. A narrow context window makes the model accurate but not really usable in complex tasks that require some background. Conversely, with an expanded context, LLMs can process and generate better responses but either take more time to do so or return sup-par results. Handling longer contexts efficiently could enable LLMs to absorb whole documents or multiple documents as background when generating text. This may lead to outputs that are more knowledge-grounded and consistent across long conversations.
Claude 2 Is Out—How Does Anthropic’s AI Chatbot Compare to ChatGPT and Google Bard?
However, the gains are not perfectly proportional to the scale factors.
It’s still necessary to fine tune strategies because scaling alone doesn’t guarantee high quality outputs. The Abacus team is also exploring advanced position encoding schemes from recent papers to further extend context capacity.
Their work suggests that scaling up existing LLMs is a viable path to expanding usable context length. This could democratize access to Large Language Models capable of handling lots of context at once.
Abacus AI has opened the doors of their repository “for research purposes only,” sharing code specific to their fine-tuning projects. This makes it possible to further iterate on its development and apply the fine tuning methods on virtually any open source Large Language Model.
With applications from personalized chatbots to creative writing aids, more memory-empowered LLMs could soon enable next-generation AI assistants that are conversant across diverse topics. For now, researchers are progressing rapidly to overcome technical constraints in pursuit of artificial general intelligence —meaning, generalized human cognitive abilities in an AI model. Maybe someday our digital friends will handle as many tabs as we humans can, but without the headache!
Meta is reportedly building chatbots with different personalities for its social media sites.
Salesforce introduced its AI layer, dubbed Einstein, in 2016. More recently, at the Salesforce World Tour event in NYC in May, all the company talked about was generative AI and Data Cloud, its in-house data lake. Today, it announced the next step in that journey with the release of Einstein Studio and the ability to bring your own model (BYOM).
There was a bit of a hubbub in February as it emerged that OpenAI had seemingly purchased AI.com in order to redirect it to the ChatGPT web interface. Redirect it to the obvious candidate and then dangle it in front of their competitors for a 50% markup.
Upload multiple files, suggested prompts and replies, and a whole lot more. OpenAI just released a bevy of updates for ChatGPT.
Fisker, the automotive startup-turned publicly traded company founded by famed automotive designer Henrik Fisker, revealed Thursday night a dizzying array of EV prototypes, from a beefy pickup truck with an expandable bed and grand tourer sports car with up to 600 miles of range to a souped-up, off-road version of its Ocean SUV and the mysterious PEAR vehicle that it plans to build with Foxconn. The showcase, which was part of an event in Huntington Beach, California for investors and media, was the first time that Fisker has shown off all of its future vehicles in prototype form and laid out its expansive and ambitious future product plan. While the EVs shown Thursday are only prototypes and the specifics are scarce, the company did share some new details that provide a clearer picture of what road — or roads — Fisker is taking.
It’s clear now that we saw the market bottom out last October. The S&P 500 is up about 1,000 points, or ~27%, from that trough. The question for investors now is, what happens next? Mike Wilson, the well-known strategist from Morgan Stanley, has a reputation as an uber-bear, but he’s pulling back from that in a recent note. “The data we have today suggests to us that we are in a policy-driven, late-cycle rally,” Wilson says. He goes on to note several supportive factors, including a reduced rate
(Bloomberg) — Apple Inc., Samsung Electronics Co. and HP Inc. are among the biggest names freezing new imports of laptops and tablets to India after the South Asian country abruptly banned inbound shipments without a license.Most Read from BloombergChina Embassy Rips ‘Brutal’ Russia Border Incident in Rare MoveS&P 500 Wipes Out Almost 1% Gain; Bond Yields Drop: Markets WrapTrump Cites Self Incrimination Concern in Lawsuit Against CohenSouth Africa Spurns US Pressure to Stop Using China’s Huawei
After Apple's third consecutive quarter of declining revenue, some Wall Street analysts say its valuation is looking stretched.
The Sunnyvale cybersecurity company posted better-than-expected earnings for the second quarter. But a warning about slowing sales spooked investors.
Treasury bill yields are above 5% after the Federal Reserve lifted its benchmark lending rate by a quarter-point last week.
For the week, the Dow lost 1%, the S&P 500 fell 2.3%, and the Nasdaq dropped 2.8%.
(Bloomberg) — Qualcomm Inc., NXP Semiconductors NV and other chipmakers are forming a new company to hasten the development of RISC-V, a standard that could challenge the near-ubiquitous technology of Arm Ltd.Most Read from BloombergChina Embassy Rips ‘Brutal’ Russia Border Incident in Rare MoveS&P 500 Wipes Out Almost 1% Gain; Bond Yields Drop: Markets WrapTrump Cites Self Incrimination Concern in Lawsuit Against CohenSouth Africa Spurns US Pressure to Stop Using China’s Huawei TechnologyElon
The outlook for these e-commerce players is strengthening and this should remain a viable space to invest in for 2023 and beyond.
Block reported second-quarter earnings that topped estimates as the consumer Cash App business turned in a strong quarter.
I'm afraid of the stock market. With my first investment, I lost 60% of my money. So I'm strictly into bonds. With interest rates low, what's your advice? Should I stay or try something else? -Jerold It's reasonable to be … Continue reading → The post Ask an Advisor: ‘I'm Strictly Into Bonds' and Afraid of the Stock Market. Is This a Strategy I Should Stick With? appeared first on SmartAsset Blog.
Don’t look for lower inflation or interest-rate cuts anytime soon. Instead, higher bond yields probably are in the offing.
Google is trying to entice hybrid workers back to the office by offering 'cheap' stays at its campus hotel.
(Bloomberg) — The biggest US lenders expect to pay almost $8.9 billion to help replenish the US government’s bedrock Deposit Insurance Fund after it was tapped to backstop uninsured depositors at Silicon Valley Bank and Signature Bank. Most Read from BloombergChina Embassy Rips ‘Brutal’ Russia Border Incident in Rare MoveS&P 500 Wipes Out Almost 1% Gain; Bond Yields Drop: Markets WrapTrump Cites Self Incrimination Concern in Lawsuit Against CohenSouth Africa Spurns US Pressure to Stop Using Chi
Share prices skyrocket, then sink, for companies that don’t even have a direct link to room-temperature superconductors.
There's "obviously a mismatch" between Anheuser-Busch's market value and their employee numbers, former company exec Anson Frericks says while predicting more layoffs are coming.