Large Language Models

Walmart has been actively exploring the use of AI for some time now, and they recently began testing their own large-scale language model (LLM) called Wallaby. Designed specifically for the retail industry, this model has been trained with decades of Walmart data to better understand how employees and customers communicate. It also responds in a more natural way and is designed to align with Walmart's core values of customer service. Desirée Gosby, Walmart's vice president of emerging technologies for global technology, said in an interview that Wallaby is currently in an internal testing phase, esp...

OpenAI's ChatGPT5 has been long awaited, and speculation about GPT-5's features, new features, and release date continues to heat up. A lot of information about ChatGPT5 has been circulated. This article will combine the news on the market to integrate and predict the release date, features, and price of ChatGPT5. What is ChatGPT5? ChatGPT-5 is the latest version of ChatGPT. ChatGPT is a large language model (LLM) that attracted worldwide attention when it was launched in November 2022…

Have you ever thought that your research papers may have been used to train AI? That's right, many academic publishers are "packaging and selling" their results to technology companies that develop AI models. There is no doubt that this move has caused quite a stir in the research community, especially when authors are unaware of it. Experts say that if your work has not yet been used by a large language model (LLM), it is likely to be "used" in the near future. Recently, British academic publisher Taylor & Francis reached an agreement with Microsoft...

Google has officially launched a new member of its Gemma2 series, the Gemma22B model. This model, with 2 billion parameters, has demonstrated excellent performance in a variety of hardware environments. In addition to the powerful model itself, Google has also launched the ShieldGemma security classifier to filter harmful content and provides the Gemma Scope tool for researchers to analyze the model's decision-making process. Gemma22B performed particularly well in the "Chatbot Arena" rankings, with a high score of 1130, successfully surpassing GPT-…

Llama3.1 has been leaked! You heard it right, this open source model with 405 billion parameters has caused a stir on Reddit. This may be the open source model closest to GPT-4o so far, and even surpasses it in some aspects. Llama3.1 is a large language model developed by Meta (formerly Facebook). Although it has not been officially released, the leaked version has caused a sensation in the community. This model not only includes the base model, but also benchmark results of 8B, 70B, and 405B with the largest parameters. Performance comparison: Ll…

In an interview with a Sequoia Capital podcast last week, Microsoft CTO Kevin Scott reiterated his belief that the "law of scale" of large language models (LLMs) will continue to drive progress in artificial intelligence, despite suspicions among some in the field that progress has stagnated. Scott played a key role in Microsoft's $13 billion technology-sharing agreement with OpenAI. "Others may disagree, but I don't think that scale has reached a critical point of diminishing returns," Scott said. "I want people to understand that there is a...

Groq, an innovative company focused on increasing the speed of AI computing, has recently received widespread attention for its large models and unique hardware, the Language Processing Unit (LPU). Let's summarize Groq's features and its usage scenarios in layman's terms. First, Groq has developed a piece of hardware called the LPU, which is specifically designed to handle language-related tasks. Unlike traditional graphics processing units (GPUs), the LPU is a tensor streaming processor (TSP) designed for high-speed and efficient AI and machine learning model inference. In other words, it is a core designed to understand...

A new study from the University of Chicago shows that the large language model GPT-4 can outperform human analysts in predicting a company's future earnings growth, and the artificial intelligence model only uses the company's financial statements without additional information. Image source Pexels In the past, financial analysts relied on expertise and experience to assess a company's financial situation and predict future earnings. However, this study shows that artificial intelligence models can perform this task just as well, or even better. It is understood that the researchers provided the model with anonymized financial data, including balance sheets and income statements, and…

Apple is working on a large language model (LLM) that runs on-device, according to Bloomberg reporter Mark Gurman, to improve the responsiveness and privacy protection of its upcoming generative AI features. Gurman mentioned in his "Power On" newsletter that Apple's LLM will be the basis for the company's future generative AI features. Unlike most cloud-based AI services today, all signs indicate that the model will run entirely on the user's device. By running on the device, Apple's...

Once upon a time, Nvidia's H100 GPU for artificial intelligence computing was in short supply. However, according to Digitimes, TSMC's Taiwan general manager Terence Liao said that the delivery waiting time for Nvidia H100 has been greatly shortened in the past few months, from the initial 3-4 months to the current 2-3 months (8-12 weeks). Server foundry manufacturers also revealed that compared to the situation in 2023 when Nvidia H100 was almost impossible to buy, the current supply bottleneck is gradually easing. Although the delivery waiting time has…

AMD recently released a series of benchmarks, claiming that its Ryzen mobile 7040 Phoenix series and 8040 series processors can outperform Intel's latest Core Ultra Meteor Lake CPUs by up to 79% in running large language models (LLMs). This test compares the AMD Ryzen 7 7840U and Intel Core Ultra 7 155H processors, both of which are equipped with hardware neural network processing units (NPUs). AM…

In the AI industry, especially in the field of generative AI, the rapid development of technology and the broad prospects of application have attracted a lot of investment and attention. However, the high cost of this field has also caused widespread discussion in the industry. Recently, a report in the Wall Street Journal pointed out that companies in the AI industry spend 17 times their revenue on purchasing Nvidia GPUs, which is a shocking figure and has also triggered in-depth thinking about the future development of the industry. AI startup Cognition Labs, backed by well-known investor Peter Thiel, is seeking a valuation of $2 billion, and its valuation…

OpenAI may have accidentally leaked information about its upcoming GPT-4.5 Turbo large language model. Netizens discovered that an article about GPT-4.5 Turbo was posted on the OpenAI blog, but the page has been withdrawn and search engines can no longer retrieve relevant information. The leaked information shows that GPT-4.5 Turbo will surpass its predecessor GPT-4 Turbo in terms of speed, accuracy, and scalability. More importantly, the leaked description mentions the model's "knowledge deadline...

Google's latest research reveals a method to attack large language models and successfully steal key information from the OpenAI GPT-3.5-turbo model. According to Google's statement, they not only restored the entire projection matrix of the OpenAI large model, but also obtained the exact size of the hidden dimension, all with less than 2,000 clever API queries and a cost as low as 150 yuan. The core target of the attack is the model's embedding projection layer, which is the last layer of the model and is responsible for mapping the hidden dimension to the logits vector. By issuing targeted queries to the model's API, it is possible to extract...

Medical artificial intelligence (AI) technologies are rapidly evolving, aiming to revolutionize healthcare delivery by leveraging the vast potential of large language models (LLMs). These technological advances promise to improve diagnostic accuracy, personalize treatment options, and unlock access to comprehensive medical knowledge, fundamentally transforming patient care. Integrating AI into healthcare aims to improve the efficiency and precision of healthcare delivery, effectively bridging the gap between the technological frontier and patient-centered care. Linguistic diversity in global healthcare delivery is a key challenge. Although medical knowledge is dominated by English, medical…

Artificial intelligence startup Anthropic today announced the launch of its groundbreaking Claude 3 series of large language models (LLMs) that set new performance benchmarks on a variety of cognitive tasks. The Claude 3 series includes three sub-models, namely Claude 3 Haiku, Claude 3 Sonnet and Claude 3 Opus, which provide different levels of intelligence, speed and cost options to meet the needs of various artificial intelligence applications. According to officials, the Claude 3 series…

Data company Palantir (PLTR) surged more than $19% in after-hours trading after its earnings report on Monday, beating analysts' revenue expectations. As of the fourth quarter of 2023, the company's revenue increased by $20% year-on-year to $608 million, exceeding LSEG (formerly Refinitiv)'s expectations of $602 million. Source Note: Image generated by AI, image licensing service provider Midjourney Palantir CEO Alex Karp said in a letter to shareholders that the company's expansion and growth are "unprecedented", especially...

Oracle announced the general availability of the Oracle Cloud Infrastructure (OCI) Generative AI service, along with new innovations that make it easier for enterprises to take advantage of the latest advances in generative AI. The OCI Generative AI service is a fully managed service that seamlessly integrates Large Language Models (LLMs) from Cohere and Meta Llama2 to address a variety of business use cases. The OCI Generative AI service now features…

Google announced the termination of its contract with Australian data company Appen, which provided large language model AI training for Google products such as Bard and Search. This decision is part of Google's ongoing efforts to evaluate and adjust its many supplier partnerships under Alphabet to ensure the maximum efficiency of its supplier operations. According to The Verge, Appen said that they were not aware of this in advance and that the decision was unilaterally terminated by Google. At companies such as Appen, manual workers are usually responsible for training the less popular parts of AI...

Cohere is a company focused on building the future of artificial intelligence, providing a range of powerful language AI products that enable content generation, summarization, search, and more. Cohere is a Canadian startup founded in 2019 by three former Google AI researchers, aiming to provide enterprises with high-performance, secure, and customizable large language models (LLMs). Cohere's vision is to make it easy for everyone to use language AI to build amazing products and services. Cohere's core...

Marco Argenti, chief investment officer of financial giant Goldman Sachs, said in an interview looking ahead to 2024 that the year will focus mainly on "hybrid" AI and applications built on large language models. He emphasized that "hybrid AI" will become the dominant trend, in which large models act as the "brain" to interpret prompts and user needs, or as orchestrators of planning tasks, assigning tasks to multiple working models dedicated to specific tasks, similar to ChatGPT. Source Note: Image generated by AI,…

Travel site Expedia is taking steps to encourage people to start their travel searches using artificial intelligence (AI) features on its site, rather than using external search engines. Expedia already uses AI in some customer service features and to help property owners describe their homes and hotels. The company hopes that in the future AI will be able to recommend travel destinations based on users’ previous travel experiences and drive more direct traffic to its site. It’s a long-term plan to change the balance of power on the web — though it’s still in its early stages for the company…

Researchers at Purdue University in Indiana have designed a new method to successfully induce large language models (LLMs) to generate harmful content, revealing potential harm hidden in compliant responses. When conversing with a chatbot, the researchers found that by leveraging probabilistic data and soft labels made public by the model maker, the model could be forced to generate harmful content with a success rate of up to 98%. Source Note: The image is generated by AI, and the image is authorized by Midjourney. Traditional jailbreaking methods usually require providing prompts to bypass security features, while this new method uses probabilistic data and soft labels to force...

Google released its latest and most advanced large-scale language model, Gemini, on Wednesday, and announced that starting today, users will be able to experience the specially adjusted Gemini Pro English version in Google Bard. Click to go to the official website of Google Gemini Gemini was first announced at Google I/O in May this year, and three different sizes were released: Gemini Ultra, Gemini Pro and Gemini Nano. Each size of Gemini …