
{"id":28403,"date":"2023-08-28T20:49:27","date_gmt":"2023-08-28T18:49:27","guid":{"rendered":"https:\/\/careerfoundry.inbearbeitung.de\/en\/?p=28403"},"modified":"2023-08-28T19:08:40","modified_gmt":"2023-08-28T17:08:40","slug":"large-language-models","status":"publish","type":"post","link":"https:\/\/careerfoundry.inbearbeitung.de\/en\/blog\/data-analytics\/large-language-models\/","title":{"rendered":"What Are Large Language Models? A Complete Guide"},"content":{"rendered":"<p><strong>As the artificial intelligence boom captures minds and reshapes old ways of working, the spotlight is now upon the technology powering these transformative tools.\u00a0<\/strong><\/p>\n<p><span style=\"font-weight: 400;\">Because behind the scenes, it&#8217;s not some vaguely described \u201cartificial intelligence\u201d doing the heavy lifting. It&#8217;s a new technology called <\/span><b>large language models, or LLMs<\/b><span style=\"font-weight: 400;\">. And for those considering <\/span><a href=\"https:\/\/careerfoundry.inbearbeitung.de\/en\/blog\/data-analytics\/what-is-data-analytics\/\" target=\"_blank\" rel=\"noopener\"><span style=\"font-weight: 400;\">a career in data analytics<\/span><\/a><span style=\"font-weight: 400;\"> or tech, understanding the capabilities and applications of LLMs is essential.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">In this comprehensive but jargon-free guide, we\u2019ll define large language models, how they work, and their potential use cases.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Because while they\u2019re already well-known for their ability to generate human-like text, LLMs also have many other applications. Excitingly, this technology is still in its infancy, meaning the most promising innovations are yet to come.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">To help you get to grips with what is an LLM, we\u2019ll answer the following questions:<\/span><\/p>\n<ol>\n<li><strong><a href=\"#what-is-a-large-language-model\">What is a large language model?<\/a><\/strong><\/li>\n<li><strong><a href=\"#examples-of-large-language-models\">7 popular large language model examples<\/a><\/strong><\/li>\n<li><strong><a href=\"#use-cases-of-llms\">Use cases of LLMs<\/a><\/strong><\/li>\n<li><strong><a href=\"#how-do-llms-work\">How do LLMs work?<\/a><\/strong>\n<ul>\n<li><a href=\"#benefits-of-llms\"><span style=\"font-weight: 400;\">Benefits of LLMs<\/span><\/a><\/li>\n<li><a href=\"#limitations-of-llms\"><span style=\"font-weight: 400;\">Limitations of LLMs<\/span><\/a><\/li>\n<\/ul>\n<\/li>\n<li><strong><a href=\"#how-to-use-large-language-models-in-data-analysis\">How to use large language models in data analysis<\/a><\/strong><\/li>\n<li><strong><a href=\"#large-language-models-faq\">Large language models FAQ<\/a><\/strong><\/li>\n<\/ol>\n<p><span style=\"font-weight: 400;\">Ready to dive deep into the world of large language models? Then let\u2019s jump in.<\/span><\/p>\n<h2 id=\"what-is-a-large-language-model\"><span style=\"font-weight: 400;\">1. What is a large language model?<\/span><\/h2>\n<p><span style=\"font-weight: 400;\">In a nutshell, a large language model (LLM) is a natural language processing computer program.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">LLMs are primarily known for driving popular AI tools such as Open AI\u2019s ChatGPT and Google\u2019s Gemini.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Trained using <\/span><b>artificial neural networks<\/b><span style=\"font-weight: 400;\">\u2014which aim to mimic the intelligence of the human brain\u2014large language models can generate natural-sounding and largely accurate text outputs. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">And, unless you\u2019ve been living off the grid, you\u2019ll no doubt be aware that these tools are transforming everything in the world around us, from how we conduct online searches and consume information to how we carry out our jobs.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">But behind these seemingly simple tools, large language models are the real workhorses. These sophisticated systems can execute complex constellations of algorithms to understand text inputs and generate human-like text in response.\u00a0<\/span><\/p>\n<h3><span style=\"font-weight: 400;\">Transformer models<\/span><\/h3>\n<p><span style=\"font-weight: 400;\">Large language models belong to a family of systems called <\/span><b>\u201ctransformer models\u201d<\/b>.<\/p>\n<p>This is <span style=\"font-weight: 400;\">a type of architecture initially proposed in 2017, and subsequently developed to analyze existing text and generate new content. Transformer models are powerful language engines that break sentences down into tiny chunks, digest the meaning of each chunk, and then reassemble them into coherent text. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">This machine-learning breakthrough paved the way for LLMs, which took the transformer concept and ran with it, expanding it to an incredible scale.<\/span><\/p>\n<h3><span style=\"font-weight: 400;\">Did LLMs eat the internet?<\/span><\/h3>\n<p><span style=\"font-weight: 400;\">The large language models we now use have been trained on mountains of text. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">You may occasionally hear that they\u2019ve \u201cread the whole internet\u201d. While this may not be precisely true, it gives a good sense of their size. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">Being trained on such massive datasets, LLMs have absorbed grammar rules, vocabulary, and the nuances of language. And with this immense knowledge, they can assist us in writing, answering questions, making predictions, and even understanding the mood behind words. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">The best part? They\u2019re constantly improving, making them indispensable tools both in the world of data analytics and beyond.<\/span><\/p>\n<h2 id=\"examples-of-large-language-models\"><span style=\"font-weight: 400;\">2. 7 popular large language model examples<\/span><\/h2>\n<p><span style=\"font-weight: 400;\">OK, so we\u2019ve got an idea of what large language models are. But what are some examples? <\/span><\/p>\n<p><span style=\"font-weight: 400;\">You\u2019ll no doubt have heard of AI assistants like ChatGPT and Google\u2019s Gemini (formerly Bard). However, these household names merely represent the interactive frontend of the technology\u2014the interface that makes them accessible for all to use. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">Behind these accessible frontends loom various large language models. Some are open source, and some are privately developed. <\/span><span style=\"font-weight: 400;\">Here are seven well-known ones:<\/span><\/p>\n<h3><span style=\"font-weight: 400;\">GPT (Generative Pre-trained Transformer)<\/span><\/h3>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-28477\" src=\"https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/chatgpt-logo.png\" alt=\"The logo of famous LLM, ChatGPT.\" width=\"886\" height=\"408\" title=\"\" srcset=\"https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/chatgpt-logo.png 886w, https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/chatgpt-logo-300x138.png 300w, https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/chatgpt-logo-768x354.png 768w\" sizes=\"auto, (max-width: 886px) 100vw, 886px\" \/><\/p>\n<p><span style=\"font-weight: 400;\">The GPT family of models, developed by OpenAI, are powerful language models known for their ability to generate coherent and contextually relevant text.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">The model first associated with <\/span><a href=\"https:\/\/chat.openai.com\/\" target=\"_blank\" rel=\"noopener\"><span style=\"font-weight: 400;\">ChatGPT<\/span><\/a><span style=\"font-weight: 400;\">\u2014when it launched in November 2022\u2014was GPT-3. Meanwhile, the newer<\/span> <span style=\"font-weight: 400;\">GPT-4 is more accurate, creative, and reliable at complex problem-solving. Future models will no doubt refine GPT\u2019s abilities further still.<\/span><\/p>\n<h3><span style=\"font-weight: 400;\">LaMDA<\/span><\/h3>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone wp-image-28479 size-full\" src=\"https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/Google-Bard-LLM-1.png\" alt=\"Screenshot from Google&#039;s Bard AI tool, using the LaMDA LLM.\" width=\"1200\" height=\"239\" title=\"\" srcset=\"https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/Google-Bard-LLM-1.png 1200w, https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/Google-Bard-LLM-1-300x60.png 300w, https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/Google-Bard-LLM-1-1024x204.png 1024w, https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/Google-Bard-LLM-1-768x153.png 768w\" sizes=\"auto, (max-width: 1200px) 100vw, 1200px\" \/><\/p>\n<p><span style=\"font-weight: 400;\">LaMDA is a large language model developed by Google.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">It was designed\u2014like OpenAI\u2019s GPT models\u2014to engage in more nuanced and coherent conversations with Google\u2019s search users via its <\/span><a href=\"https:\/\/gemini.google.com\/\" rel=\"noopener\"><span style=\"font-weight: 400;\">Gemini<\/span><\/a><span style=\"font-weight: 400;\"> tool.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Within six months of Bard\u2019s launch, the LLM behind the technology was replaced by Google\u2019s more sophisticated<\/span> <a href=\"https:\/\/ai.google\/discover\/palm2\/\" target=\"_blank\" rel=\"noopener\"><span style=\"font-weight: 400;\">PaLM 2 model<\/span><\/a><span style=\"font-weight: 400;\">. In early 2024, it was replaced again by Gemini, which is multi-modal (it can interact with image and video as well as text).<\/span><\/p>\n<h3><span style=\"font-weight: 400;\">LLaMA<\/span><\/h3>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-28480\" src=\"https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/Meta-Llama-2.png\" alt=\"Screenshot of Meta&#039;s Llama 2 AI tool based on their LLM.\" width=\"1200\" height=\"317\" title=\"\" srcset=\"https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/Meta-Llama-2.png 1200w, https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/Meta-Llama-2-300x79.png 300w, https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/Meta-Llama-2-1024x271.png 1024w, https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/Meta-Llama-2-768x203.png 768w\" sizes=\"auto, (max-width: 1200px) 100vw, 1200px\" \/><\/p>\n<p><span style=\"font-weight: 400;\">LLaMA (Language Model for Multilingual Audience) is one of Meta (formerly Facebook)\u2019s entries into the large language model market!\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">It\u2019s notable for being one of the first proprietary models to provide multilingual communication, bridging language barriers to enable smoother cross-language conversation. Unlike the other entries on our list so far, it&#8217;s only used for non-commercial research purposes.<\/span><\/p>\n<h3><span style=\"font-weight: 400;\">BLOOM<\/span><\/h3>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-28481\" src=\"https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/Bloom-llm.png\" alt=\"Image of the Bloom LLM logo.\" width=\"1170\" height=\"358\" title=\"\" srcset=\"https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/Bloom-llm.png 1170w, https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/Bloom-llm-300x92.png 300w, https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/Bloom-llm-1024x313.png 1024w, https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/Bloom-llm-768x235.png 768w\" sizes=\"auto, (max-width: 1170px) 100vw, 1170px\" \/><\/p>\n<p><span style=\"font-weight: 400;\">Although it has been largely superseded by models like LLaMA, <a href=\"https:\/\/huggingface.co\/bigscience\/bloom\" target=\"_blank\" rel=\"noopener\">BLOOM<\/a> is another large language model focused on multilingual communication.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Crucially, BLOOM is the first on our list not to have been developed by a private big tech company\u2014rather, it&#8217;s an open-source model developed collaboratively by researchers, with its code and resources available for public consumption.<\/span><\/p>\n<h3><span style=\"font-weight: 400;\">XLM-RoBERTa<\/span><\/h3>\n<p><span style=\"font-weight: 400;\">XLM-RoBERTa is an extension of its predecessor, the RoBERTa model, which in turn is built upon the BERT model.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Again, this model has been pre-trained on vast amounts of multilingual text data. It&#8217;s open source and has been modified numerous times, making it a prime example of how different models inform the creation of new, more sophisticated ones.<\/span><\/p>\n<h3><span style=\"font-weight: 400;\">XLNet<\/span><\/h3>\n<p><span style=\"font-weight: 400;\"><a href=\"https:\/\/arxiv.org\/abs\/1906.08237\" target=\"_blank\" rel=\"noopener\">First proposed and released in 2019<\/a>, XLNet is one of the earliest modern LLMs to garner attention.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">While it&#8217;s largely considered out of date compared to newer models (although this is up for discussion, as it depends on what the model is used for) it was one of the first to build upon the new transformer architecture.<\/span><\/p>\n<h3><span style=\"font-weight: 400;\">ELIZA<\/span><\/h3>\n<p><span style=\"font-weight: 400;\">ELIZA, one of the earliest chatbots (from the 1960s), is not a large language model and is incredibly basic by today\u2019s standards.<\/span><\/p>\n<p><span style=\"font-weight: 400;\"> However, it pioneered the concept of natural language interaction, so I felt it deserved a little mention on our list. You can even read the <a href=\"https:\/\/dl.acm.org\/doi\/10.1145\/365153.365168\" target=\"_blank\" rel=\"noopener\">original proposal of the model<\/a> from 1966!<\/span><\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-28482\" src=\"https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/llms-are-the-source-of-the-current-AI-hype.jpeg\" alt=\"A machine learning engineer works with LLMs on his laptop in an office.\" width=\"1200\" height=\"600\" title=\"\" srcset=\"https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/llms-are-the-source-of-the-current-AI-hype.jpeg 1200w, https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/llms-are-the-source-of-the-current-AI-hype-300x150.jpeg 300w, https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/llms-are-the-source-of-the-current-AI-hype-1024x512.jpeg 1024w, https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/llms-are-the-source-of-the-current-AI-hype-768x384.jpeg 768w\" sizes=\"auto, (max-width: 1200px) 100vw, 1200px\" \/><\/p>\n<p><span style=\"font-weight: 400;\">It&#8217;s important to note that the seven LLM examples I&#8217;ve just given you are just a handful of the available ones that exist. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">But what should be clear from this list is that <strong>large language models evolve at a dizzying rate<\/strong>. New ones are constantly emerging, building upon what came before. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">As time goes by, even more advanced models will appear on the scene, so keep your eyes peeled for the latest developments!<\/span><\/p>\n<h2 id=\"use-cases-of-llms\"><span style=\"font-weight: 400;\">3. Use cases of LLMs<\/span><\/h2>\n<p><span style=\"font-weight: 400;\">While initially making headlines for their uncanny ability to write human-like text, LLMs have capabilities far beyond mere content generation.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">While this remains their core function, their ability to analyze human language and understand context means they apply to a range of tasks, including:<\/span><\/p>\n<p><b>Text summarization: <\/b><span style=\"font-weight: 400;\">Large language models are excellent at analyzing long pieces of text and producing coherent summaries, making it easier to extract key information from lengthy articles, documents, or reports.<\/span><\/p>\n<p><b>Chatbots and conversational assistants<\/b><span style=\"font-weight: 400;\">: LLMs help chatbots engage in natural language conversations with their users. In this case, they\u2019re often used for customer support, information retrieval, or task automation.<\/span><\/p>\n<p><b>Spelling and grammar correction:<\/b><span style=\"font-weight: 400;\"> LLMs can correct spelling and grammar errors in written text, improving readability with minimal effort.<\/span><\/p>\n<p><b>Translation:<\/b><span style=\"font-weight: 400;\"> Large language models are very good at translating text from one language to another, enabling effective communication regardless of linguistic barriers.<\/span><\/p>\n<p><b>Recommendation systems: <\/b><span style=\"font-weight: 400;\">Although we often associate large language models with consumer-oriented applications, they have uses in the business domain, as well.\u00a0 A great example is how online retailers and streaming services use them to analyze user preferences. They can then provide personalized recommendations for products (e.g. Amazon) movies (e.g. Netflix) and music (e.g. Spotify).<\/span><\/p>\n<p><b>Code generation:<\/b><span style=\"font-weight: 400;\"> LLMs can help programmers generate code, taking the heavy lifting out of their work and allowing them to focus on the more complex and creative aspects of the job.<\/span><\/p>\n<p><b>Image annotation:<\/b><span style=\"font-weight: 400;\"> Newer LLMs can recognize not just text, but image prompts. As such, they can generate descriptive captions and alt text, enhancing image accessibility and searchability.<\/span><\/p>\n<p><b>Data analytics:<\/b><span style=\"font-weight: 400;\"> While all tasks carried out by large language models technically involve analyzing data, the models can also support specific data analytics tasks, such as sentiment analysis. But we\u2019ll cover this in more detail in section 3.<\/span><\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-28483\" src=\"https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/llms-have-huge-value-but-also-come-with-risks.jpeg\" alt=\"A team of tech workers laughing in a meeting room.\" width=\"1200\" height=\"600\" title=\"\" srcset=\"https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/llms-have-huge-value-but-also-come-with-risks.jpeg 1200w, https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/llms-have-huge-value-but-also-come-with-risks-300x150.jpeg 300w, https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/llms-have-huge-value-but-also-come-with-risks-1024x512.jpeg 1024w, https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/llms-have-huge-value-but-also-come-with-risks-768x384.jpeg 768w\" sizes=\"auto, (max-width: 1200px) 100vw, 1200px\" \/><\/p>\n<p><span style=\"font-weight: 400;\">This limited list barely scratches the surface, and many applications of LLM-powered artificial intelligence systems are still emerging. Imagine, for example, a voice-powered virtual assistant (such as Alexa or Siri) integrated with a large language model.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">The level of speech recognition and natural response this could provide would represent a huge leap forward for this kind of consumer product. And you won\u2019t have to imagine it for long\u2014this application of LLMs is <\/span><a href=\"https:\/\/techreport.com\/news\/alexas-improved-llm-sets-new-standards-for-ai-voice-assistants\/\" rel=\"noopener\"><span style=\"font-weight: 400;\">already being developed<\/span><\/a><span style=\"font-weight: 400;\">.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Furthermore, with the speed of these models\u2019 evolution, and the ability of LLMs to create their own training data, further exciting use cases are no doubt just around the corner. Watch this space!<\/span><\/p>\n<h2 id=\"how-do-llms-work\"><span style=\"font-weight: 400;\">4. How do LLMs work?<\/span><\/h2>\n<p><span style=\"font-weight: 400;\">So we\u2019ve got a good feel for what LLMs are and what they can do. But how do they actually work?<\/span><\/p>\n<p><span style=\"font-weight: 400;\">At its core, an LLM is just a computer program that can understand and generate text. However, what goes on under the hood is quite complex. To achieve their objectives, LLMs use<\/span><a href=\"https:\/\/careerfoundry.inbearbeitung.de\/en\/blog\/data-analytics\/machine-learning-vs-deep-learning\/\"> <span style=\"font-weight: 400;\">deep learning<\/span><\/a><span style=\"font-weight: 400;\"> to train themselves on the nuances of human language so they can predict and output suitable responses.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Achieving this involves training large language models in two main phases: <strong>pretraining<\/strong> and<strong> fine-tuning<\/strong>.<\/span><\/p>\n<p><b>Learn more:<\/b><a href=\"https:\/\/careerfoundry.inbearbeitung.de\/en\/blog\/data-analytics\/machine-learning-models\/\"> <span style=\"font-weight: 400;\">A Beginner\u2019s Guide To Machine Learning Models<\/span><\/a><\/p>\n<h3>Pretraining phase<\/h3>\n<p><span style=\"font-weight: 400;\">The first phase is \u201cpretraining&#8221;. During this phase, the model is exposed to massive amounts of text-based training data from the internet and other sources. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">By applying statistical analysis and <a href=\"https:\/\/careerfoundry.inbearbeitung.de\/en\/blog\/data-analytics\/what-are-nlp-algorithms\/\" target=\"_blank\" rel=\"noopener\">natural language processing<\/a>, it \u201clearns\u201d grammar, vocabulary, and some general understanding of language.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">At this stage, a large language model won\u2019t know the specifics about individual documents in its training dataset, although it may pick up some facts from what it analyzes. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">This is the same to learning to read\u2014you have to consume a lot of books to perfect your language skills, and you\u2019ll pick up new words and facts along the way.<\/span><\/p>\n<h3>Fine-tuning phase<\/h3>\n<p><span style=\"font-weight: 400;\">After pretraining, LLMs can be tailored for more specific tasks by learning from particular examples and instructions.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Fine-tuning involves taking the basic knowledge that the model has learned from all of its training data and then teaching it to contextualize this to specific tasks such as answering questions, translating languages, or any of the other jobs associated with the use cases we went through earlier.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">This process allows engineers to customize large language models, controlling their output, and adapting them to specific domains. This is where LLMs will increasingly diversify over time, as organizations devise ever-more sophisticated new ways of applying them.<\/span><\/p>\n<h3 id=\"benefits-of-llms\"><span style=\"font-weight: 400;\">Benefits of LLMs<\/span><\/h3>\n<p><span style=\"font-weight: 400;\">We\u2019ve already covered some of the benefits of large language models, but here\u2019s a more detailed list:<\/span><\/p>\n<h4><span style=\"font-weight: 400;\">Natural language understanding<\/span><\/h4>\n<p><span style=\"font-weight: 400;\">LLMs are the first technology developed specifically to comprehend complex language structures, idioms, and context. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">This is valuable for a wide range of tasks, as we\u2019ve already explored.<\/span><\/p>\n<h4><span style=\"font-weight: 400;\">Improved productivity<\/span><\/h4>\n<p><span style=\"font-weight: 400;\">While there is much doom-mongering about AI automation replacing jobs, <\/span><a href=\"https:\/\/www.ben-evans.com\/benedictevans\/2023\/7\/2\/working-with-ai\" target=\"_blank\" rel=\"noopener\"><span style=\"font-weight: 400;\">in reality, they are helping improve productivity<\/span><\/a><span style=\"font-weight: 400;\">.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">By automating menial tasks like data entry and text analysis, LLMs free up workers to focus on more creative and impactful business tasks that require human-specific skills such as critical thinking and problem-solving.<\/span><\/p>\n<h4><span style=\"font-weight: 400;\">Customization<\/span><\/h4>\n<p><span style=\"font-weight: 400;\">As discussed, LLMs can be fine-tuned to complete tasks that align with specific tones, styles, target audiences, and even business domains.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">In future, this may enable businesses to create personalized AI tools with relative ease, even streamlining tasks that are highly specific to their business or domain area.<\/span><\/p>\n<h4><span style=\"font-weight: 400;\">Research<\/span><\/h4>\n<p><span style=\"font-weight: 400;\">LLMs are more than mere content creation tools\u2014their pattern recognition abilities are ideal for assisting researchers in analyzing vast amounts of information.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Large language models can provide quick access to relevant content and even suggest possible avenues for further research.<\/span><\/p>\n<h3 id=\"limitations-of-llms\"><span style=\"font-weight: 400;\">Limitations of LLMs:<\/span><\/h3>\n<p><span style=\"font-weight: 400;\">Right, so <a href=\"https:\/\/techcrunch.com\/2023\/06\/30\/ai-is-not-a-panacea-for-software-development\/\" target=\"_blank\" rel=\"noopener\">we won\u2019t pretend that LLMs are a panacea<\/a>, pouring only positive change into the world.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">As with any new technology, large language models also have some limitations and concerns. These include:<\/span><\/p>\n<h4><span style=\"font-weight: 400;\">The need for technical expertise<\/span><\/h4>\n<p><span style=\"font-weight: 400;\">Developing and fine-tuning LLMs requires <\/span><a href=\"https:\/\/careerfoundry.inbearbeitung.de\/en\/blog\/data-analytics\/what-is-machine-learning\/\" target=\"_blank\" rel=\"noopener\"><span style=\"font-weight: 400;\">high levels of technical knowledge in machine learning<\/span><\/a><span style=\"font-weight: 400;\">, natural language processing, and data preprocessing.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">While these skills are available, they aren\u2019t yet widely spread enough to meet the high demand. This makes them a great career choice, though!<\/span><\/p>\n<h4><span style=\"font-weight: 400;\">Financial and environmental impact<\/span><\/h4>\n<p><span style=\"font-weight: 400;\">Training and deploying LLMs requires huge computational resources, leading to high costs in terms of hardware and energy consumption.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">These costs <\/span><a href=\"https:\/\/www.nextplatform.com\/2022\/12\/01\/counting-the-cost-of-training-large-language-models\/\" target=\"_blank\" rel=\"noopener\"><span style=\"font-weight: 400;\">are financial<\/span><\/a><span style=\"font-weight: 400;\">, yes, but even more crucially they are also environmental. Deploying <\/span><a href=\"https:\/\/shrinkthatfootprint.com\/carbon-footprint-of-training-gpt-3-and-large-language-models\/\" target=\"_blank\" rel=\"noopener\"><span style=\"font-weight: 400;\">LLMs raises concerns about the carbon footprint<\/span><\/a><span style=\"font-weight: 400;\"> associated with extensive computing.<\/span><\/p>\n<h4><span style=\"font-weight: 400;\">Lack of data when fine-tuning<\/span><\/h4>\n<p><span style=\"font-weight: 400;\">Although LLMs are data-hungry, there are domains and languages with limited available data for fine-tuning.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">While this is not a huge problem yet, it has the potential to result in models producing suboptimal or, worse yet, biased outputs when dealing with specialized topics.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Expect to be hearing more about <\/span><a href=\"https:\/\/www.ft.com\/content\/053ee253-820e-453a-a1d5-0f24985258de\" target=\"_blank\" rel=\"noopener\"><span style=\"font-weight: 400;\">the issue of synthetic data<\/span><\/a><span style=\"font-weight: 400;\"> in future, where the data used in the pretraining phase was created by AIs themselves.<\/span><\/p>\n<h4><span style=\"font-weight: 400;\">Ethical concerns<\/span><\/h4>\n<p><span style=\"font-weight: 400;\">LLMs are trained on vast datasets that, if we\u2019re honest, often contain biased and harmful content. This means these biases can then be perpetuated by the model.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Furthermore, while the current commercial LLMs have \u201cguard rails\u201d in place, <\/span><a href=\"https:\/\/finance.yahoo.com\/news\/researchers-way-easily-bypass-guardrails-183009628.html?guccounter=1&amp;guce_referrer=aHR0cHM6Ly93d3cuZ29vZ2xlLmNvbS8&amp;guce_referrer_sig=AQAAAFOA8ZoTE686PbnO7WbZ6sMO9IhfBEiBl0qa4vwJOYcXO0U-gFB3ynC0q2RN-_du-tbHzPSIoyo0a0fwtw3eaAhtcIxrnWh6anrzmM-GC03I-EqBdn8u3hV91-_V1uVlUknN3r834v-SSMWMSbqwPX37RWikLt3tDqrp-hknHJfc\" target=\"_blank\" rel=\"noopener\"><span style=\"font-weight: 400;\">researchers have been able to manipulate these to produce harmful content<\/span><\/a><span style=\"font-weight: 400;\">. This has major implications for the creation and dissemination of hate speech, political propaganda, and misinformation.<\/span><\/p>\n<p>You can learn more about these ethical issues and how to be mindful of them in <a href=\"https:\/\/careerfoundry.inbearbeitung.de\/en\/blog\/data-analytics\/bias-in-machine-learning\/\" target=\"_blank\" rel=\"noopener\">our guide to bias in machine learning<\/a>.<\/p>\n<h4><span style=\"font-weight: 400;\">Accuracy<\/span><\/h4>\n<p><span style=\"font-weight: 400;\">While LLMs have impressive capabilities, they are not infallible.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">They often generate responses that seem plausible but are completely fabricated. This is especially concerning in domains like healthcare, law, and finance, where inaccurate misinformation can have serious consequences.<\/span><\/p>\n<h4><span style=\"font-weight: 400;\">Erosion of human skills<\/span><\/h4>\n<p><span style=\"font-weight: 400;\">One concern is that an overreliance on LLMs might, in the future, lead to the erasure of human expertise. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">A reduced value placed on human skills and creativity could mean the loss of vital skills such as critical thinking, emotional intelligence, and nuanced decision-making.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Fortunately, hoards of data scientists and <\/span><a href=\"https:\/\/careerfoundry.inbearbeitung.de\/en\/blog\/data-analytics\/machine-learning-engineer\/\" target=\"_blank\" rel=\"noopener\"><span style=\"font-weight: 400;\">machine learning engineers<\/span><\/a><span style=\"font-weight: 400;\"> are already looking at ways to solve these and other problems associated with large language models. But it\u2019s nevertheless necessary to be aware of them.<\/span><\/p>\n<h2 id=\"how-to-use-large-language-models-in-data-analysis\"><span style=\"font-weight: 400;\">5. How to use large language models in data analysis<\/span><\/h2>\n<p><span style=\"font-weight: 400;\">Among their many other uses, language models are ideally suited to streamlining various aspects <a href=\"https:\/\/careerfoundry.inbearbeitung.de\/en\/blog\/data-analytics\/the-data-analysis-process-step-by-step\/\" target=\"_blank\" rel=\"noopener\">of the data analytics process<\/a>. Here are some ways they can be applied in this field:<\/span><\/p>\n<h3><span style=\"font-weight: 400;\">Sentiment analysis<\/span><\/h3>\n<p><span style=\"font-weight: 400;\">Sentiment analysis is the process of determining the emotional tone of a piece of text, and whether it\u2019s positive, negative, or neutral.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Large language models excel at this, making them well-suited for such a task. For example, if you had the following customer review: \u201cThe product is amazing! I love it,\u201d an LLM could process the text and predict a sentiment label like \u201cpositive\u201d.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Marketing analysts, for example, could then use this to gain insights into public opinion about their products or services. Learn more in <a href=\"https:\/\/careerfoundry.inbearbeitung.de\/en\/blog\/data-analytics\/sentiment-analysis\/\" target=\"_blank\" rel=\"noopener\">our full guide to sentiment analysis<\/a>.<\/span><\/p>\n<h3><span style=\"font-weight: 400;\">Classification<\/span><\/h3>\n<p><span style=\"font-weight: 400;\">Like sentiment analysis, classification involves categorizing text data into predefined groups. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">Via fine-tuning, large language models can be trained to perform specific classification tasks such as spam detection, topic categorization, or customer support ticketing. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">Using this approach, data analysts can even use LLMs to categorize numbers and figures within a large spreadsheet, for example, saving themselves a lot of time in the data cleaning process.<\/span><\/p>\n<h3><span style=\"font-weight: 400;\">Code generation<\/span><\/h3>\n<p><span style=\"font-weight: 400;\">Most data analytics tasks require at least some level of coding. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">Fortunately, large language models can assist programmers with this too, generating snippets based on natural language prompts. This is particularly useful for speeding up the development of new algorithms. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">By simply describing the functionality they need, data analysts can reduce the time they spend trawling Python libraries or writing code from scratch.\u00a0<\/span><\/p>\n<p>There&#8217;s a whole host of<a href=\"https:\/\/careerfoundry.inbearbeitung.de\/en\/blog\/web-development\/ai-programming-tools\/\" target=\"_blank\" rel=\"noopener\"> AI programming tools<\/a> out there, with GitHub CoPilot being just one example.<\/p>\n<h3><span style=\"font-weight: 400;\">Information extraction<\/span><\/h3>\n<p><span style=\"font-weight: 400;\">Extracting relevant information from unstructured text data, such as news articles or research papers, can be very time-consuming when carried out manually. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">Large language models can generate concise summaries of lengthy documents, helping data analysts quickly grasp the main points they need without spending excessive time reading. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">This is particularly useful for industries that rely on analyzing complex technical documents, such as the finance or legal sectors.<\/span><\/p>\n<h3><span style=\"font-weight: 400;\">Named entity recognition (NER)<\/span><\/h3>\n<p><span style=\"font-weight: 400;\">Named entity recognition (NER) is a subtask of information extraction that involves identifying and classifying entities mentioned in a text, such as names of people, organizations, locations, dates, and more. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">This has obvious implications for data analytics, which often needs to distinguish between different data points. LLMs excel in these tasks due to their contextual understanding and language modeling capabilities.<\/span><\/p>\n<h3><span style=\"font-weight: 400;\">Healthcare diagnostics<\/span><\/h3>\n<p><span style=\"font-weight: 400;\">To underscore the potential of fine-tuning large language models, consider their role in addressing data analytics tasks within the healthcare domain. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">Fine-tuning a model on medical data can assist <a href=\"https:\/\/careerfoundry.inbearbeitung.de\/en\/blog\/data-analytics\/what-is-a-healthcare-data-analyst\/\" target=\"_blank\" rel=\"noopener\">healthcare analysts<\/a> in diagnosing medical conditions based on symptoms and patient history. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">They can also be used to support more administrative tasks such as patient appointment scheduling.<\/span><\/p>\n<h2><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-28485\" src=\"https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/llms-are-particularly-useful-for-healthcare-analyst.jpeg\" alt=\"A healthcare data analyst works in an office with LLMs.\" width=\"1200\" height=\"714\" title=\"\" srcset=\"https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/llms-are-particularly-useful-for-healthcare-analyst.jpeg 1200w, https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/llms-are-particularly-useful-for-healthcare-analyst-300x179.jpeg 300w, https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/llms-are-particularly-useful-for-healthcare-analyst-1024x609.jpeg 1024w, https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-content\/uploads\/2023\/08\/llms-are-particularly-useful-for-healthcare-analyst-768x457.jpeg 768w\" sizes=\"auto, (max-width: 1200px) 100vw, 1200px\" \/><\/h2>\n<h2 id=\"large-language-models-faq\"><span style=\"font-weight: 400;\">6. Large language models FAQs<\/span><\/h2>\n<p><span style=\"font-weight: 400;\">Now we\u2019ve covered all the must-know information, you might have some more questions about large language models!\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Here are answers to some of the most common ones.<\/span><\/p>\n<h3><span style=\"font-weight: 400;\">What is the best large language model?<\/span><\/h3>\n<p><span style=\"font-weight: 400;\">Deciding the \u201cbest\u201d large language model is a bit like selecting the \u201cbest\u201d flavor of fruit\u2014it&#8217;s subjective and depends on the task at hand. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">Models like GPT-4 have garnered attention for their impressive performance across various natural language processing tasks. However, the definition of \u201cbest\u201d still depends on factors like the model\u2019s intended use, task complexity, and so on.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">As large language models evolve, those tailored to specific tasks will emerge. For creative text generation, GPT-4 might excel, but other models could shine in fields like sentiment analysis or medical diagnostics. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">In short, <strong>the best LLM model will ultimately depend on the use case<\/strong>.<\/span><\/p>\n<h3><span style=\"font-weight: 400;\">How are large language models different from natural language processing?<\/span><\/h3>\n<p><span style=\"font-weight: 400;\">Natural language processing (NLP) is a broad field that studies how computers understand and process human language.<\/span><\/p>\n<p><span style=\"font-weight: 400;\"> NLP might involve using various techniques, including rule-based methods, machine learning, and deep learning, to process, analyze, and generate human language.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Meanwhile, <strong>LLMs are a specific application of NLP<\/strong>. As an advanced machine learning model, a large language model\u2019s massive amount of text-based training data makes it ideally suited to complex natural language processing tasks, like those described throughout this blog post. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">Essentially, while NLP encapsulates a comprehensive study of language and its computational aspects, LLMs exemplify a pinnacle of this endeavor, wielding their extensive training data to accomplish intricate linguistic feats.<\/span><\/p>\n<h3><span style=\"font-weight: 400;\">What is the largest language model in the world?<\/span><\/h3>\n<p><span style=\"font-weight: 400;\">The largest publicly-known language model is currently OpenAI&#8217;s GPT-4 (Generative Pre-trained Transformer 4), which, it&#8217;s rumored, <strong>has a massive 1.7 trillion parameters<\/strong> (parameters being the learned weights and biases that determine how the model understands context and responds to different inputs). <\/span><\/p>\n<p><span style=\"font-weight: 400;\">Considering that GPT-3 had a <\/span><span style=\"font-weight: 400;\">\u201c<\/span><span style=\"font-weight: 400;\">mere<\/span><span style=\"font-weight: 400;\">\u201d<\/span><span style=\"font-weight: 400;\"> 175 billion parameters, this tells you just how fast these large language models are growing in size and complexity. Larger models will no doubt soon emerge.<\/span><\/p>\n<h2><span style=\"font-weight: 400;\">7. Final thoughts<\/span><\/h2>\n<p><span style=\"font-weight: 400;\">There we have it! Everything you need to know about large language models.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">In this post, we\u2019ve explored the new technology driving the AI revolution. As we\u2019ve seen, the emergence of LLMs marks a significant milestone in the advancement of artificial intelligence, with many hailing it as the greatest change in society since the industrial revolution.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Whether or not this hyperbole stands the test of time remains to be seen. But it\u2019s already clear that large language models transcend their reputation as mere text generators, finding applications in fields ranging from marketing and healthcare to data analytics.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">From sentiment analysis to classification tasks, the ability of large language models to decipher context and nuances will empower evermore accurate and efficient data processing. And their role in code generation is already speeding up programming tasks, improving the efficiency of data scientists the world over.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">If you\u2019re curious to learn more about what a possible career in data analytics or data science might involve, or simply want to capitalize on the potential of AI within this field, why not check out CareerFoundry&#8217;s <\/span><strong><a href=\"https:\/\/careerfoundry.inbearbeitung.de\/en\/short-courses\/become-a-data-analyst\/?popup-tracking=WYSDN-short-course-DAT\">free, 5-day data analytics short course<\/a><\/strong><span style=\"font-weight: 400;\">? Or, if you prefer to read on, check out the following introductory guides to learn more:<\/span><\/p>\n<ul>\n<li><a href=\"https:\/\/careerfoundry.inbearbeitung.de\/en\/blog\/data-analytics\/best-machine-learning-languages\/\">What\u2019s the Best Language for Machine Learning?<\/a><\/li>\n<li><a href=\"https:\/\/careerfoundry.inbearbeitung.de\/en\/blog\/data-analytics\/machine-learning-engineer-salary\/\">What\u2019s the Average Machine Learning Engineer Salary?<\/a><\/li>\n<li><a href=\"https:\/\/careerfoundry.inbearbeitung.de\/en\/blog\/web-development\/ai-and-web-development\/\">Why AI Won&#8217;t Replace Web Developers Anytime Soon<\/a><\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"<p>Large language models, or LLMs, are the power behind the current AI buzz. Learn what they are, how they work, and how they help data analysts.<\/p>\n","protected":false},"author":101,"featured_media":28521,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"_lmt_disableupdate":"yes","_lmt_disable":"","footnotes":""},"categories":[3],"tags":[],"class_list":["post-28403","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-data-analytics"],"acf":{"homepage_category_featured":false,"cards_inner_programs_lists_right":"","cards_inner_programs_lists_left":"","related_plan_cards":""},"modified_by":"Matthew Deery","_links":{"self":[{"href":"https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-json\/wp\/v2\/posts\/28403","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-json\/wp\/v2\/users\/101"}],"replies":[{"embeddable":true,"href":"https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-json\/wp\/v2\/comments?post=28403"}],"version-history":[{"count":9,"href":"https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-json\/wp\/v2\/posts\/28403\/revisions"}],"predecessor-version":[{"id":33422,"href":"https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-json\/wp\/v2\/posts\/28403\/revisions\/33422"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-json\/wp\/v2\/media\/28521"}],"wp:attachment":[{"href":"https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-json\/wp\/v2\/media?parent=28403"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-json\/wp\/v2\/categories?post=28403"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/careerfoundry.inbearbeitung.de\/en\/wp-json\/wp\/v2\/tags?post=28403"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}