{"id":553,"date":"2023-03-16T00:00:00","date_gmt":"2023-03-16T00:00:00","guid":{"rendered":"https:\/\/www.nextias.com\/current_affairs\/uncategorized\/16-03-2023\/what-is-gpt-4-and-how-is-it-different-from-chatgpt\/"},"modified":"2023-03-16T00:00:00","modified_gmt":"2023-03-16T00:00:00","slug":"what-is-gpt-4-and-how-is-it-different-from-chatgpt","status":"publish","type":"post","link":"https:\/\/www.nextias.com\/ca\/current-affairs\/16-03-2023\/what-is-gpt-4-and-how-is-it-different-from-chatgpt","title":{"rendered":"What is GPT-4 and how is it different from ChatGPT?"},"content":{"rendered":"<p><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"><strong><u>In News<\/u><\/strong><\/span><\/span><\/span><\/p>\n<ul>\n<li style=\"list-style-type:disc\"><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\">OpenAI has announced <\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"><strong>GPT4,<\/strong><\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"> t<\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"><strong>he latest version of its large language model<\/strong><\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"> that powers key applications like ChatGPT and the new Bing.<\/span><\/span><\/span><\/li>\n<\/ul>\n<p><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"><strong><u>What is ChatGPT?<\/u><\/strong><\/span><\/span><\/span><\/p>\n<ul>\n<li style=\"list-style-type:disc\"><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\">ChatGPT is an <\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"><strong>artificial-intelligence chatbot<\/strong><\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"> developed by San Francisco-based AI research company <\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"><strong>OpenAI in 2022.<\/strong><\/span><\/span><\/span><\/li>\n<li style=\"list-style-type:disc\"><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\">It is a <\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"><strong>trained model<\/strong><\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"> which interacts in a <\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"><strong>conversational way.<\/strong><\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"> The dialogue format makes it possible to answer followup questions, admit its mistakes, challenge incorrect premises, and reject inappropriate requests.<\/span><\/span><\/span><\/li>\n<li style=\"list-style-type:disc\"><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\">It can have <\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"><strong>conversations on topics<\/strong><\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"> from history to philosophy, generate lyrics and suggest edits to computer programming code.<\/span><\/span><\/span><\/li>\n<\/ul>\n<p><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"><strong><u>Technology Used<\/u><\/strong><\/span><\/span><\/span><\/p>\n<ul>\n<li style=\"list-style-type:disc\"><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\">The technology that underlies ChatGPT is referenced in the second half of its name, GPT, which stands for <\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"><strong>Generative Pre-trained Transformer.\u00a0<\/strong><\/span><\/span><\/span><\/li>\n<li style=\"list-style-type:disc\"><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\">Transformers are <\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"><strong>specialized algorithms <\/strong><\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\">for finding <\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"><strong>long-range patterns<\/strong><\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"> in sequences of data.\u00a0<\/span><\/span><\/span><\/li>\n<li style=\"list-style-type:disc\"><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\">A transformer <\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"><strong>learns to predict<\/strong><\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"> not just the next word in a sentence but also the next sentence in a paragraph and the next paragraph in an essay. This is what allows it to <\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"><strong>stay on topic for long stretches of text.<\/strong><\/span><\/span><\/span><\/li>\n<\/ul>\n<p><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"><strong><u>Shortcomings of the Previous Models<\/u><\/strong><\/span><\/span><\/span><\/p>\n<ul>\n<li style=\"list-style-type:disc\"><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"><strong>The chatbot isn\u2019t always accurate: <\/strong><\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\">Their sources aren\u2019t fact-checked, and they rely on human feedback to improve its accuracy. They can also get facts mixed up and produce misinformation.<\/span><\/span><\/span><\/li>\n<li style=\"list-style-type:disc\"><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\">GPT-3 and ChatGPT\u2019s GPT-3.5 were limited to textual input and output, meaning they <\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"><strong>could only read and write.\u00a0<\/strong><\/span><\/span><\/span><\/li>\n<li style=\"list-style-type:disc\"><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\">GPT-3 and GPT-3.5 only operated in one modality, text, meaning users<\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"><strong> could only ask questions by typing them out.<\/strong><\/span><\/span><\/span><\/li>\n<\/ul>\n<p><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"><strong><u>What is GPT-4?<\/u><\/strong><\/span><\/span><\/span><\/p>\n<ul>\n<li style=\"list-style-type:disc\"><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\">GPT-4 is a large <\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"><strong>multimodal model<\/strong><\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"> which means it can encompass more than just text and<\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"><strong> accepts images as inputs.<\/strong><\/span><\/span><\/span><\/li>\n<li style=\"list-style-type:disc\"><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\">GPT-4 also exhibits <\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"><strong>human-level performance<\/strong><\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"> on various professional and academic benchmarks.<\/span><\/span><\/span><\/li>\n<li style=\"list-style-type:disc\"><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\">For example, it can answer tax-related questions, schedule a meeting among three busy people, or learn a user\u2019s creative writing style.<\/span><\/span><\/span><\/li>\n<li style=\"list-style-type:disc\"><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\">GPT-4 is also capable of<\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"><strong> handling over 25,000 words of text, <\/strong><\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\">opening up a greater number of use cases that now also include <\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"><strong>long-form content creation,<\/strong><\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"> <\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"><strong>document search and analysis, and extended conversations.<\/strong><\/span><\/span><\/span><\/li>\n<\/ul>\n<p><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"><strong><u>How is GPT-4 different from GPT-3?<\/u><\/strong><\/span><\/span><\/span><\/p>\n<ul>\n<li style=\"list-style-type:disc\"><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"><strong>GPT-4 can see images now:<\/strong><\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"> GPT-4 can be fed images and asked to output information accordingly whereas GPT-3 and ChatGPT\u2019s GPT-3.5 were limited to textual input and output.<\/span><\/span><\/span><\/li>\n<li style=\"list-style-type:disc\"><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\">GPT-4 can process a lot more information at a time.<\/span><\/span><\/span><\/li>\n<li style=\"list-style-type:disc\"><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"><strong>GPT-4 has an improved accuracy:<\/strong><\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"> GPT-4 significantly reduces hallucinations relative to previous models and scores 40 per cent higher than GPT-3.5 on factuality evaluations.\u00a0<\/span><\/span><\/span><\/li>\n<li style=\"list-style-type:disc\"><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"><strong>Hard to trick: <\/strong><\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\">It will be a lot harder to trick GPT-4 into producing undesirable outputs such as hate speech and misinformation.<\/span><\/span><\/span><\/li>\n<li style=\"list-style-type:disc\"><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\"><strong>Better at understanding languages other than English: <\/strong><\/span><\/span><\/span><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#000000\">GPT-4 is more multilingual which means that users will be able to use chatbots based on GPT-4 to produce outputs with greater clarity and higher accuracy in their native languages.<\/span><\/span><\/span><\/li>\n<\/ul>\n<p><a href=\"https:\/\/indianexpress.com\/article\/explained\/explained-sci-tech\/gpt-4-new-different-chatgpt-openai-bing-explained-8500024\/\" style=\"text-decoration:none\" target=\"_blank\" rel=\"noopener\"><span style=\"font-size:12pt\"><span style=\"font-family:'Book Antiqua',serif\"><span style=\"color:#1155cc\"><u>IE<\/u><\/span><\/span><\/span><\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>In News OpenAI has announced GPT4, the latest version of its large language model that powers key applications like ChatGPT and the new Bing. What is ChatGPT? ChatGPT is an artificial-intelligence chatbot developed by San Francisco-based AI research company OpenAI in 2022. It is a trained model which interacts in a conversational way. The dialogue [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":554,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[21],"tags":[26,33],"class_list":["post-553","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-current-affairs","tag-gs-3","tag-science-technology"],"acf":[],"jetpack_featured_media_url":"https:\/\/wp-images.nextias.com\/cdn-cgi\/image\/format=auto\/ca\/uploads\/2023\/07\/411174Screenshot_6.png","_links":{"self":[{"href":"https:\/\/www.nextias.com\/ca\/wp-json\/wp\/v2\/posts\/553","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.nextias.com\/ca\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.nextias.com\/ca\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.nextias.com\/ca\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.nextias.com\/ca\/wp-json\/wp\/v2\/comments?post=553"}],"version-history":[{"count":0,"href":"https:\/\/www.nextias.com\/ca\/wp-json\/wp\/v2\/posts\/553\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.nextias.com\/ca\/wp-json\/wp\/v2\/media\/554"}],"wp:attachment":[{"href":"https:\/\/www.nextias.com\/ca\/wp-json\/wp\/v2\/media?parent=553"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.nextias.com\/ca\/wp-json\/wp\/v2\/categories?post=553"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.nextias.com\/ca\/wp-json\/wp\/v2\/tags?post=553"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}