{"id":24882,"date":"2024-12-11T09:37:50","date_gmt":"2024-12-11T01:37:50","guid":{"rendered":"https:\/\/www.1ai.net\/?p=24882"},"modified":"2024-12-11T09:41:02","modified_gmt":"2024-12-11T01:41:02","slug":"lg-%e5%8f%91%e5%b8%83-exaone-3-5-%e5%bc%80%e6%ba%90-ai-%e6%a8%a1%e5%9e%8b%ef%bc%9a%e9%95%bf%e6%96%87%e6%9c%ac%e5%a4%84%e7%90%86%e5%88%a9%e5%99%a8%e3%80%81%e7%8b%ac%e7%89%b9%e6%8a%80%e6%9c%af%e6%9c%89","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/24882.html","title":{"rendered":"LG Releases EXAONE 3.5 Open Source AI Model: Long Text Processing Tool, Unique Technology to Reduce \"Hallucinations\""},"content":{"rendered":"<p>December 11th.<a href=\"https:\/\/www.1ai.net\/en\/tag\/lg\" title=\"_OTHER ORGANISER\" target=\"_blank\" >LG<\/a> Artificial Intelligence Research Institute releases EXAONE 3.5 open source this Monday, December 9 <a href=\"https:\/\/www.1ai.net\/en\/tag\/ai%e6%a8%a1%e5%9e%8b\" title=\"[View articles tagged with [AI models]]\" target=\"_blank\" >AI Models<\/a>and the simultaneous launch of enterprise-grade AI for LG employees. <a href=\"https:\/\/www.1ai.net\/en\/tag\/%e6%99%ba%e8%83%bd%e4%bd%93\" title=\"[View articles tagged with [intelligent body]]\" target=\"_blank\" >Agent<\/a>Service ChatEXAONE.<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-24883\" title=\"cc6f921bj00sob35u003vd000sg00fdp\" src=\"https:\/\/www.1ai.net\/wp-content\/uploads\/2024\/12\/cc6f921bj00sob35u003vd000sg00fdp.jpg\" alt=\"cc6f921bj00sob35u003vd000sg00fdp\" width=\"1024\" height=\"553\" \/><\/p>\n<p><strong>EXAONE 3.5<\/strong><\/p>\n<p>This release of EXAONE 3.5 comes just four months after version 3.0, and the new model is available in three versions:<strong>An ultra-lightweight device-side model with 2.4 billion parameters, a lightweight general-purpose model with 7.8 billion parameters, and a high-performance specialized model with 32 billion parameters.<\/strong><\/p>\n<p>LG claims that EXAONE 3.5 processes 100 pages of text in a single pass and excels in 20 benchmarks, leading other open-source AI models especially in practical usability, long text processing, coding and math.<\/p>\n<p>The highlight of EXAONE 3.5 is the unique search-enhanced generation technology and multi-step reasoning capability, which effectively reduces the \"illusion\" phenomenon and improves answer accuracy.<\/p>\n<p>LG AI Research plans to expand the context window of EXAONE 3.5 for tokens from 32,000 to 128,000 in the first half of 2025, and will release a large-scale action model (LAM) in the same year to further enhance AI capabilities and continue to lead the innovation in the AI field.<\/p>\n<p><strong>ChatEXAONE<\/strong><\/p>\n<p>More than just a standard AI intelligence, ChatEXAONE is a powerful enterprise-grade tool with sophisticated query analysis and user-defined search capabilities, equipped with encryption and privacy-protecting technology to ensure secure use within the company.<\/p>\n<p data-vmark=\"f0c2\"><span class=\"referenceTitle\">1AI Attach reference address<\/span><\/p>\n<ul class=\"custom_reference list-paddingleft-1\">\n<li class=\"list-undefined list-reference-paddingleft\">\n<p data-vmark=\"c66b\"><a href=\"https:\/\/www.lgresearch.ai\/blog\/view?seq=507\" target=\"_blank\" rel=\"noopener\">Open-sourcing Three EXAONE 3.5 Models : Frontier-level Model, Top-tier Performance in Instruction Following and Long Context Capabilities<\/a><\/p>\n<\/li>\n<li class=\"list-undefined list-reference-paddingleft\">\n<p data-vmark=\"0b54\"><a href=\"https:\/\/arxiv.org\/abs\/2412.04862\" target=\"_blank\" rel=\"noopener\">EXAONE 3.5: Series of Large Language Models for Real-world Use Cases<\/a><\/p>\n<\/li>\n<li class=\"list-undefined list-reference-paddingleft\">\n<p data-vmark=\"8f10\"><a href=\"https:\/\/huggingface.co\/LGAI-EXAONE\" target=\"_blank\" rel=\"noopener\">LGAI-EXAONE<\/a><\/p>\n<\/li>\n<\/ul>","protected":false},"excerpt":{"rendered":"<p>On December 11, the LG Institute of Artificial Intelligence released the EXAONE 3.5 Open Source AI model on Monday (9 December) and simultaneously launched the Enterprise AI Smart Service for LG employees. EXAONE 3.5 The new model provides three versions of EXAONE 3.5 months away from version 3.0: the superweight end-end model for 2.4 billion parameters, the lightweight generic model for 7.8 billion parameters and the high performance special model for 32 billion parameters. LG claims that EXAONE 3.5 One-time processing 100-page text performed excellently in 20 benchmark tests, especially in terms of physical availability, long text processing<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[146],"tags":[167,1416,1355],"collection":[],"class_list":["post-24882","post","type-post","status-publish","format-standard","hentry","category-news","tag-ai","tag-lg","tag-1355"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/24882","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/comments?post=24882"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/24882\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=24882"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=24882"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=24882"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=24882"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}