{"id":7671,"date":"2023-11-23T15:10:15","date_gmt":"2023-11-23T15:10:15","guid":{"rendered":"https:\/\/dailyai.com\/?p=7671"},"modified":"2023-11-23T15:13:38","modified_gmt":"2023-11-23T15:13:38","slug":"system-2-attention-improves-accuracy-of-llm-responses","status":"publish","type":"post","link":"https:\/\/dailyai.com\/de\/2023\/11\/system-2-attention-improves-accuracy-of-llm-responses\/","title":{"rendered":"System 2 Aufmerksamkeit verbessert die Genauigkeit der LLM-Antworten"},"content":{"rendered":"<p><strong>Gro\u00dfe Sprachmodelle (Large Language Models, LLM) werden oft durch Verzerrungen oder irrelevanten Kontext in einer Eingabeaufforderung in die Irre gef\u00fchrt. Forscher von Meta haben einen scheinbar einfachen Weg gefunden, dies zu beheben.<\/strong><\/p>\n<p>Wenn die Kontextfenster gr\u00f6\u00dfer werden, k\u00f6nnen die Aufforderungen, die wir in ein LLM eingeben, l\u00e4nger und immer detaillierter werden. LLMs sind besser darin geworden, die Nuancen oder kleineren Details in unseren Aufforderungen zu erkennen, aber manchmal kann sie das verwirren.<\/p>\n<p>Fr\u00fche maschinelle Lernverfahren verwendeten einen \"Hard-Attention\"-Ansatz, bei dem der relevanteste Teil einer Eingabe herausgefiltert und nur auf diesen reagiert wurde. Das funktioniert gut, wenn Sie versuchen, ein Bild zu beschriften, aber schlecht, wenn Sie einen Satz \u00fcbersetzen oder eine vielschichtige Frage beantworten wollen.<\/p>\n<p>Die meisten LLMs verwenden heute einen \"Soft-Attention\"-Ansatz, bei dem der gesamte Prompt mit Token versehen und gewichtet wird.<\/p>\n<p>Meta schl\u00e4gt einen Ansatz vor, der <a href=\"https:\/\/arxiv.org\/pdf\/2311.11829.pdf\" target=\"_blank\" rel=\"noopener\">System 2 Aufmerksamkeit<\/a> (S2A), um das Beste aus beiden Welten zu erhalten. S2A nutzt die nat\u00fcrlichen Sprachverarbeitungsf\u00e4higkeiten eines LLM, um Ihre Eingabeaufforderung aufzugreifen und voreingenommene und irrelevante Informationen zu entfernen, bevor es an die Arbeit geht, eine Antwort zu geben.<\/p>\n<p>Hier ist ein Beispiel.<\/p>\n<figure id=\"attachment_7673\" aria-describedby=\"caption-attachment-7673\" style=\"width: 1200px\" class=\"wp-caption aligncenter\"><img loading=\"lazy\" decoding=\"async\" class=\"wp-image-7673 size-full\" src=\"https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/S2A-Math-example.png\" alt=\"\" width=\"1200\" height=\"750\" srcset=\"https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/S2A-Math-example.png 1200w, https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/S2A-Math-example-300x188.png 300w, https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/S2A-Math-example-1024x640.png 1024w, https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/S2A-Math-example-768x480.png 768w, https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/S2A-Math-example-370x231.png 370w, https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/S2A-Math-example-800x500.png 800w, https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/S2A-Math-example-20x13.png 20w, https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/S2A-Math-example-740x463.png 740w, https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/S2A-Math-example-77x48.png 77w\" sizes=\"auto, (max-width: 1200px) 100vw, 1200px\" \/><figcaption id=\"caption-attachment-7673\" class=\"wp-caption-text\">S2A Math Beispiel. Quelle: arXiv<\/figcaption><\/figure>\n<p>S2A l\u00f6scht die Informationen zu Max, da sie f\u00fcr die Frage irrelevant sind. S2A generiert eine optimierte Eingabeaufforderung neu, bevor es mit der Bearbeitung der Frage beginnt. LLMs sind notorisch schlecht in <a href=\"https:\/\/dailyai.com\/de\/2023\/10\/chatgpts-accounting-skills-are-put-to-the-test\/\">Mathe<\/a> Daher ist es eine gro\u00dfe Hilfe, die Aufforderung weniger verwirrend zu gestalten.<\/p>\n<p>LLMs sind sehr menschenfreundlich und stimmen Ihnen gerne zu, selbst wenn Sie falsch liegen. S2A entfernt alle Verzerrungen in einer Eingabeaufforderung und verarbeitet dann nur die relevanten Teile der Aufforderung. Dies reduziert das, was KI-Forscher als \"Kriecherei\" bezeichnen, oder die Neigung eines KI-Modells zum Arschkriechen.<\/p>\n<figure id=\"attachment_7674\" aria-describedby=\"caption-attachment-7674\" style=\"width: 1190px\" class=\"wp-caption aligncenter\"><img loading=\"lazy\" decoding=\"async\" class=\"wp-image-7674 size-full\" src=\"https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/S2A-sycophancy-reduction.png\" alt=\"\" width=\"1190\" height=\"584\" srcset=\"https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/S2A-sycophancy-reduction.png 1190w, https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/S2A-sycophancy-reduction-300x147.png 300w, https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/S2A-sycophancy-reduction-1024x503.png 1024w, https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/S2A-sycophancy-reduction-768x377.png 768w, https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/S2A-sycophancy-reduction-370x182.png 370w, https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/S2A-sycophancy-reduction-800x393.png 800w, https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/S2A-sycophancy-reduction-740x363.png 740w, https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/S2A-sycophancy-reduction-20x10.png 20w, https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/S2A-sycophancy-reduction-98x48.png 98w\" sizes=\"auto, (max-width: 1190px) 100vw, 1190px\" \/><figcaption id=\"caption-attachment-7674\" class=\"wp-caption-text\">S2A sycophancy Reduktion. Quelle: arXiv<\/figcaption><\/figure>\n<p>S2A ist eigentlich nur eine Systemaufforderung, die den LLM anweist, die urspr\u00fcngliche Aufforderung ein wenig zu verfeinern, bevor er sich an die Arbeit macht. Die Ergebnisse, die die Forscher mit Mathematik-, Sach- und Langformfragen erzielten, waren beeindruckend.<\/p>\n<p>Hier ein Beispiel f\u00fcr die Verbesserungen, die S2A bei Sachfragen erzielte. Die Ausgangsbasis waren Antworten auf Fragen, die Verzerrungen enthielten, w\u00e4hrend die Oracle-Aufforderung eine von Menschen verfeinerte ideale Aufforderung war.<\/p>\n<p>S2A kommt den Ergebnissen der Oracle-Eingabeaufforderung sehr nahe und bietet eine Verbesserung der Genauigkeit um fast 50% gegen\u00fcber der Grundeinstellung.<\/p>\n<figure id=\"attachment_7675\" aria-describedby=\"caption-attachment-7675\" style=\"width: 586px\" class=\"wp-caption aligncenter\"><img loading=\"lazy\" decoding=\"async\" class=\"wp-image-7675 size-full\" src=\"https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/S2A-Comparison-of-results.png\" alt=\"\" width=\"586\" height=\"342\" srcset=\"https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/S2A-Comparison-of-results.png 586w, https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/S2A-Comparison-of-results-300x175.png 300w, https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/S2A-Comparison-of-results-370x216.png 370w, https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/S2A-Comparison-of-results-20x12.png 20w, https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/S2A-Comparison-of-results-82x48.png 82w\" sizes=\"auto, (max-width: 586px) 100vw, 586px\" \/><figcaption id=\"caption-attachment-7675\" class=\"wp-caption-text\">S2A Vergleich der Ergebnisse. Quelle: arXiv<\/figcaption><\/figure>\n<p>Wo ist also der Haken? Durch die Vorverarbeitung der urspr\u00fcnglichen Eingabeaufforderung vor der Beantwortung entstehen zus\u00e4tzliche Rechenanforderungen f\u00fcr den Prozess. Wenn die Eingabeaufforderung lang ist und viele relevante Informationen enth\u00e4lt, kann die Neuerstellung der Eingabeaufforderung erhebliche Kosten verursachen.<\/p>\n<p>Es ist unwahrscheinlich, dass die Nutzer besser darin werden, gut formulierte Prompts zu schreiben, so dass S2A ein guter Weg sein k\u00f6nnte, dies zu umgehen.<\/p>\n<p>Wird Meta S2A in sein Programm aufnehmen? <a href=\"https:\/\/dailyai.com\/de\/2023\/07\/meta-and-microsoft-release-advanced-ai-llama-2-for-free\/\">Lama<\/a> Modell? Wir wissen es nicht, aber Sie k\u00f6nnen den S2A-Ansatz selbst nutzen.<\/p>\n<p>Wenn Sie darauf achten, dass Sie in Ihren Aufforderungen keine Meinungen oder Vorschl\u00e4ge machen, ist es wahrscheinlicher, dass Sie genaue Antworten von diesen Modellen erhalten.<\/p>","protected":false},"excerpt":{"rendered":"<p>Gro\u00dfe Sprachmodelle (Large Language Models, LLM) werden oft durch Verzerrungen oder irrelevanten Kontext in einer Eingabeaufforderung in die Irre gef\u00fchrt. Meta-Forscher haben einen scheinbar einfachen Weg gefunden, dies zu beheben. Wenn die Kontextfenster gr\u00f6\u00dfer werden, k\u00f6nnen die Eingabeaufforderungen, die wir in ein LLM eingeben, l\u00e4nger und immer detaillierter werden. LLMs sind besser darin geworden, Nuancen oder kleinere Details in unseren Aufforderungen zu erkennen, aber manchmal kann sie das verwirren. Fr\u00fche maschinelle Lernverfahren verwendeten einen Ansatz der \"harten Aufmerksamkeit\", bei dem der wichtigste Teil einer Eingabe herausgegriffen wurde und nur auf diesen reagiert wurde. Das funktioniert gut, wenn Sie versuchen, ein Bild zu beschriften,<\/p>","protected":false},"author":6,"featured_media":7676,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[84],"tags":[118,131],"class_list":["post-7671","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-industry","tag-llms","tag-meta"],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v27.4 - https:\/\/yoast.com\/product\/yoast-seo-wordpress\/ -->\n<title>System 2 Attention improves accuracy of LLM responses | DailyAI<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/dailyai.com\/de\/2023\/11\/system-2-attention-improves-accuracy-of-llm-responses\/\" \/>\n<meta property=\"og:locale\" content=\"de_DE\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"System 2 Attention improves accuracy of LLM responses | DailyAI\" \/>\n<meta property=\"og:description\" content=\"Large Language Models (LLM) are often mislead by bias or irrelevant context in a prompt. Researchers at Meta have found a seemingly simple way to fix that. As context windows increase the prompts that we enter into an LLM can become longer and increasingly detailed. LLMs have become better at picking up on the nuances or smaller details in our prompts, but sometimes this can confuse them. Early machine learning used a \u201chard attention\u201d approach that singled out the most relevant part of an input and responded only to that. This works fine when you\u2019re trying to caption an image,\" \/>\n<meta property=\"og:url\" content=\"https:\/\/dailyai.com\/de\/2023\/11\/system-2-attention-improves-accuracy-of-llm-responses\/\" \/>\n<meta property=\"og:site_name\" content=\"DailyAI\" \/>\n<meta property=\"article:published_time\" content=\"2023-11-23T15:10:15+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2023-11-23T15:13:38+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/Simplify.jpg\" \/>\n\t<meta property=\"og:image:width\" content=\"1000\" \/>\n\t<meta property=\"og:image:height\" content=\"666\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/jpeg\" \/>\n<meta name=\"author\" content=\"Eugene van der Watt\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:creator\" content=\"@DailyAIOfficial\" \/>\n<meta name=\"twitter:site\" content=\"@DailyAIOfficial\" \/>\n<meta name=\"twitter:label1\" content=\"Verfasst von\" \/>\n\t<meta name=\"twitter:data1\" content=\"Eugene van der Watt\" \/>\n\t<meta name=\"twitter:label2\" content=\"Gesch\u00e4tzte Lesezeit\" \/>\n\t<meta name=\"twitter:data2\" content=\"3\u00a0Minuten\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"NewsArticle\",\"@id\":\"https:\\\/\\\/dailyai.com\\\/2023\\\/11\\\/system-2-attention-improves-accuracy-of-llm-responses\\\/#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/dailyai.com\\\/2023\\\/11\\\/system-2-attention-improves-accuracy-of-llm-responses\\\/\"},\"author\":{\"name\":\"Eugene van der Watt\",\"@id\":\"https:\\\/\\\/dailyai.com\\\/#\\\/schema\\\/person\\\/7ce525c6d0c79838b7cc7cde96993cfa\"},\"headline\":\"System 2 Attention improves accuracy of LLM responses\",\"datePublished\":\"2023-11-23T15:10:15+00:00\",\"dateModified\":\"2023-11-23T15:13:38+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/dailyai.com\\\/2023\\\/11\\\/system-2-attention-improves-accuracy-of-llm-responses\\\/\"},\"wordCount\":520,\"publisher\":{\"@id\":\"https:\\\/\\\/dailyai.com\\\/#organization\"},\"image\":{\"@id\":\"https:\\\/\\\/dailyai.com\\\/2023\\\/11\\\/system-2-attention-improves-accuracy-of-llm-responses\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/dailyai.com\\\/wp-content\\\/uploads\\\/2023\\\/11\\\/Simplify.jpg\",\"keywords\":[\"LLMS\",\"Meta\"],\"articleSection\":[\"Industry\"],\"inLanguage\":\"de\"},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/dailyai.com\\\/2023\\\/11\\\/system-2-attention-improves-accuracy-of-llm-responses\\\/\",\"url\":\"https:\\\/\\\/dailyai.com\\\/2023\\\/11\\\/system-2-attention-improves-accuracy-of-llm-responses\\\/\",\"name\":\"System 2 Attention improves accuracy of LLM responses | DailyAI\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/dailyai.com\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/dailyai.com\\\/2023\\\/11\\\/system-2-attention-improves-accuracy-of-llm-responses\\\/#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/dailyai.com\\\/2023\\\/11\\\/system-2-attention-improves-accuracy-of-llm-responses\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/dailyai.com\\\/wp-content\\\/uploads\\\/2023\\\/11\\\/Simplify.jpg\",\"datePublished\":\"2023-11-23T15:10:15+00:00\",\"dateModified\":\"2023-11-23T15:13:38+00:00\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/dailyai.com\\\/2023\\\/11\\\/system-2-attention-improves-accuracy-of-llm-responses\\\/#breadcrumb\"},\"inLanguage\":\"de\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/dailyai.com\\\/2023\\\/11\\\/system-2-attention-improves-accuracy-of-llm-responses\\\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"de\",\"@id\":\"https:\\\/\\\/dailyai.com\\\/2023\\\/11\\\/system-2-attention-improves-accuracy-of-llm-responses\\\/#primaryimage\",\"url\":\"https:\\\/\\\/dailyai.com\\\/wp-content\\\/uploads\\\/2023\\\/11\\\/Simplify.jpg\",\"contentUrl\":\"https:\\\/\\\/dailyai.com\\\/wp-content\\\/uploads\\\/2023\\\/11\\\/Simplify.jpg\",\"width\":1000,\"height\":666},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/dailyai.com\\\/2023\\\/11\\\/system-2-attention-improves-accuracy-of-llm-responses\\\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/dailyai.com\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"System 2 Attention improves accuracy of LLM responses\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/dailyai.com\\\/#website\",\"url\":\"https:\\\/\\\/dailyai.com\\\/\",\"name\":\"DailyAI\",\"description\":\"Your Daily Dose of AI News\",\"publisher\":{\"@id\":\"https:\\\/\\\/dailyai.com\\\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/dailyai.com\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"de\"},{\"@type\":\"Organization\",\"@id\":\"https:\\\/\\\/dailyai.com\\\/#organization\",\"name\":\"DailyAI\",\"url\":\"https:\\\/\\\/dailyai.com\\\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"de\",\"@id\":\"https:\\\/\\\/dailyai.com\\\/#\\\/schema\\\/logo\\\/image\\\/\",\"url\":\"https:\\\/\\\/dailyai.com\\\/wp-content\\\/uploads\\\/2023\\\/06\\\/Daily-Ai_TL_colour.png\",\"contentUrl\":\"https:\\\/\\\/dailyai.com\\\/wp-content\\\/uploads\\\/2023\\\/06\\\/Daily-Ai_TL_colour.png\",\"width\":4501,\"height\":934,\"caption\":\"DailyAI\"},\"image\":{\"@id\":\"https:\\\/\\\/dailyai.com\\\/#\\\/schema\\\/logo\\\/image\\\/\"},\"sameAs\":[\"https:\\\/\\\/x.com\\\/DailyAIOfficial\",\"https:\\\/\\\/www.linkedin.com\\\/company\\\/dailyaiofficial\\\/\",\"https:\\\/\\\/www.youtube.com\\\/@DailyAIOfficial\"]},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/dailyai.com\\\/#\\\/schema\\\/person\\\/7ce525c6d0c79838b7cc7cde96993cfa\",\"name\":\"Eugene van der Watt\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"de\",\"@id\":\"https:\\\/\\\/dailyai.com\\\/wp-content\\\/uploads\\\/2023\\\/07\\\/Eugine_Profile_Picture-96x96.png\",\"url\":\"https:\\\/\\\/dailyai.com\\\/wp-content\\\/uploads\\\/2023\\\/07\\\/Eugine_Profile_Picture-96x96.png\",\"contentUrl\":\"https:\\\/\\\/dailyai.com\\\/wp-content\\\/uploads\\\/2023\\\/07\\\/Eugine_Profile_Picture-96x96.png\",\"caption\":\"Eugene van der Watt\"},\"description\":\"Eugene comes from an electronic engineering background and loves all things tech. When he takes a break from consuming AI news you'll find him at the snooker table.\",\"sameAs\":[\"www.linkedin.com\\\/in\\\/eugene-van-der-watt-16828119\"],\"url\":\"https:\\\/\\\/dailyai.com\\\/de\\\/author\\\/eugene\\\/\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"System 2 Aufmerksamkeit verbessert die Genauigkeit von LLM-Antworten | DailyAI","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/dailyai.com\/de\/2023\/11\/system-2-attention-improves-accuracy-of-llm-responses\/","og_locale":"de_DE","og_type":"article","og_title":"System 2 Attention improves accuracy of LLM responses | DailyAI","og_description":"Large Language Models (LLM) are often mislead by bias or irrelevant context in a prompt. Researchers at Meta have found a seemingly simple way to fix that. As context windows increase the prompts that we enter into an LLM can become longer and increasingly detailed. LLMs have become better at picking up on the nuances or smaller details in our prompts, but sometimes this can confuse them. Early machine learning used a \u201chard attention\u201d approach that singled out the most relevant part of an input and responded only to that. This works fine when you\u2019re trying to caption an image,","og_url":"https:\/\/dailyai.com\/de\/2023\/11\/system-2-attention-improves-accuracy-of-llm-responses\/","og_site_name":"DailyAI","article_published_time":"2023-11-23T15:10:15+00:00","article_modified_time":"2023-11-23T15:13:38+00:00","og_image":[{"width":1000,"height":666,"url":"https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/Simplify.jpg","type":"image\/jpeg"}],"author":"Eugene van der Watt","twitter_card":"summary_large_image","twitter_creator":"@DailyAIOfficial","twitter_site":"@DailyAIOfficial","twitter_misc":{"Verfasst von":"Eugene van der Watt","Gesch\u00e4tzte Lesezeit":"3\u00a0Minuten"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"NewsArticle","@id":"https:\/\/dailyai.com\/2023\/11\/system-2-attention-improves-accuracy-of-llm-responses\/#article","isPartOf":{"@id":"https:\/\/dailyai.com\/2023\/11\/system-2-attention-improves-accuracy-of-llm-responses\/"},"author":{"name":"Eugene van der Watt","@id":"https:\/\/dailyai.com\/#\/schema\/person\/7ce525c6d0c79838b7cc7cde96993cfa"},"headline":"System 2 Attention improves accuracy of LLM responses","datePublished":"2023-11-23T15:10:15+00:00","dateModified":"2023-11-23T15:13:38+00:00","mainEntityOfPage":{"@id":"https:\/\/dailyai.com\/2023\/11\/system-2-attention-improves-accuracy-of-llm-responses\/"},"wordCount":520,"publisher":{"@id":"https:\/\/dailyai.com\/#organization"},"image":{"@id":"https:\/\/dailyai.com\/2023\/11\/system-2-attention-improves-accuracy-of-llm-responses\/#primaryimage"},"thumbnailUrl":"https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/Simplify.jpg","keywords":["LLMS","Meta"],"articleSection":["Industry"],"inLanguage":"de"},{"@type":"WebPage","@id":"https:\/\/dailyai.com\/2023\/11\/system-2-attention-improves-accuracy-of-llm-responses\/","url":"https:\/\/dailyai.com\/2023\/11\/system-2-attention-improves-accuracy-of-llm-responses\/","name":"System 2 Aufmerksamkeit verbessert die Genauigkeit von LLM-Antworten | DailyAI","isPartOf":{"@id":"https:\/\/dailyai.com\/#website"},"primaryImageOfPage":{"@id":"https:\/\/dailyai.com\/2023\/11\/system-2-attention-improves-accuracy-of-llm-responses\/#primaryimage"},"image":{"@id":"https:\/\/dailyai.com\/2023\/11\/system-2-attention-improves-accuracy-of-llm-responses\/#primaryimage"},"thumbnailUrl":"https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/Simplify.jpg","datePublished":"2023-11-23T15:10:15+00:00","dateModified":"2023-11-23T15:13:38+00:00","breadcrumb":{"@id":"https:\/\/dailyai.com\/2023\/11\/system-2-attention-improves-accuracy-of-llm-responses\/#breadcrumb"},"inLanguage":"de","potentialAction":[{"@type":"ReadAction","target":["https:\/\/dailyai.com\/2023\/11\/system-2-attention-improves-accuracy-of-llm-responses\/"]}]},{"@type":"ImageObject","inLanguage":"de","@id":"https:\/\/dailyai.com\/2023\/11\/system-2-attention-improves-accuracy-of-llm-responses\/#primaryimage","url":"https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/Simplify.jpg","contentUrl":"https:\/\/dailyai.com\/wp-content\/uploads\/2023\/11\/Simplify.jpg","width":1000,"height":666},{"@type":"BreadcrumbList","@id":"https:\/\/dailyai.com\/2023\/11\/system-2-attention-improves-accuracy-of-llm-responses\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/dailyai.com\/"},{"@type":"ListItem","position":2,"name":"System 2 Attention improves accuracy of LLM responses"}]},{"@type":"WebSite","@id":"https:\/\/dailyai.com\/#website","url":"https:\/\/dailyai.com\/","name":"DailyAI","description":"Ihre t\u00e4gliche Dosis an AI-Nachrichten","publisher":{"@id":"https:\/\/dailyai.com\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/dailyai.com\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"de"},{"@type":"Organization","@id":"https:\/\/dailyai.com\/#organization","name":"DailyAI","url":"https:\/\/dailyai.com\/","logo":{"@type":"ImageObject","inLanguage":"de","@id":"https:\/\/dailyai.com\/#\/schema\/logo\/image\/","url":"https:\/\/dailyai.com\/wp-content\/uploads\/2023\/06\/Daily-Ai_TL_colour.png","contentUrl":"https:\/\/dailyai.com\/wp-content\/uploads\/2023\/06\/Daily-Ai_TL_colour.png","width":4501,"height":934,"caption":"DailyAI"},"image":{"@id":"https:\/\/dailyai.com\/#\/schema\/logo\/image\/"},"sameAs":["https:\/\/x.com\/DailyAIOfficial","https:\/\/www.linkedin.com\/company\/dailyaiofficial\/","https:\/\/www.youtube.com\/@DailyAIOfficial"]},{"@type":"Person","@id":"https:\/\/dailyai.com\/#\/schema\/person\/7ce525c6d0c79838b7cc7cde96993cfa","name":"Eugene van der Watt","image":{"@type":"ImageObject","inLanguage":"de","@id":"https:\/\/dailyai.com\/wp-content\/uploads\/2023\/07\/Eugine_Profile_Picture-96x96.png","url":"https:\/\/dailyai.com\/wp-content\/uploads\/2023\/07\/Eugine_Profile_Picture-96x96.png","contentUrl":"https:\/\/dailyai.com\/wp-content\/uploads\/2023\/07\/Eugine_Profile_Picture-96x96.png","caption":"Eugene van der Watt"},"description":"Eugene kommt aus der Elektronikbranche und liebt alles, was mit Technik zu tun hat. Wenn er eine Pause vom Konsum von KI-Nachrichten einlegt, findet man ihn am Snookertisch.","sameAs":["www.linkedin.com\/in\/eugene-van-der-watt-16828119"],"url":"https:\/\/dailyai.com\/de\/author\/eugene\/"}]}},"_links":{"self":[{"href":"https:\/\/dailyai.com\/de\/wp-json\/wp\/v2\/posts\/7671","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/dailyai.com\/de\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/dailyai.com\/de\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/dailyai.com\/de\/wp-json\/wp\/v2\/users\/6"}],"replies":[{"embeddable":true,"href":"https:\/\/dailyai.com\/de\/wp-json\/wp\/v2\/comments?post=7671"}],"version-history":[{"count":7,"href":"https:\/\/dailyai.com\/de\/wp-json\/wp\/v2\/posts\/7671\/revisions"}],"predecessor-version":[{"id":7682,"href":"https:\/\/dailyai.com\/de\/wp-json\/wp\/v2\/posts\/7671\/revisions\/7682"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/dailyai.com\/de\/wp-json\/wp\/v2\/media\/7676"}],"wp:attachment":[{"href":"https:\/\/dailyai.com\/de\/wp-json\/wp\/v2\/media?parent=7671"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/dailyai.com\/de\/wp-json\/wp\/v2\/categories?post=7671"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/dailyai.com\/de\/wp-json\/wp\/v2\/tags?post=7671"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}