{"id":12519,"date":"2024-05-22T18:16:07","date_gmt":"2024-05-22T18:16:07","guid":{"rendered":"https:\/\/dailyai.com\/?p=12519"},"modified":"2024-05-24T10:16:44","modified_gmt":"2024-05-24T10:16:44","slug":"decoding-the-ai-mind-anthropic-researchers-peer-inside-the-black-box","status":"publish","type":"post","link":"https:\/\/dailyai.com\/de\/2024\/05\/decoding-the-ai-mind-anthropic-researchers-peer-inside-the-black-box\/","title":{"rendered":"Entschl\u00fcsselung des KI-Gehirns: Anthropologische Forscher blicken in die \"Black Box\""},"content":{"rendered":"<p><b><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\">Anthropic<\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span> haben die Forscher erfolgreich Millionen von Konzepten innerhalb <span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\">Claude<\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span> Sonnet, einer ihrer fortgeschrittenen LLMs.<\/b><\/p>\n<p><span style=\"font-weight: 400;\">KI-Modelle werden oft als Blackboxen betrachtet, d. h. man kann nicht in sie hineinsehen, um zu verstehen, wie sie genau funktionieren.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Wenn Sie einem LLM eine Eingabe machen, erzeugt es eine Antwort, aber die Gr\u00fcnde f\u00fcr seine Entscheidungen sind nicht klar. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">Sie geben etwas ein, und das Ergebnis kommt heraus - und nicht einmal die KI-Entwickler selbst verstehen wirklich, was in dieser \"Box\" passiert.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Neuronale Netze erstellen ihre eigenen internen Repr\u00e4sentationen von Informationen, wenn sie w\u00e4hrend des Datentrainings Eingaben auf Ausgaben abbilden.\u00a0<\/span><span style=\"font-weight: 400;\">Die Bausteine dieses Prozesses, die so genannten \"Neuronenaktivierungen\", werden durch numerische Werte dargestellt. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">Jedes Konzept ist auf mehrere Neuronen verteilt, und jedes Neuron tr\u00e4gt zur Darstellung mehrerer Konzepte bei, was es schwierig macht, Konzepte direkt einzelnen Neuronen zuzuordnen.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Dies ist im Gro\u00dfen und Ganzen vergleichbar mit unseren menschlichen Gehirnen. So wie unser Gehirn Sinneseindr\u00fccke verarbeitet und Gedanken, Verhaltensweisen und Erinnerungen erzeugt, sind die Milliarden, ja Billionen von Prozessen, die hinter diesen Funktionen stehen, der Wissenschaft noch weitgehend unbekannt.<\/span><\/p>\n<p><a href=\"https:\/\/www.anthropic.com\/research\/mapping-mind-language-model\"><span style=\"font-weight: 400;\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\">Anthropic<\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span>Studie<\/span><\/a><span style=\"font-weight: 400;\"> versucht, mit einer Technik namens \"W\u00f6rterbuchlernen\" einen Blick in die Blackbox der KI zu werfen.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Dabei werden komplexe Muster in einem KI-Modell in lineare Bausteine oder \"Atome\" zerlegt, die f\u00fcr den Menschen intuitiv verst\u00e4ndlich sind.<\/span><\/p>\n<h2>Abbildung von LLMs mit Dictionary Learning<\/h2>\n<p><span style=\"font-weight: 400;\">Im Oktober 2023, <span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\">Anthropic<\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span> wandte diese Methode auf ein kleines \"Spielzeug\"-Sprachmodell an und fand koh\u00e4rente Merkmale, die Konzepten wie Gro\u00dfbuchstaben, DNA-Sequenzen, Nachnamen in Zitaten, mathematischen Substantiven oder Funktionsargumenten in Python-Code entsprechen.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">In dieser neuesten Studie wird die Technik so erweitert, dass sie auch f\u00fcr die gr\u00f6\u00dferen KI-Sprachmodelle von heute geeignet ist, <span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\">Anthropic<\/span><\/span><\/span><\/span>'s <span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\">Claude<\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span> 3 Sonett.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Im Folgenden wird Schritt f\u00fcr Schritt beschrieben, wie die Studie ablief:<\/span><\/p>\n<h4><b>Erkennen von Mustern mit W\u00f6rterbuchlernen<\/b><\/h4>\n<p><span style=\"font-weight: 400;\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\">Anthropic<\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span> verwendet W\u00f6rterbuchlernen, um Neuronenaktivierungen in verschiedenen Kontexten zu analysieren und gemeinsame Muster zu erkennen. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">Beim W\u00f6rterbuchlernen werden diese Aktivierungen zu einer kleineren Menge aussagekr\u00e4ftiger \"Merkmale\" zusammengefasst, die vom Modell erlernte Konzepte auf h\u00f6herer Ebene darstellen. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">Durch die Identifizierung dieser Merkmale k\u00f6nnen die Forscher besser verstehen, wie das Modell Informationen verarbeitet und darstellt.<\/span><\/p>\n<h4><b>Extraktion von Merkmalen aus der mittleren Schicht<\/b><\/h4>\n<p><span style=\"font-weight: 400;\">Die Forscher konzentrierten sich auf die mittlere Schicht der <span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\">Claude<\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span> 3.0 Sonnet, das als kritischer Punkt in der Verarbeitungspipeline des Modells dient. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">Durch die Anwendung des W\u00f6rterbuchlernens auf diese Ebene werden Millionen von Merkmalen extrahiert, die die internen Repr\u00e4sentationen des Modells und die gelernten Konzepte in diesem Stadium erfassen. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">Die Extraktion von Merkmalen aus der mittleren Schicht erm\u00f6glicht es den Forschern, das Informationsverst\u00e4ndnis des Modells zu untersuchen <em>nach<\/em> er hat die Eingabe verarbeitet <em>vor<\/em> die Erzeugung der endg\u00fcltigen Ausgabe.<\/span><\/p>\n<h4><b>Entdeckung vielf\u00e4ltiger und abstrakter Konzepte<\/b><\/h4>\n<p><span style=\"font-weight: 400;\">Die extrahierten Merkmale zeigten eine breite Palette von Konzepten, die von <span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\">Claude<\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span>von konkreten Entit\u00e4ten wie St\u00e4dten und Menschen bis hin zu abstrakten Begriffen, die sich auf wissenschaftliche Bereiche und Programmiersyntax beziehen. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">Interessanterweise erwiesen sich die Merkmale als multimodal und reagierten sowohl auf textuelle als auch auf visuelle Eingaben, was darauf hindeutet, dass das Modell Konzepte \u00fcber verschiedene Modalit\u00e4ten hinweg lernen und darstellen kann. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">Dar\u00fcber hinaus deuten die mehrsprachigen Merkmale darauf hin, dass das Modell Konzepte erfassen kann, die in verschiedenen Sprachen ausgedr\u00fcckt werden.<\/span><\/p>\n<p><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"598\" class=\"wp-image-12521 size-large\" src=\"https:\/\/dailyai.com\/wp-content\/uploads\/2024\/05\/c896a301ad3d1ef4237cb05b68d78b467c444097-2200x1284-1-1024x598.webp\" alt=\"&lt;span class=\" srcset=\"https:\/\/dailyai.com\/wp-content\/uploads\/2024\/05\/c896a301ad3d1ef4237cb05b68d78b467c444097-2200x1284-1-1024x598.webp 1024w, https:\/\/dailyai.com\/wp-content\/uploads\/2024\/05\/c896a301ad3d1ef4237cb05b68d78b467c444097-2200x1284-1-300x175.webp 300w, https:\/\/dailyai.com\/wp-content\/uploads\/2024\/05\/c896a301ad3d1ef4237cb05b68d78b467c444097-2200x1284-1-768x448.webp 768w, https:\/\/dailyai.com\/wp-content\/uploads\/2024\/05\/c896a301ad3d1ef4237cb05b68d78b467c444097-2200x1284-1-1536x896.webp 1536w, https:\/\/dailyai.com\/wp-content\/uploads\/2024\/05\/c896a301ad3d1ef4237cb05b68d78b467c444097-2200x1284-1-2048x1195.webp 2048w, https:\/\/dailyai.com\/wp-content\/uploads\/2024\/05\/c896a301ad3d1ef4237cb05b68d78b467c444097-2200x1284-1-18x12.webp 18w, https:\/\/dailyai.com\/wp-content\/uploads\/2024\/05\/c896a301ad3d1ef4237cb05b68d78b467c444097-2200x1284-1-60x35.webp 60w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><\/p>\n<h3><b>Analyse der Organisation von Konzepten<\/b><\/h3>\n<p><span style=\"font-weight: 400;\">Um zu verstehen, wie das Modell die verschiedenen Konzepte organisiert und miteinander in Beziehung setzt, analysierten die Forscher die \u00c4hnlichkeit zwischen den Merkmalen auf der Grundlage ihrer Aktivierungsmuster. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">Sie entdeckten, dass Merkmale, die verwandte Konzepte darstellen, dazu neigen, sich zu gruppieren. So wiesen beispielsweise Merkmale, die mit St\u00e4dten oder wissenschaftlichen Disziplinen in Verbindung stehen, eine gr\u00f6\u00dfere \u00c4hnlichkeit untereinander auf als Merkmale, die nicht verwandte Konzepte darstellen. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">Dies deutet darauf hin, dass die interne Organisation der Konzepte des Modells bis zu einem gewissen Grad mit der menschlichen Intuition \u00fcber konzeptionelle Beziehungen \u00fcbereinstimmt.<\/span><\/p>\n<figure id=\"attachment_12520\" aria-describedby=\"caption-attachment-12520\" style=\"width: 1024px\" class=\"wp-caption aligncenter\"><img loading=\"lazy\" decoding=\"async\" class=\"wp-image-12520 size-large\" src=\"https:\/\/dailyai.com\/wp-content\/uploads\/2024\/05\/fe4d42c004bf43efda0f5921adfedd2f8f42e417-2200x2140-1-1024x996.webp\" alt=\"anthropisch\" width=\"1024\" height=\"996\" srcset=\"https:\/\/dailyai.com\/wp-content\/uploads\/2024\/05\/fe4d42c004bf43efda0f5921adfedd2f8f42e417-2200x2140-1-1024x996.webp 1024w, https:\/\/dailyai.com\/wp-content\/uploads\/2024\/05\/fe4d42c004bf43efda0f5921adfedd2f8f42e417-2200x2140-1-300x292.webp 300w, https:\/\/dailyai.com\/wp-content\/uploads\/2024\/05\/fe4d42c004bf43efda0f5921adfedd2f8f42e417-2200x2140-1-768x747.webp 768w, https:\/\/dailyai.com\/wp-content\/uploads\/2024\/05\/fe4d42c004bf43efda0f5921adfedd2f8f42e417-2200x2140-1-1536x1494.webp 1536w, https:\/\/dailyai.com\/wp-content\/uploads\/2024\/05\/fe4d42c004bf43efda0f5921adfedd2f8f42e417-2200x2140-1-2048x1992.webp 2048w, https:\/\/dailyai.com\/wp-content\/uploads\/2024\/05\/fe4d42c004bf43efda0f5921adfedd2f8f42e417-2200x2140-1-12x12.webp 12w, https:\/\/dailyai.com\/wp-content\/uploads\/2024\/05\/fe4d42c004bf43efda0f5921adfedd2f8f42e417-2200x2140-1-60x58.webp 60w, https:\/\/dailyai.com\/wp-content\/uploads\/2024\/05\/fe4d42c004bf43efda0f5921adfedd2f8f42e417-2200x2140-1-24x24.webp 24w, https:\/\/dailyai.com\/wp-content\/uploads\/2024\/05\/fe4d42c004bf43efda0f5921adfedd2f8f42e417-2200x2140-1-48x48.webp 48w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><figcaption id=\"caption-attachment-12520\" class=\"wp-caption-text\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\">Anthropic<\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span> es geschafft, abstrakte Begriffe wie \"innerer Konflikt\" abzubilden. Quelle: <a href=\"https:\/\/www.anthropic.com\/research\/mapping-mind-language-model\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\">Anthropic<\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/a>.<\/figcaption><\/figure>\n<h4><b>\u00dcberpr\u00fcfen der Merkmale<\/b><\/h4>\n<p><span style=\"font-weight: 400;\">Um zu best\u00e4tigen, dass die identifizierten Merkmale einen direkten Einfluss auf das Verhalten und die Ergebnisse des Modells haben, f\u00fchrten die Forscher Experimente zur \"Merkmalssteuerung\" durch. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">Dabei wurde die Aktivierung bestimmter Merkmale w\u00e4hrend der Verarbeitung des Modells selektiv verst\u00e4rkt oder unterdr\u00fcckt und die Auswirkungen auf die Reaktionen des Modells beobachtet. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">Durch die Manipulation einzelner Merkmale konnten die Forscher eine direkte Verbindung zwischen einzelnen Merkmalen und dem Verhalten des Modells herstellen. So f\u00fchrte beispielsweise die Verst\u00e4rkung eines Merkmals, das sich auf eine bestimmte Stadt bezog, dazu, dass das Modell auch in irrelevanten Kontexten stadtbezogene Ergebnisse erzeugte.<\/span><\/p>\n<p>Lesen Sie die vollst\u00e4ndige Studie <a href=\"https:\/\/transformer-circuits.pub\/2024\/scaling-monosemanticity\/index.html\" target=\"_blank\" rel=\"noopener\">hier<\/a>.<\/p>\n<p><iframe loading=\"lazy\" title=\"W\u00f6rterbuchlernen zu Claude 3 Sonnet\" width=\"1080\" height=\"608\" src=\"https:\/\/www.youtube.com\/embed\/CJIbCV92d88?feature=oembed\" frameborder=\"0\" allow=\"accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share\" referrerpolicy=\"strict-origin-when-cross-origin\" allowfullscreen><\/iframe><\/p>\n<h2>Warum Interpretierbarkeit f\u00fcr die KI-Sicherheit entscheidend ist<\/h2>\n<p><span style=\"font-weight: 400;\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\">Anthropic<\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span>Die Forschungen der Gruppe sind von grundlegender Bedeutung f\u00fcr die Interpretierbarkeit von KI und damit auch f\u00fcr die Sicherheit.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Zu verstehen, wie LLMs Informationen verarbeiten und darstellen, hilft Forschern, Risiken zu verstehen und zu mindern. Es <\/span><span style=\"font-weight: 400;\">legt den Grundstein f\u00fcr die Entwicklung transparenter und erkl\u00e4rbarer KI-Systeme.\u00a0<\/span><\/p>\n<p>Als <span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\">Anthropic<\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span> erkl\u00e4rt: \"Wir hoffen, dass wir und andere diese Entdeckungen nutzen k\u00f6nnen, um Modelle sicherer zu machen. Zum Beispiel k\u00f6nnte es m\u00f6glich sein, die hier beschriebenen Techniken zu nutzen, um KI-Systeme auf bestimmte gef\u00e4hrliche Verhaltensweisen zu \u00fcberwachen (z. B. T\u00e4uschung des Nutzers), sie auf w\u00fcnschenswerte Ergebnisse zu lenken (Debiasierung) oder bestimmte gef\u00e4hrliche Themen ganz zu entfernen.<\/p>\n<p><span style=\"font-weight: 400;\">Ein besseres Verst\u00e4ndnis des KI-Verhaltens ist von entscheidender Bedeutung, da KI in kritischen Entscheidungsprozessen in Bereichen wie dem Gesundheitswesen, dem Finanzwesen und der Strafjustiz allgegenw\u00e4rtig ist. Es hilft auch bei der Aufdeckung der Ursache von <a href=\"https:\/\/dailyai.com\/de\/2023\/06\/anthropic-releases-paper-highlighting-the-potential-bias-of-large-language-models\/\">Vorspannung<\/a>Halluzinationen und andere unerw\u00fcnschte oder unvorhersehbare Verhaltensweisen.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Zum Beispiel, ein <\/span><a href=\"https:\/\/dailyai.com\/de\/2024\/01\/researchers-aim-to-determine-how-black-box-drug-discovery-models-work\/\"><span style=\"font-weight: 400;\">aktuelle Studie<\/span><\/a><span style=\"font-weight: 400;\"> von der Universit\u00e4t Bonn haben aufgedeckt, dass Graphische Neuronale Netze (GNNs), die f\u00fcr die Entdeckung von Medikamenten verwendet werden, sich stark auf das Abrufen von \u00c4hnlichkeiten aus Trainingsdaten verlassen, anstatt wirklich komplexe neue chemische Interaktionen zu lernen. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">Das macht es schwierig zu verstehen, wie genau diese Modelle neue Verbindungen von Interesse bestimmen.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Letztes Jahr hat die <\/span><a href=\"https:\/\/dailyai.com\/de\/2023\/09\/the-uk-government-wants-to-see-inside-ais-black-box\/\"><span style=\"font-weight: 400;\">Die britische Regierung verhandelte<\/span><\/a><span style=\"font-weight: 400;\"> mit gro\u00dfen Tech-Giganten wie <span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\">OpenAI<\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span> und <span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\">DeepMind<\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span>die Zugang zu den internen Entscheidungsprozessen ihrer KI-Systeme suchen.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Verordnung wie die <a href=\"https:\/\/dailyai.com\/de\/2024\/03\/the-eu-ai-act-passed-in-a-landslide-and-will-come-into-force-this-year\/\">Das KI-Gesetz der EU<\/a> wird die KI-Unternehmen zu mehr Transparenz zwingen, auch wenn Gesch\u00e4ftsgeheimnisse sicher unter Verschluss bleiben werden.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\">Anthropic<\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span>Die Forschungen der Universit\u00e4t M\u00fcnchen geben einen Einblick in das Innere der Box, indem sie Informationen \u00fcber das Modell \"abbilden\".\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Die Wahrheit ist jedoch, dass diese Modelle so umfangreich sind, dass sie <span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\"><span class=\"noTranslate\" data-no-translation=\"\">Anthropic<\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span>Wir halten es f\u00fcr ziemlich wahrscheinlich, dass wir um Gr\u00f6\u00dfenordnungen zu kurz kommen und dass wir, wenn wir alle Merkmale - in allen Schichten! - sehr viel mehr Rechenleistung ben\u00f6tigen w\u00fcrden, als f\u00fcr das Training der zugrunde liegenden Modelle erforderlich ist.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Das ist ein interessanter Punkt - das Reverse Engineering eines Modells ist rechnerisch komplexer als das Engineering des Modells selbst.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Es erinnert an enorm teure neurowissenschaftliche Projekte wie das <a href=\"https:\/\/dailyai.com\/de\/2023\/08\/exploring-impact-of-the-human-brain-project-hbp-on-science-and-ai\/\">Human Brain Project (HBP)<\/a>die Milliarden in die Kartierung unserer eigenen menschlichen Gehirne gesteckt hat, nur um letztendlich zu scheitern.\u00a0<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Untersch\u00e4tzen Sie nie, wie viel im Inneren der Blackbox steckt.\u00a0<\/span><\/p>","protected":false},"excerpt":{"rendered":"<p>Anthropic-Forscher identifizierten erfolgreich Millionen von Konzepten in Claude Sonnet, einem ihrer fortgeschrittenen LLMs. KI-Modelle werden oft als \"Black Boxes\" betrachtet, was bedeutet, dass man nicht in sie hineinsehen kann, um zu verstehen, wie sie genau funktionieren.  Wenn Sie ein LLM mit einer Eingabe versorgen, erzeugt es eine Antwort, aber die Gr\u00fcnde f\u00fcr seine Entscheidungen sind nicht klar. Die Eingabe geht hinein, und die Ausgabe kommt heraus - und nicht einmal die KI-Entwickler selbst verstehen wirklich, was in dieser \"Box\" passiert.  Neuronale Netze erstellen ihre eigenen internen Repr\u00e4sentationen von Informationen, wenn sie w\u00e4hrend des Datentrainings Eingaben auf Ausgaben abbilden. Die Bausteine dieser<\/p>","protected":false},"author":2,"featured_media":12522,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[84],"tags":[148,105],"class_list":["post-12519","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-industry","tag-anthropic","tag-machine-learning"],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v27.4 - https:\/\/yoast.com\/product\/yoast-seo-wordpress\/ -->\n<title>Decoding the AI mind: Anthropic researchers peer inside the &quot;black box&quot; | 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\/2024\/05\/decoding-the-ai-mind-anthropic-researchers-peer-inside-the-black-box\/\" \/>\n<meta property=\"og:locale\" content=\"de_DE\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Decoding the AI mind: Anthropic researchers peer inside the &quot;black box&quot; | DailyAI\" \/>\n<meta property=\"og:description\" content=\"Anthropic researchers successfully identified millions of concepts within Claude Sonnet, one of their advanced LLMs. AI models are often considered black boxes, meaning you can&#8217;t &#8216;see&#8217; inside them to understand exactly how they work.\u00a0 When you provide an LLM with an input, it generates a response, but the reasoning behind its choices isn\u2019t clear. Your input goes in, and the output comes out \u2013 and even the AI developers themselves don&#8217;t truly understand what happens inside that &#8216;box.&#8217;\u00a0 Neural networks create their own internal representations of information when they map inputs to outputs during data training.\u00a0The building blocks of this\" \/>\n<meta property=\"og:url\" content=\"https:\/\/dailyai.com\/de\/2024\/05\/decoding-the-ai-mind-anthropic-researchers-peer-inside-the-black-box\/\" \/>\n<meta property=\"og:site_name\" content=\"DailyAI\" \/>\n<meta property=\"article:published_time\" content=\"2024-05-22T18:16:07+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2024-05-24T10:16:44+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/dailyai.com\/wp-content\/uploads\/2024\/05\/DALL\u00b7E-2024-05-22-19.15.50-Anthropic-researchers-studying-artificial-intelligence-inside-a-high-tech-laboratory-with-a-prominent-black-box-in-the-center.-The-scene-features-adva.webp\" \/>\n\t<meta property=\"og:image:width\" content=\"1792\" \/>\n\t<meta property=\"og:image:height\" content=\"1024\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/webp\" \/>\n<meta name=\"author\" content=\"Sam Jeans\" \/>\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=\"Sam Jeans\" \/>\n\t<meta name=\"twitter:label2\" content=\"Gesch\u00e4tzte Lesezeit\" \/>\n\t<meta name=\"twitter:data2\" content=\"6\u00a0Minuten\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"NewsArticle\",\"@id\":\"https:\\\/\\\/dailyai.com\\\/2024\\\/05\\\/decoding-the-ai-mind-anthropic-researchers-peer-inside-the-black-box\\\/#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/dailyai.com\\\/2024\\\/05\\\/decoding-the-ai-mind-anthropic-researchers-peer-inside-the-black-box\\\/\"},\"author\":{\"name\":\"Sam Jeans\",\"@id\":\"https:\\\/\\\/dailyai.com\\\/#\\\/schema\\\/person\\\/711e81f945549438e8bbc579efdeb3c9\"},\"headline\":\"Decoding the AI mind: Anthropic researchers peer inside the &#8220;black box&#8221;\",\"datePublished\":\"2024-05-22T18:16:07+00:00\",\"dateModified\":\"2024-05-24T10:16:44+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/dailyai.com\\\/2024\\\/05\\\/decoding-the-ai-mind-anthropic-researchers-peer-inside-the-black-box\\\/\"},\"wordCount\":1075,\"publisher\":{\"@id\":\"https:\\\/\\\/dailyai.com\\\/#organization\"},\"image\":{\"@id\":\"https:\\\/\\\/dailyai.com\\\/2024\\\/05\\\/decoding-the-ai-mind-anthropic-researchers-peer-inside-the-black-box\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/dailyai.com\\\/wp-content\\\/uploads\\\/2024\\\/05\\\/DALL\u00b7E-2024-05-22-19.15.50-Anthropic-researchers-studying-artificial-intelligence-inside-a-high-tech-laboratory-with-a-prominent-black-box-in-the-center.-The-scene-features-adva.webp\",\"keywords\":[\"Anthropic\",\"machine learning\"],\"articleSection\":[\"Industry\"],\"inLanguage\":\"de\"},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/dailyai.com\\\/2024\\\/05\\\/decoding-the-ai-mind-anthropic-researchers-peer-inside-the-black-box\\\/\",\"url\":\"https:\\\/\\\/dailyai.com\\\/2024\\\/05\\\/decoding-the-ai-mind-anthropic-researchers-peer-inside-the-black-box\\\/\",\"name\":\"Decoding the AI mind: Anthropic researchers peer inside the \\\"black box\\\" | DailyAI\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/dailyai.com\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/dailyai.com\\\/2024\\\/05\\\/decoding-the-ai-mind-anthropic-researchers-peer-inside-the-black-box\\\/#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/dailyai.com\\\/2024\\\/05\\\/decoding-the-ai-mind-anthropic-researchers-peer-inside-the-black-box\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/dailyai.com\\\/wp-content\\\/uploads\\\/2024\\\/05\\\/DALL\u00b7E-2024-05-22-19.15.50-Anthropic-researchers-studying-artificial-intelligence-inside-a-high-tech-laboratory-with-a-prominent-black-box-in-the-center.-The-scene-features-adva.webp\",\"datePublished\":\"2024-05-22T18:16:07+00:00\",\"dateModified\":\"2024-05-24T10:16:44+00:00\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/dailyai.com\\\/2024\\\/05\\\/decoding-the-ai-mind-anthropic-researchers-peer-inside-the-black-box\\\/#breadcrumb\"},\"inLanguage\":\"de\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/dailyai.com\\\/2024\\\/05\\\/decoding-the-ai-mind-anthropic-researchers-peer-inside-the-black-box\\\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"de\",\"@id\":\"https:\\\/\\\/dailyai.com\\\/2024\\\/05\\\/decoding-the-ai-mind-anthropic-researchers-peer-inside-the-black-box\\\/#primaryimage\",\"url\":\"https:\\\/\\\/dailyai.com\\\/wp-content\\\/uploads\\\/2024\\\/05\\\/DALL\u00b7E-2024-05-22-19.15.50-Anthropic-researchers-studying-artificial-intelligence-inside-a-high-tech-laboratory-with-a-prominent-black-box-in-the-center.-The-scene-features-adva.webp\",\"contentUrl\":\"https:\\\/\\\/dailyai.com\\\/wp-content\\\/uploads\\\/2024\\\/05\\\/DALL\u00b7E-2024-05-22-19.15.50-Anthropic-researchers-studying-artificial-intelligence-inside-a-high-tech-laboratory-with-a-prominent-black-box-in-the-center.-The-scene-features-adva.webp\",\"width\":1792,\"height\":1024},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/dailyai.com\\\/2024\\\/05\\\/decoding-the-ai-mind-anthropic-researchers-peer-inside-the-black-box\\\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/dailyai.com\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Decoding the AI mind: Anthropic researchers peer inside the &#8220;black box&#8221;\"}]},{\"@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\\\/711e81f945549438e8bbc579efdeb3c9\",\"name\":\"Sam Jeans\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"de\",\"@id\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/a24a4a8f8e2a1a275b7491dc9c9f032c401eabf23c3206da4628dc84b6dac5c8?s=96&d=robohash&r=g\",\"url\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/a24a4a8f8e2a1a275b7491dc9c9f032c401eabf23c3206da4628dc84b6dac5c8?s=96&d=robohash&r=g\",\"contentUrl\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/a24a4a8f8e2a1a275b7491dc9c9f032c401eabf23c3206da4628dc84b6dac5c8?s=96&d=robohash&r=g\",\"caption\":\"Sam Jeans\"},\"description\":\"Sam is a science and technology writer who has worked in various AI startups. When he\u2019s not writing, he can be found reading medical journals or digging through boxes of vinyl records.\",\"sameAs\":[\"https:\\\/\\\/www.linkedin.com\\\/in\\\/sam-jeans-6746b9142\\\/\"],\"url\":\"https:\\\/\\\/dailyai.com\\\/de\\\/author\\\/samjeans\\\/\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"Entschl\u00fcsselung des KI-Verstandes: Anthropologische Forscher blicken in die \"Black Box\" | 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\/2024\/05\/decoding-the-ai-mind-anthropic-researchers-peer-inside-the-black-box\/","og_locale":"de_DE","og_type":"article","og_title":"Decoding the AI mind: Anthropic researchers peer inside the \"black box\" | DailyAI","og_description":"Anthropic researchers successfully identified millions of concepts within Claude Sonnet, one of their advanced LLMs. AI models are often considered black boxes, meaning you can&#8217;t &#8216;see&#8217; inside them to understand exactly how they work.\u00a0 When you provide an LLM with an input, it generates a response, but the reasoning behind its choices isn\u2019t clear. Your input goes in, and the output comes out \u2013 and even the AI developers themselves don&#8217;t truly understand what happens inside that &#8216;box.&#8217;\u00a0 Neural networks create their own internal representations of information when they map inputs to outputs during data training.\u00a0The building blocks of this","og_url":"https:\/\/dailyai.com\/de\/2024\/05\/decoding-the-ai-mind-anthropic-researchers-peer-inside-the-black-box\/","og_site_name":"DailyAI","article_published_time":"2024-05-22T18:16:07+00:00","article_modified_time":"2024-05-24T10:16:44+00:00","og_image":[{"width":1792,"height":1024,"url":"https:\/\/dailyai.com\/wp-content\/uploads\/2024\/05\/DALL\u00b7E-2024-05-22-19.15.50-Anthropic-researchers-studying-artificial-intelligence-inside-a-high-tech-laboratory-with-a-prominent-black-box-in-the-center.-The-scene-features-adva.webp","type":"image\/webp"}],"author":"Sam Jeans","twitter_card":"summary_large_image","twitter_creator":"@DailyAIOfficial","twitter_site":"@DailyAIOfficial","twitter_misc":{"Verfasst von":"Sam Jeans","Gesch\u00e4tzte Lesezeit":"6\u00a0Minuten"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"NewsArticle","@id":"https:\/\/dailyai.com\/2024\/05\/decoding-the-ai-mind-anthropic-researchers-peer-inside-the-black-box\/#article","isPartOf":{"@id":"https:\/\/dailyai.com\/2024\/05\/decoding-the-ai-mind-anthropic-researchers-peer-inside-the-black-box\/"},"author":{"name":"Sam Jeans","@id":"https:\/\/dailyai.com\/#\/schema\/person\/711e81f945549438e8bbc579efdeb3c9"},"headline":"Decoding the AI mind: Anthropic researchers peer inside the &#8220;black box&#8221;","datePublished":"2024-05-22T18:16:07+00:00","dateModified":"2024-05-24T10:16:44+00:00","mainEntityOfPage":{"@id":"https:\/\/dailyai.com\/2024\/05\/decoding-the-ai-mind-anthropic-researchers-peer-inside-the-black-box\/"},"wordCount":1075,"publisher":{"@id":"https:\/\/dailyai.com\/#organization"},"image":{"@id":"https:\/\/dailyai.com\/2024\/05\/decoding-the-ai-mind-anthropic-researchers-peer-inside-the-black-box\/#primaryimage"},"thumbnailUrl":"https:\/\/dailyai.com\/wp-content\/uploads\/2024\/05\/DALL\u00b7E-2024-05-22-19.15.50-Anthropic-researchers-studying-artificial-intelligence-inside-a-high-tech-laboratory-with-a-prominent-black-box-in-the-center.-The-scene-features-adva.webp","keywords":["Anthropic","machine learning"],"articleSection":["Industry"],"inLanguage":"de"},{"@type":"WebPage","@id":"https:\/\/dailyai.com\/2024\/05\/decoding-the-ai-mind-anthropic-researchers-peer-inside-the-black-box\/","url":"https:\/\/dailyai.com\/2024\/05\/decoding-the-ai-mind-anthropic-researchers-peer-inside-the-black-box\/","name":"Entschl\u00fcsselung des KI-Verstandes: Anthropologische Forscher blicken in die \"Black Box\" | DailyAI","isPartOf":{"@id":"https:\/\/dailyai.com\/#website"},"primaryImageOfPage":{"@id":"https:\/\/dailyai.com\/2024\/05\/decoding-the-ai-mind-anthropic-researchers-peer-inside-the-black-box\/#primaryimage"},"image":{"@id":"https:\/\/dailyai.com\/2024\/05\/decoding-the-ai-mind-anthropic-researchers-peer-inside-the-black-box\/#primaryimage"},"thumbnailUrl":"https:\/\/dailyai.com\/wp-content\/uploads\/2024\/05\/DALL\u00b7E-2024-05-22-19.15.50-Anthropic-researchers-studying-artificial-intelligence-inside-a-high-tech-laboratory-with-a-prominent-black-box-in-the-center.-The-scene-features-adva.webp","datePublished":"2024-05-22T18:16:07+00:00","dateModified":"2024-05-24T10:16:44+00:00","breadcrumb":{"@id":"https:\/\/dailyai.com\/2024\/05\/decoding-the-ai-mind-anthropic-researchers-peer-inside-the-black-box\/#breadcrumb"},"inLanguage":"de","potentialAction":[{"@type":"ReadAction","target":["https:\/\/dailyai.com\/2024\/05\/decoding-the-ai-mind-anthropic-researchers-peer-inside-the-black-box\/"]}]},{"@type":"ImageObject","inLanguage":"de","@id":"https:\/\/dailyai.com\/2024\/05\/decoding-the-ai-mind-anthropic-researchers-peer-inside-the-black-box\/#primaryimage","url":"https:\/\/dailyai.com\/wp-content\/uploads\/2024\/05\/DALL\u00b7E-2024-05-22-19.15.50-Anthropic-researchers-studying-artificial-intelligence-inside-a-high-tech-laboratory-with-a-prominent-black-box-in-the-center.-The-scene-features-adva.webp","contentUrl":"https:\/\/dailyai.com\/wp-content\/uploads\/2024\/05\/DALL\u00b7E-2024-05-22-19.15.50-Anthropic-researchers-studying-artificial-intelligence-inside-a-high-tech-laboratory-with-a-prominent-black-box-in-the-center.-The-scene-features-adva.webp","width":1792,"height":1024},{"@type":"BreadcrumbList","@id":"https:\/\/dailyai.com\/2024\/05\/decoding-the-ai-mind-anthropic-researchers-peer-inside-the-black-box\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/dailyai.com\/"},{"@type":"ListItem","position":2,"name":"Decoding the AI mind: Anthropic researchers peer inside the &#8220;black box&#8221;"}]},{"@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\/711e81f945549438e8bbc579efdeb3c9","name":"Sam Jeans","image":{"@type":"ImageObject","inLanguage":"de","@id":"https:\/\/secure.gravatar.com\/avatar\/a24a4a8f8e2a1a275b7491dc9c9f032c401eabf23c3206da4628dc84b6dac5c8?s=96&d=robohash&r=g","url":"https:\/\/secure.gravatar.com\/avatar\/a24a4a8f8e2a1a275b7491dc9c9f032c401eabf23c3206da4628dc84b6dac5c8?s=96&d=robohash&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/a24a4a8f8e2a1a275b7491dc9c9f032c401eabf23c3206da4628dc84b6dac5c8?s=96&d=robohash&r=g","caption":"Sam Jeans"},"description":"Sam ist ein Wissenschafts- und Technologiewissenschaftler, der in verschiedenen KI-Startups gearbeitet hat. Wenn er nicht gerade schreibt, liest er medizinische Fachzeitschriften oder kramt in Kisten mit Schallplatten.","sameAs":["https:\/\/www.linkedin.com\/in\/sam-jeans-6746b9142\/"],"url":"https:\/\/dailyai.com\/de\/author\/samjeans\/"}]}},"_links":{"self":[{"href":"https:\/\/dailyai.com\/de\/wp-json\/wp\/v2\/posts\/12519","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\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/dailyai.com\/de\/wp-json\/wp\/v2\/comments?post=12519"}],"version-history":[{"count":11,"href":"https:\/\/dailyai.com\/de\/wp-json\/wp\/v2\/posts\/12519\/revisions"}],"predecessor-version":[{"id":12573,"href":"https:\/\/dailyai.com\/de\/wp-json\/wp\/v2\/posts\/12519\/revisions\/12573"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/dailyai.com\/de\/wp-json\/wp\/v2\/media\/12522"}],"wp:attachment":[{"href":"https:\/\/dailyai.com\/de\/wp-json\/wp\/v2\/media?parent=12519"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/dailyai.com\/de\/wp-json\/wp\/v2\/categories?post=12519"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/dailyai.com\/de\/wp-json\/wp\/v2\/tags?post=12519"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}